<?xml version="1.0"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="zh-Hant-TW">
	<id>https://tommy0804.ddns.net/api.php?action=feedcontributions&amp;feedformat=atom&amp;user=Tommy</id>
	<title>flip the world - 使用者貢獻 [zh-tw]</title>
	<link rel="self" type="application/atom+xml" href="https://tommy0804.ddns.net/api.php?action=feedcontributions&amp;feedformat=atom&amp;user=Tommy"/>
	<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php/%E7%89%B9%E6%AE%8A:%E4%BD%BF%E7%94%A8%E8%80%85%E8%B2%A2%E7%8D%BB/Tommy"/>
	<updated>2026-11-20T20:56:30Z</updated>
	<subtitle>使用者貢獻</subtitle>
	<generator>MediaWiki 1.43.9</generator>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=862</id>
		<title>Homebrewing</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=862"/>
		<updated>2026-11-20T15:39:44Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​#0012 section: drop duplicate glass photo (kept in top gallery)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&#039;&#039;&#039;Homebrewing&#039;&#039;&#039; — eleven batches of hazy / New England IPA brewed at home in 10-litre batches, December 2025 to October 2026. Every batch is logged with gravity readings, pH, hop timing and a tasting note, and each recipe is version-controlled in git so every change can be traced back to the batch that caused it. This page summarises what was brewed, what went wrong, and the process rules that came out of it.&lt;br /&gt;
&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_0012_glass.jpg|Orange Glow v2.6 (#0012), the best batch so far&lt;br /&gt;
File:Homebrew_glass_orange_glow.jpg|Orange Glow v2.3 (#0007), five weeks after brew day&lt;br /&gt;
File:Homebrew_glass_in_hand.jpg|A finished hazy IPA, in the glass&lt;br /&gt;
File:Homebrew_keg_and_bottles.jpg|Keg and bottles, with the Strange Haze label in front&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Batches !! Period !! Batch size !! Styles !! Recipes (current version)&lt;br /&gt;
|-&lt;br /&gt;
| 11 (12 brews) || 2025-12 → 2026-10 || 6–11.5 L || Hazy IPA, DDH NEIPA, Milkshake IPA || 妃子起笑 Crazy Concubine (v4.4), Orange Glow (v2.6), Strange Haze (v1.1), Milkshake IPA (v1)&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The data ==&lt;br /&gt;
[[File:Homebrew_og_fg_by_batch.png|900px|center|Original and final gravity of every batch]]&lt;br /&gt;
The first six batches had no written recipe and wild original gravities (up to 1.110 — a 10.5% &amp;quot;IPA&amp;quot;). From #0007 on, every batch follows a versioned recipe and the OG has been pulled back into the 1.050–1.076 range the recipes are designed for.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_abv_attenuation.png|900px|center|ABV and apparent attenuation per batch]]&lt;br /&gt;
Attenuation went from ~71% (early batches, under-pitched, no oxygen) to a consistent 81–88% once pure-oxygen pitching, yeast nutrient and modern yeast strains were adopted. #0011&#039;s 62% is not a stuck fermentation: it is a kit recipe with 50 g of unfermentable lactose, and it landed exactly on the kit&#039;s printed OG 1.052 / FG 1.020.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_gravity_curves.png|900px|center|Fermentation curves of four batches]]&lt;br /&gt;
The single most useful chart. With one pack of Verdant and no oxygen (#0006) fermentation drags on for 12 days. With LalBrew Pomona, or two packs plus pure oxygen, ~90% of the sugar is gone within &#039;&#039;&#039;2 days&#039;&#039;&#039;. Any schedule written as &amp;quot;dry hop on Day 3&amp;quot; is simply wrong for these yeasts.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_plan_vs_actual.png|800px|center|Planned vs actual timing of key events]]&lt;br /&gt;
Since #0009 each batch starts with a one-line written plan (dry-hop day and gravity, green-light day, transfer day, expected FG). The log then measures the error. The first two plans were badly off — fermentation finished 8–12 days early — and each new plan is corrected from the previous error. #0011&#039;s DH1, DH2 and FG all landed within a day of plan (1.0, 0.8 and 0.8 days early).&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_ph.png|900px|center|Mash pH and finished-beer pH]]&lt;br /&gt;
pH turned out to be the hidden cause of &amp;quot;too bitter&amp;quot; — and, in #0011, of &amp;quot;slightly sour&amp;quot;. The same tap water gave a mash pH of 4.76 in one batch and 5.81 in another, so the correction has to go in either direction. #0011&#039;s mash was corrected to 5.37, but the wort measured 4.83 after the boil and was not corrected again; the beer finished at 3.94. In both #0008 (4.75 → 3.83) and #0011 (4.83 → 3.94) pH fell about &#039;&#039;&#039;0.9&#039;&#039;&#039; during fermentation, not the textbook 0.3–0.5 — this soft, low-alkalinity water has almost no buffer.&lt;br /&gt;
&lt;br /&gt;
== Fermentation with the RAPT Pill (#0012) ==&lt;br /&gt;
#0012 (Orange Glow v2.6, brewed 2026-11-10, LalBrew Pomona, OG 1.068) is the first batch logged by the [[#Equipment|RAPT Pill]]: a gravity and temperature reading every 15 minutes instead of a hydrometer sample every 12 h. The three charts below come from that batch and show how the fermentation was steered from the data.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_gravity_vs_days.png|900px|center|#0012: gravity vs days since pitch, with the expected DH1 window and the actual events]]&lt;br /&gt;
The plan is written in gravity, not days: DH1 goes in when gravity is half-way from OG to FG (1.039–1.029 for OG 1.068), and the diacetyl rest starts at ≤ 1.020. Krausen took ~13 h to appear (the wort was pitched at 22.8 °C and then cooled to 17–18 °C), but once going this yeast was faster than #0009 at the same stage. DH1 (56 g Nectaron + Citra, bagged) went in at 1.041 on Day 1.2 — deliberately ~2 points early, because dry-hopping early costs nothing in biotransformation and dry-hopping late is how #0010 missed its window. The bath went up to 21.5 °C for the diacetyl rest at 1.022 on Day 1.5, without waiting for the last 2 points.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_pill_telemetry.png|900px|center|#0012: RAPT Pill raw gravity, offset-corrected gravity, beer temperature and hydrometer checks]]&lt;br /&gt;
* &#039;&#039;&#039;The Pill reads high by a fixed offset.&#039;&#039;&#039; Two degassed hydrometer checks (at pitch and on Day 1) both put it ~13.5 points above the true gravity, so it is a constant offset, not a scaling error. The dashed line is Pill − 13.5. Recalibration is deliberately left until transfer: recalibrating in the middle of fermentation would put a step in the curve, and nobody reading it later could tell that step from fermentation. Alarm thresholds are converted to Pill units instead (DH1 window 1.039–1.029 → Pill 1.053–1.043).&lt;br /&gt;
* &#039;&#039;&#039;Ignore the Pill while krausen sits on it.&#039;&#039;&#039; Around hour 13 the raw reading &#039;&#039;rose&#039;&#039; from 1.0814 to 1.0869. Gravity cannot go up, and warming would make a tilt hydrometer read lower, not higher, so the cause is CO₂ bubbles and krausen pushing the Pill more upright. The bump marks the start of fermentation; it is not data.&lt;br /&gt;
* &#039;&#039;&#039;Degas every hydrometer sample.&#039;&#039;&#039; The Day 1 sample read 1.056 straight from the tap and 1.051 after degassing and temperature correction. What matters is bubbles stuck on the hydrometer, not whether the sample looks fizzy.&lt;br /&gt;
* &#039;&#039;&#039;Take beer temperature from the Pill, not from the sample.&#039;&#039;&#039; A sample warms up in the room within minutes. After the bath was set to 21.5 °C the beer climbed only ~0.3 °C/h, so the 3-day diacetyl rest is counted from when the &#039;&#039;beer&#039;&#039; reaches 21.5 °C, not from when the setpoint was changed.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_pill_rate.png|900px|center|#0012: gravity drop rate in points per day, with the 2 points/day green-light threshold]]&lt;br /&gt;
The rate is the change in Pill gravity smoothed over a ±1.5 h centred window (differences between neighbouring 15-minute readings are almost pure noise), in points per day — the same unit as RAPT&#039;s Gravity Velocity alert. It rose from ~31 points/day at hour 18 to ~73 around the start of the diacetyl rest, and has been falling since. When the line stays under 2 points/day, the gravity has stopped changing. That is one of the two green-light conditions; the other is 3 days at 20–22 °C.&lt;br /&gt;
&lt;br /&gt;
As of 2026-11-12 the raw Pill reading has levelled off at ~1.025, which is ~1.012 true — inside the 1.011–1.014 FG target, for about 7.4% ABV. Next: pull the DH1 bag and take a degassed hydrometer sample on 11/14 (to confirm FG and check whether the offset has drifted), then the green-light check and DH2 (84 g, loose) on 11/15.&lt;br /&gt;
&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_0012_keg_glass.jpg|#0012 poured from the mini keg, 2026-11-20&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== Batch log ==&lt;br /&gt;
{| class=&amp;quot;wikitable sortable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! # !! Date !! Beer / version !! Yeast !! OG → FG !! ABV !! Result / key lesson&lt;br /&gt;
|-&lt;br /&gt;
| 0001 || 2025-12-03 || DDH Hazy IPA || LalBrew New England || 1.062 → 1.018 || 5.8% || Bottle-conditioned. Lesson: add priming sugar, don&#039;t refrigerate during bottle conditioning.&lt;br /&gt;
|-&lt;br /&gt;
| 0002 || 2025-12-28 || Hazy IPA || Mauribrew Weiss || 1.040 → ? || – || Poor mash efficiency; dry-hop order wrong; forgot to measure FG before bottling.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-1 || 2026-02-26 || IPA || Mauribrew ALE514 || 1.050 → 1.010 || 5.3% || Flat aroma and taste; too much priming sugar, vegetal smell.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-2 || 2026-02-28 || IPA + 900 g oats || ALE514 || 1.056 → 1.016 || 5.3% || Under-pitched; bottled while airlock still active → gushers on opening.&lt;br /&gt;
|-&lt;br /&gt;
| 0004 || 2026-03-15 || Milkshake IPA || LalBrew New England || 1.110 → 1.030 || 10.5% || Lid not sealed (airlock silent 36 h); bottled while still fermenting; weak carbonation, slightly sour.&lt;br /&gt;
|-&lt;br /&gt;
| 0005 || 2026-04-23 || Milkshake IPA &amp;quot;light&amp;quot; || Verdant IPA || 1.102 → 1.030 || ~9.5% || &#039;&#039;&#039;Failed.&#039;&#039;&#039; Bittering hops 22 g instead of 6 g; bottles never carbonated — at ~9.5% ABV the yeast was too stressed to referment, and the cold crash removed what was left. Most poured away.&lt;br /&gt;
|-&lt;br /&gt;
| 0006 || 2026-05-09 || Hazy IPA || Verdant IPA || 1.062 → 1.008 || 7.1% || Second dry hop at Day 9 triggered &#039;&#039;&#039;hop creep&#039;&#039;&#039; (1.014 → 1.008). Clean, bitterness fine, aroma weak. First keg + forced carbonation.&lt;br /&gt;
|-&lt;br /&gt;
| 0007 || 2026-06-12 || Orange Glow v2.3 || Verdant IPA ||1.080 → 1.015 || 8.5% || First versioned recipe. Not hot, bitterness right after 3 weeks, but aroma lacking and colour unappealing.&lt;br /&gt;
|-&lt;br /&gt;
| 0008 || 2026-06-25 || Orange Glow v2.4 || LalBrew Pomona || 1.090 → 1.016 || 9.7% || Pomona hit 1.016 by Day 3 — dry hop missed its window. Finished pH 3.83 → harsh bitterness drowned the juice.&lt;br /&gt;
|-&lt;br /&gt;
| 0009 || 2026-07-29 || Orange Glow v2.5 &amp;quot;clean&amp;quot; || Pomona || 1.050 → 1.006 || 5.8% || Boiled off 4–5 L, OG shot to 1.110, over-diluted back to 1.050. pH 4.30 — bitterness finally right. Thin body, under-carbonated (keg fridge is 7–8 °C, not 3 °C). Passable daily hazy.&lt;br /&gt;
|-&lt;br /&gt;
| 0010 || 2026-08-26 || Strange Haze v1 (blueberry vanilla) || Verdant ×2 || 1.070 → 1.012 || 7.6% || Pre-boil OG check caught an over-concentrated wort in time. No harshness, no diacetyl — but not enough juice, sweetness, purple colour or vanilla; grainy from 20% oats.&lt;br /&gt;
|-&lt;br /&gt;
| 0011 || 2026-10-08 || 妃子起笑 v4.3 (PB Craft kit) || Weiss || 1.052 → 1.020 || 4.2% || Clean process — no diacetyl, oxidation or harsh bitterness, no bottle bombs — but weak aroma, thin &amp;quot;watery&amp;quot; body, pale (3–4 SRM) and slightly sour (pH 3.94). Every flaw traced to three fixable numbers: post-boil pH not corrected, CaCl₂ only 0.5 g, both dry hops added while still fermenting. Led to v4.4 and to recipe-wide changes.&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The beers ==&lt;br /&gt;
; 妃子起笑 · Crazy Concubine — Hazy Pale Ale (OG 1.052, ~5%)&lt;br /&gt;
: The first recipe. v1&#039;s second dry hop went in after fermentation had finished and hop creep ate the body (FG 1.014 → 1.008, 5.8% instead of 5%). v3 fixed it with zero new cost — one dry hop during active fermentation, cooler mash, pure-oxygen pitch, no cold crash. v4 added a 78 °C whirlpool for front-end juiciness. #0011 brewed the PB Craft kit exactly as printed (v4.3) and hit the kit&#039;s numbers, which showed the kit itself is the limit. v4.4: 10 L instead of 11.5 L (OG back to 1.060), CaCl₂ 3 g, a post-boil pH check, and the dry hops swapped — Idaho Gem first, Citra + Amarillo last. #0002, the same kit, had used that &amp;quot;wrong&amp;quot; order by mistake — and had the lychee aroma #0011 lacked.&lt;br /&gt;
; Orange Glow — DDH NEIPA (OG ~1.066, ~7%)&lt;br /&gt;
: The main development line: v1 → v2.6 across three brews. Swapped glucose for DME, Extra Pale for Golden Promise, Verdant for Pomona, added ALDC and Fermaid-O, moved to a split dry hop. #0009 drifted to a pale 5.8%; v2.6 brings back Munich for the orange-gold colour and uses a Strata / Mosaic / Nelson / Peacharine hop bill.&lt;br /&gt;
; Strange Haze — Imperial Milkshake IPA (blueberry vanilla)&lt;br /&gt;
: Built around a pre-printed label (9% ABV, purple nebula), so the fruit is blueberry for colour. v1 was &amp;quot;neither here nor there&amp;quot;; v1.1 cuts bitterness, adds lactose back, raises mash to 69–70 °C and splits 800 g of blueberry between mid-fermentation and post-fermentation.&lt;br /&gt;
; Milkshake IPA — low residual sugar (6 L)&lt;br /&gt;
: Body from oat/wheat protein instead of lots of lactose; unsweetened mango + passion fruit purée. Not yet brewed under this recipe.&lt;br /&gt;
&lt;br /&gt;
== Label and logo ==&lt;br /&gt;
The beers go out under the name &#039;&#039;&#039;piz brewing&#039;&#039;&#039;. The logo started from our dog, became a pencil sketch of just the ears and nose, and was finished with AI image generation. The labels were made the same way: many drafts generated and refined until one fit the beer.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_logo_dog.jpg|The original&lt;br /&gt;
File:Homebrew_logo_sketch.jpg|Pencil sketch: ears and nose&lt;br /&gt;
File:Homebrew_logo_final.jpg|Final logo&lt;br /&gt;
File:Homebrew_label_strange_haze.jpg|Strange Haze label, made with AI image generation&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== What I learned ==&lt;br /&gt;
=== Fermentation ===&lt;br /&gt;
* &#039;&#039;&#039;Dry hop by gravity, not by the calendar.&#039;&#039;&#039; The first dry hop goes in when gravity has dropped about half-way from OG to FG, while yeast is still active (biotransformation + it mops up hop-creep sugar). With fast yeast that can be Day 1, so start measuring every 12 h from Day 1.&lt;br /&gt;
* &#039;&#039;&#039;Split dry hop.&#039;&#039;&#039; DH1 during active fermentation for biotransformation; DH2 only after fermentation has stopped, so the CO₂ doesn&#039;t scrub the aroma away. After DH2, confirm gravity is stable again (second wave of hop creep). #0011 put both dry hops in while it was still fermenting, and the empty glass smelled only of malt.&lt;br /&gt;
* &#039;&#039;&#039;The last dry hop is the aroma you smell.&#039;&#039;&#039; It soaks latest, is scrubbed least and has the least time to oxidise, so put the signature hop (e.g. Citra) there, loose rather than in a bag. Only bag the dry hop you plan to pull out.&lt;br /&gt;
* &#039;&#039;&#039;Hop creep is real.&#039;&#039;&#039; Hop enzymes break down dextrins, so a late dry hop restarts fermentation: thinner body, higher ABV, and in bottles — gushers or bombs.&lt;br /&gt;
* &#039;&#039;&#039;&amp;quot;Green light&amp;quot; before packaging:&#039;&#039;&#039; two gravity readings ~48 h apart that differ by ≤ 0.001, &#039;&#039;&#039;and&#039;&#039;&#039; at least 3 days of diacetyl rest at a beer temperature of 20–22 °C (water bath 21.5 °C). Since 2026-09 the rest time replaces the forced diacetyl test: below 18 °C the yeast barely cleans up diacetyl, above 24 °C it only adds fusel alcohols, so 3 days at 20–22 °C is the guarantee. After DH2 or fruit, hop creep can make new diacetyl, so hold 20–22 °C for another 48 h and re-check gravity before cold-crashing. #0011 was packaged before creep had stopped and still had no butter flavour. If a beer does taste buttery, warm the keg back to room temperature for 3 days. Never package on a fixed day count — #0003-2, #0004 and #0008 all did, and all had carbonation problems.&lt;br /&gt;
* &#039;&#039;&#039;Pure oxygen 60–90 s before pitching&#039;&#039;&#039; lets a single dry pack ferment a 1.07+ wort; no starter needed. After pitching, oxygen is the enemy.&lt;br /&gt;
* Fermenting beer runs 1–2 °C above the water bath during the first 48 h. Set the bath ~1.5 °C below the target beer temperature at peak, then raise it to 21.5 °C for the diacetyl rest.&lt;br /&gt;
&lt;br /&gt;
=== Brew day ===&lt;br /&gt;
* &#039;&#039;&#039;Check pre-boil gravity and volume before boiling.&#039;&#039;&#039; Weigh the pot (total − empty pot weight ≈ litres), measure pre-boil gravity, and compute the target end volume = pre-boil points × litres ÷ target points. Boil to that volume, not to a timer. #0009 boiled off 4–5 L (normal is 1–2 L) and hit 1.110.&lt;br /&gt;
* &#039;&#039;&#039;Boil gently, lid off.&#039;&#039;&#039; Hard boiling over-concentrates bitterness and scorches. Excess evaporation turned out to be a big part of the &amp;quot;nasty bitterness&amp;quot; in #0007/#0008.&lt;br /&gt;
* &#039;&#039;&#039;Measure mash pH, not water pH, and be ready to go both ways.&#039;&#039;&#039; Sample 10–15 min after mash-in, cool to 20–25 °C, target 5.2–5.4. Keep both baking soda and phosphoric/lactic acid on hand. This soft, low-alkalinity water swings hard: ~1 g baking soda ≈ +0.67, 1 ml phosphoric ≈ −0.28 — use half steps and re-measure. Calibrate the pH pen every brew day and store the electrode wet.&lt;br /&gt;
* &#039;&#039;&#039;Post-boil pH is a second, separate checkpoint.&#039;&#039;&#039; A long boil concentrates the acidity too: #0011 boiled 14.55 L down to 11.5 L and dropped from 5.37 to 4.83. Correct the cooled wort to 5.2–5.4 before pitching; with a ~0.9 drop in fermentation that lands the beer at ~4.4. (#0010 split the same 2 g of baking soda between mash and post-boil and had no sourness.)&lt;br /&gt;
* &#039;&#039;&#039;Salt: CaCl₂·2H₂O 0.2 g per litre of total brewing water&#039;&#039;&#039; (≈ Ca 54 / Cl 96 ppm), no gypsum. The filtered water has almost no minerals; #0011&#039;s 0.5 g gave ~17 ppm chloride and ~9 ppm calcium, and the beer tasted like &amp;quot;sweet water&amp;quot;.&lt;br /&gt;
* Pale-malt grists (Extra Pale/Pilsner) tend to mash high (~5.7); grists with Munich plus gypsum tend to mash low.&lt;br /&gt;
&lt;br /&gt;
=== Bitterness and aroma ===&lt;br /&gt;
* &#039;&#039;&#039;Fix bitterness at the source, never by cutting hops.&#039;&#039;&#039; The culprits were pH (finished pH 3.83 = throat-biting), gypsum (sulfate sharpens bitterness), over-boiling, hops left in too long, and drinking too young. Bittering hops go in a bag, on time, and come out after the boil; dry-hop bags come out after 2–3 days.&lt;br /&gt;
* &#039;&#039;&#039;Whirlpool and dry hop do different jobs.&#039;&#039;&#039; Whirlpool (75–78 °C) gives juice on the palate; dry hop gives the aroma in the nose. Cutting either one loses something the other can&#039;t replace.&lt;br /&gt;
* Aging softens bitterness but kills hop aroma at the same time — fix bitterness at the source and drink fresh (2–3 weeks) instead.&lt;br /&gt;
* 0.5 kg oats + 1 kg wheat in 4.5 kg of grain is already enough haze; 1 kg of oats (20%) tasted grainy.&lt;br /&gt;
* &#039;&#039;&#039;Lactose: go big or skip it.&#039;&#039;&#039; 50 g in 11.5 L (4.3 g/L) is far below the ~20–25 g/L you can taste — it added nothing.&lt;br /&gt;
* &#039;&#039;&#039;A kit guarantees beer, not good beer.&#039;&#039;&#039; The PB Craft kit has no whirlpool hops, no specialty malt, and says nothing about salt or pH. Brewed exactly as printed (#0011), it gives a pale, thin, low-aroma beer; those four things have to be added.&lt;br /&gt;
&lt;br /&gt;
=== Packaging ===&lt;br /&gt;
* &#039;&#039;&#039;Keg + forced carbonation&#039;&#039;&#039; is now the default (#0009 on). Bottle conditioning kept failing: too much sugar (#0003-1), bottling before fermentation finished (#0003-2, #0004), and #0005 at ~9.5% ABV, where the yeast could not referment.&lt;br /&gt;
* &#039;&#039;&#039;Cold crash is now standard&#039;&#039;&#039; (since #0011): 2–4 °C for 24–48 h before transfer. #0006–#0011 never really cold-crashed, and all six got the same notes — weak aroma, &amp;quot;don&#039;t know what I&#039;m drinking&amp;quot;, grainy. Suspended yeast fits that picture; it is still a hypothesis, to be tested by drinking #0011&#039;s keg (no yeast) next to its bottles (live yeast). For bottle conditioning after a cold crash: under 7% ABV enough yeast remains; above 8%, add 1–2 g of an alcohol-tolerant yeast (e.g. EC-1118) at bottling.&lt;br /&gt;
* &#039;&#039;&#039;Don&#039;t package on a falling gravity with sulfite in the keg.&#039;&#039;&#039; #0011 was packaged at 1.020 while still dropping: no metabisulfite for an early package (keep the yeast alive), and halve the priming sugar for bottles (4.5 g/L — it carbonated without gushing).&lt;br /&gt;
* &#039;&#039;&#039;Never bottle the gravity samples.&#039;&#039;&#039; #0011&#039;s test bottle made from saved samples gushed and measured pH 3.5 — lactic infection from the open sample cup, not the fermenter (the normal bottles were 3.94). Drink samples on the spot; for a control bottle, fill one more from the main beer.&lt;br /&gt;
* Low-oxygen transfer: purge the keg with CO₂ three times, fill from the bottom through the liquid post, minimum headspace.&lt;br /&gt;
* Antioxidant in the keg: &#039;&#039;&#039;ascorbic acid 0.3 g + potassium metabisulfite 0.25 g per 10 L, always as a pair.&#039;&#039;&#039; Ascorbic acid alone makes H₂O₂, which with trace iron/copper (Fenton reaction) oxidises the beer worse than nothing. The chemicals are ~20% of the protection; the purging is the other 80%.&lt;br /&gt;
* &#039;&#039;&#039;Measure the fridge.&#039;&#039;&#039; The keg fridge is actually 7–8 °C, so 2.4–2.5 volumes needs ~15 PSI, not the textbook 11–12 PSI. #0009 and #0010 were both under-carbonated for this reason. A small 3 L keg also absorbs CO₂ slowly — #0011 was only ~70–80% carbonated after 5 days at pressure; lay it down and shake for 2–3 minutes.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_mini_keg_regulator.jpg|Mini keg with CO₂ regulator: carbonation is set by pressure, not sugar&lt;br /&gt;
File:Homebrew_mini_keg_tap.jpg|Mini keg with a wooden tap handle&lt;br /&gt;
File:Homebrew_bottles_box.jpg|A case of bottles filled from the keg&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
=== Method ===&lt;br /&gt;
* Write the plan down before brewing (one line: dry-hop day and gravity, green light, transfer, FG) and measure the error afterwards. The predictions are converging batch by batch.&lt;br /&gt;
* Log everything with a timestamp — gravity with date and time, pH with temperature. Most of the lessons above only became visible once the numbers were next to each other.&lt;br /&gt;
* Change one or two things per version. The recipe history (Crazy Concubine v1→v4.4, Orange Glow v1→v2.6) reads like a commit log: each version cites the batch that justified it.&lt;br /&gt;
&lt;br /&gt;
== Equipment ==&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_first_kit.jpg|Brew day #0001 (Dec 2025): kettle, fermenter, grain and hops&lt;br /&gt;
File:Homebrew_hop_pellets.jpg|Hop pellets, straight from the vacuum pack&lt;br /&gt;
File:Homebrew_mash.jpg|Mashing #0008, thermometer in the lid&lt;br /&gt;
File:Homebrew_ir_thermometer.jpg|IR thermometer check on the kettle&lt;br /&gt;
File:Homebrew_cooling_rig.jpg|Home-built cooling rig: thermoelectric cooler, power supply and thermostat, pumping cold water through the fermenter&lt;br /&gt;
File:Homebrew_o2_co2_cylinders.jpg|CO₂ (left) and pure O₂ (right) cylinders with regulators&lt;br /&gt;
File:Homebrew_fermenter_krausen.jpg|#0008 on Day 3: krausen, fermenting in the water bath&lt;br /&gt;
File:Homebrew_hydrometer.jpg|Gravity sample from #0009&lt;br /&gt;
File:Homebrew_rapt_pill.jpg|RAPT Pill wireless hydrometer (Nov 2026), next to its box&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
* Kettle + fermenter in a temperature-controlled water bath; weigh-the-pot volume estimation&lt;br /&gt;
* Pure O₂ cylinder with diffusion stone; refractometer + hydrometer; pH pen (4.01 / 7.00 buffers)&lt;br /&gt;
* Inkbird IHT-1P instant-read thermometer (from #0010)&lt;br /&gt;
* RAPT Pill wireless hydrometer and thermometer (from Nov 2026): floats in the fermenter and logs gravity and temperature over Bluetooth LE or 2.4 GHz Wi-Fi; range 0.95–1.12, rechargeable 18650 battery (USB-C)&lt;br /&gt;
* 8 L keg on external CO₂; 128 oz mini keg for in-keg conditioning&lt;br /&gt;
* Additives: Fermaid-O (2.5 g at pitch + 2.5 g at ~24 h), Maturex L ALDC 0.2–0.3 ml, CaCl₂&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=861</id>
		<title>Homebrewing</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=861"/>
		<updated>2026-11-20T15:37:13Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Top gallery: #0012 is the best batch so far; fix glass caption&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&#039;&#039;&#039;Homebrewing&#039;&#039;&#039; — eleven batches of hazy / New England IPA brewed at home in 10-litre batches, December 2025 to October 2026. Every batch is logged with gravity readings, pH, hop timing and a tasting note, and each recipe is version-controlled in git so every change can be traced back to the batch that caused it. This page summarises what was brewed, what went wrong, and the process rules that came out of it.&lt;br /&gt;
&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_0012_glass.jpg|Orange Glow v2.6 (#0012), the best batch so far&lt;br /&gt;
File:Homebrew_glass_orange_glow.jpg|Orange Glow v2.3 (#0007), five weeks after brew day&lt;br /&gt;
File:Homebrew_glass_in_hand.jpg|A finished hazy IPA, in the glass&lt;br /&gt;
File:Homebrew_keg_and_bottles.jpg|Keg and bottles, with the Strange Haze label in front&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Batches !! Period !! Batch size !! Styles !! Recipes (current version)&lt;br /&gt;
|-&lt;br /&gt;
| 11 (12 brews) || 2025-12 → 2026-10 || 6–11.5 L || Hazy IPA, DDH NEIPA, Milkshake IPA || 妃子起笑 Crazy Concubine (v4.4), Orange Glow (v2.6), Strange Haze (v1.1), Milkshake IPA (v1)&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The data ==&lt;br /&gt;
[[File:Homebrew_og_fg_by_batch.png|900px|center|Original and final gravity of every batch]]&lt;br /&gt;
The first six batches had no written recipe and wild original gravities (up to 1.110 — a 10.5% &amp;quot;IPA&amp;quot;). From #0007 on, every batch follows a versioned recipe and the OG has been pulled back into the 1.050–1.076 range the recipes are designed for.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_abv_attenuation.png|900px|center|ABV and apparent attenuation per batch]]&lt;br /&gt;
Attenuation went from ~71% (early batches, under-pitched, no oxygen) to a consistent 81–88% once pure-oxygen pitching, yeast nutrient and modern yeast strains were adopted. #0011&#039;s 62% is not a stuck fermentation: it is a kit recipe with 50 g of unfermentable lactose, and it landed exactly on the kit&#039;s printed OG 1.052 / FG 1.020.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_gravity_curves.png|900px|center|Fermentation curves of four batches]]&lt;br /&gt;
The single most useful chart. With one pack of Verdant and no oxygen (#0006) fermentation drags on for 12 days. With LalBrew Pomona, or two packs plus pure oxygen, ~90% of the sugar is gone within &#039;&#039;&#039;2 days&#039;&#039;&#039;. Any schedule written as &amp;quot;dry hop on Day 3&amp;quot; is simply wrong for these yeasts.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_plan_vs_actual.png|800px|center|Planned vs actual timing of key events]]&lt;br /&gt;
Since #0009 each batch starts with a one-line written plan (dry-hop day and gravity, green-light day, transfer day, expected FG). The log then measures the error. The first two plans were badly off — fermentation finished 8–12 days early — and each new plan is corrected from the previous error. #0011&#039;s DH1, DH2 and FG all landed within a day of plan (1.0, 0.8 and 0.8 days early).&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_ph.png|900px|center|Mash pH and finished-beer pH]]&lt;br /&gt;
pH turned out to be the hidden cause of &amp;quot;too bitter&amp;quot; — and, in #0011, of &amp;quot;slightly sour&amp;quot;. The same tap water gave a mash pH of 4.76 in one batch and 5.81 in another, so the correction has to go in either direction. #0011&#039;s mash was corrected to 5.37, but the wort measured 4.83 after the boil and was not corrected again; the beer finished at 3.94. In both #0008 (4.75 → 3.83) and #0011 (4.83 → 3.94) pH fell about &#039;&#039;&#039;0.9&#039;&#039;&#039; during fermentation, not the textbook 0.3–0.5 — this soft, low-alkalinity water has almost no buffer.&lt;br /&gt;
&lt;br /&gt;
== Fermentation with the RAPT Pill (#0012) ==&lt;br /&gt;
#0012 (Orange Glow v2.6, brewed 2026-11-10, LalBrew Pomona, OG 1.068) is the first batch logged by the [[#Equipment|RAPT Pill]]: a gravity and temperature reading every 15 minutes instead of a hydrometer sample every 12 h. The three charts below come from that batch and show how the fermentation was steered from the data.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_gravity_vs_days.png|900px|center|#0012: gravity vs days since pitch, with the expected DH1 window and the actual events]]&lt;br /&gt;
The plan is written in gravity, not days: DH1 goes in when gravity is half-way from OG to FG (1.039–1.029 for OG 1.068), and the diacetyl rest starts at ≤ 1.020. Krausen took ~13 h to appear (the wort was pitched at 22.8 °C and then cooled to 17–18 °C), but once going this yeast was faster than #0009 at the same stage. DH1 (56 g Nectaron + Citra, bagged) went in at 1.041 on Day 1.2 — deliberately ~2 points early, because dry-hopping early costs nothing in biotransformation and dry-hopping late is how #0010 missed its window. The bath went up to 21.5 °C for the diacetyl rest at 1.022 on Day 1.5, without waiting for the last 2 points.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_pill_telemetry.png|900px|center|#0012: RAPT Pill raw gravity, offset-corrected gravity, beer temperature and hydrometer checks]]&lt;br /&gt;
* &#039;&#039;&#039;The Pill reads high by a fixed offset.&#039;&#039;&#039; Two degassed hydrometer checks (at pitch and on Day 1) both put it ~13.5 points above the true gravity, so it is a constant offset, not a scaling error. The dashed line is Pill − 13.5. Recalibration is deliberately left until transfer: recalibrating in the middle of fermentation would put a step in the curve, and nobody reading it later could tell that step from fermentation. Alarm thresholds are converted to Pill units instead (DH1 window 1.039–1.029 → Pill 1.053–1.043).&lt;br /&gt;
* &#039;&#039;&#039;Ignore the Pill while krausen sits on it.&#039;&#039;&#039; Around hour 13 the raw reading &#039;&#039;rose&#039;&#039; from 1.0814 to 1.0869. Gravity cannot go up, and warming would make a tilt hydrometer read lower, not higher, so the cause is CO₂ bubbles and krausen pushing the Pill more upright. The bump marks the start of fermentation; it is not data.&lt;br /&gt;
* &#039;&#039;&#039;Degas every hydrometer sample.&#039;&#039;&#039; The Day 1 sample read 1.056 straight from the tap and 1.051 after degassing and temperature correction. What matters is bubbles stuck on the hydrometer, not whether the sample looks fizzy.&lt;br /&gt;
* &#039;&#039;&#039;Take beer temperature from the Pill, not from the sample.&#039;&#039;&#039; A sample warms up in the room within minutes. After the bath was set to 21.5 °C the beer climbed only ~0.3 °C/h, so the 3-day diacetyl rest is counted from when the &#039;&#039;beer&#039;&#039; reaches 21.5 °C, not from when the setpoint was changed.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_pill_rate.png|900px|center|#0012: gravity drop rate in points per day, with the 2 points/day green-light threshold]]&lt;br /&gt;
The rate is the change in Pill gravity smoothed over a ±1.5 h centred window (differences between neighbouring 15-minute readings are almost pure noise), in points per day — the same unit as RAPT&#039;s Gravity Velocity alert. It rose from ~31 points/day at hour 18 to ~73 around the start of the diacetyl rest, and has been falling since. When the line stays under 2 points/day, the gravity has stopped changing. That is one of the two green-light conditions; the other is 3 days at 20–22 °C.&lt;br /&gt;
&lt;br /&gt;
As of 2026-11-12 the raw Pill reading has levelled off at ~1.025, which is ~1.012 true — inside the 1.011–1.014 FG target, for about 7.4% ABV. Next: pull the DH1 bag and take a degassed hydrometer sample on 11/14 (to confirm FG and check whether the offset has drifted), then the green-light check and DH2 (84 g, loose) on 11/15.&lt;br /&gt;
&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_0012_glass.jpg|#0012 Orange Glow v2.6 in the glass, 2026-11-20&lt;br /&gt;
File:Homebrew_0012_keg_glass.jpg|#0012 poured from the mini keg, 2026-11-20&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== Batch log ==&lt;br /&gt;
{| class=&amp;quot;wikitable sortable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! # !! Date !! Beer / version !! Yeast !! OG → FG !! ABV !! Result / key lesson&lt;br /&gt;
|-&lt;br /&gt;
| 0001 || 2025-12-03 || DDH Hazy IPA || LalBrew New England || 1.062 → 1.018 || 5.8% || Bottle-conditioned. Lesson: add priming sugar, don&#039;t refrigerate during bottle conditioning.&lt;br /&gt;
|-&lt;br /&gt;
| 0002 || 2025-12-28 || Hazy IPA || Mauribrew Weiss || 1.040 → ? || – || Poor mash efficiency; dry-hop order wrong; forgot to measure FG before bottling.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-1 || 2026-02-26 || IPA || Mauribrew ALE514 || 1.050 → 1.010 || 5.3% || Flat aroma and taste; too much priming sugar, vegetal smell.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-2 || 2026-02-28 || IPA + 900 g oats || ALE514 || 1.056 → 1.016 || 5.3% || Under-pitched; bottled while airlock still active → gushers on opening.&lt;br /&gt;
|-&lt;br /&gt;
| 0004 || 2026-03-15 || Milkshake IPA || LalBrew New England || 1.110 → 1.030 || 10.5% || Lid not sealed (airlock silent 36 h); bottled while still fermenting; weak carbonation, slightly sour.&lt;br /&gt;
|-&lt;br /&gt;
| 0005 || 2026-04-23 || Milkshake IPA &amp;quot;light&amp;quot; || Verdant IPA || 1.102 → 1.030 || ~9.5% || &#039;&#039;&#039;Failed.&#039;&#039;&#039; Bittering hops 22 g instead of 6 g; bottles never carbonated — at ~9.5% ABV the yeast was too stressed to referment, and the cold crash removed what was left. Most poured away.&lt;br /&gt;
|-&lt;br /&gt;
| 0006 || 2026-05-09 || Hazy IPA || Verdant IPA || 1.062 → 1.008 || 7.1% || Second dry hop at Day 9 triggered &#039;&#039;&#039;hop creep&#039;&#039;&#039; (1.014 → 1.008). Clean, bitterness fine, aroma weak. First keg + forced carbonation.&lt;br /&gt;
|-&lt;br /&gt;
| 0007 || 2026-06-12 || Orange Glow v2.3 || Verdant IPA ||1.080 → 1.015 || 8.5% || First versioned recipe. Not hot, bitterness right after 3 weeks, but aroma lacking and colour unappealing.&lt;br /&gt;
|-&lt;br /&gt;
| 0008 || 2026-06-25 || Orange Glow v2.4 || LalBrew Pomona || 1.090 → 1.016 || 9.7% || Pomona hit 1.016 by Day 3 — dry hop missed its window. Finished pH 3.83 → harsh bitterness drowned the juice.&lt;br /&gt;
|-&lt;br /&gt;
| 0009 || 2026-07-29 || Orange Glow v2.5 &amp;quot;clean&amp;quot; || Pomona || 1.050 → 1.006 || 5.8% || Boiled off 4–5 L, OG shot to 1.110, over-diluted back to 1.050. pH 4.30 — bitterness finally right. Thin body, under-carbonated (keg fridge is 7–8 °C, not 3 °C). Passable daily hazy.&lt;br /&gt;
|-&lt;br /&gt;
| 0010 || 2026-08-26 || Strange Haze v1 (blueberry vanilla) || Verdant ×2 || 1.070 → 1.012 || 7.6% || Pre-boil OG check caught an over-concentrated wort in time. No harshness, no diacetyl — but not enough juice, sweetness, purple colour or vanilla; grainy from 20% oats.&lt;br /&gt;
|-&lt;br /&gt;
| 0011 || 2026-10-08 || 妃子起笑 v4.3 (PB Craft kit) || Weiss || 1.052 → 1.020 || 4.2% || Clean process — no diacetyl, oxidation or harsh bitterness, no bottle bombs — but weak aroma, thin &amp;quot;watery&amp;quot; body, pale (3–4 SRM) and slightly sour (pH 3.94). Every flaw traced to three fixable numbers: post-boil pH not corrected, CaCl₂ only 0.5 g, both dry hops added while still fermenting. Led to v4.4 and to recipe-wide changes.&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The beers ==&lt;br /&gt;
; 妃子起笑 · Crazy Concubine — Hazy Pale Ale (OG 1.052, ~5%)&lt;br /&gt;
: The first recipe. v1&#039;s second dry hop went in after fermentation had finished and hop creep ate the body (FG 1.014 → 1.008, 5.8% instead of 5%). v3 fixed it with zero new cost — one dry hop during active fermentation, cooler mash, pure-oxygen pitch, no cold crash. v4 added a 78 °C whirlpool for front-end juiciness. #0011 brewed the PB Craft kit exactly as printed (v4.3) and hit the kit&#039;s numbers, which showed the kit itself is the limit. v4.4: 10 L instead of 11.5 L (OG back to 1.060), CaCl₂ 3 g, a post-boil pH check, and the dry hops swapped — Idaho Gem first, Citra + Amarillo last. #0002, the same kit, had used that &amp;quot;wrong&amp;quot; order by mistake — and had the lychee aroma #0011 lacked.&lt;br /&gt;
; Orange Glow — DDH NEIPA (OG ~1.066, ~7%)&lt;br /&gt;
: The main development line: v1 → v2.6 across three brews. Swapped glucose for DME, Extra Pale for Golden Promise, Verdant for Pomona, added ALDC and Fermaid-O, moved to a split dry hop. #0009 drifted to a pale 5.8%; v2.6 brings back Munich for the orange-gold colour and uses a Strata / Mosaic / Nelson / Peacharine hop bill.&lt;br /&gt;
; Strange Haze — Imperial Milkshake IPA (blueberry vanilla)&lt;br /&gt;
: Built around a pre-printed label (9% ABV, purple nebula), so the fruit is blueberry for colour. v1 was &amp;quot;neither here nor there&amp;quot;; v1.1 cuts bitterness, adds lactose back, raises mash to 69–70 °C and splits 800 g of blueberry between mid-fermentation and post-fermentation.&lt;br /&gt;
; Milkshake IPA — low residual sugar (6 L)&lt;br /&gt;
: Body from oat/wheat protein instead of lots of lactose; unsweetened mango + passion fruit purée. Not yet brewed under this recipe.&lt;br /&gt;
&lt;br /&gt;
== Label and logo ==&lt;br /&gt;
The beers go out under the name &#039;&#039;&#039;piz brewing&#039;&#039;&#039;. The logo started from our dog, became a pencil sketch of just the ears and nose, and was finished with AI image generation. The labels were made the same way: many drafts generated and refined until one fit the beer.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_logo_dog.jpg|The original&lt;br /&gt;
File:Homebrew_logo_sketch.jpg|Pencil sketch: ears and nose&lt;br /&gt;
File:Homebrew_logo_final.jpg|Final logo&lt;br /&gt;
File:Homebrew_label_strange_haze.jpg|Strange Haze label, made with AI image generation&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== What I learned ==&lt;br /&gt;
=== Fermentation ===&lt;br /&gt;
* &#039;&#039;&#039;Dry hop by gravity, not by the calendar.&#039;&#039;&#039; The first dry hop goes in when gravity has dropped about half-way from OG to FG, while yeast is still active (biotransformation + it mops up hop-creep sugar). With fast yeast that can be Day 1, so start measuring every 12 h from Day 1.&lt;br /&gt;
* &#039;&#039;&#039;Split dry hop.&#039;&#039;&#039; DH1 during active fermentation for biotransformation; DH2 only after fermentation has stopped, so the CO₂ doesn&#039;t scrub the aroma away. After DH2, confirm gravity is stable again (second wave of hop creep). #0011 put both dry hops in while it was still fermenting, and the empty glass smelled only of malt.&lt;br /&gt;
* &#039;&#039;&#039;The last dry hop is the aroma you smell.&#039;&#039;&#039; It soaks latest, is scrubbed least and has the least time to oxidise, so put the signature hop (e.g. Citra) there, loose rather than in a bag. Only bag the dry hop you plan to pull out.&lt;br /&gt;
* &#039;&#039;&#039;Hop creep is real.&#039;&#039;&#039; Hop enzymes break down dextrins, so a late dry hop restarts fermentation: thinner body, higher ABV, and in bottles — gushers or bombs.&lt;br /&gt;
* &#039;&#039;&#039;&amp;quot;Green light&amp;quot; before packaging:&#039;&#039;&#039; two gravity readings ~48 h apart that differ by ≤ 0.001, &#039;&#039;&#039;and&#039;&#039;&#039; at least 3 days of diacetyl rest at a beer temperature of 20–22 °C (water bath 21.5 °C). Since 2026-09 the rest time replaces the forced diacetyl test: below 18 °C the yeast barely cleans up diacetyl, above 24 °C it only adds fusel alcohols, so 3 days at 20–22 °C is the guarantee. After DH2 or fruit, hop creep can make new diacetyl, so hold 20–22 °C for another 48 h and re-check gravity before cold-crashing. #0011 was packaged before creep had stopped and still had no butter flavour. If a beer does taste buttery, warm the keg back to room temperature for 3 days. Never package on a fixed day count — #0003-2, #0004 and #0008 all did, and all had carbonation problems.&lt;br /&gt;
* &#039;&#039;&#039;Pure oxygen 60–90 s before pitching&#039;&#039;&#039; lets a single dry pack ferment a 1.07+ wort; no starter needed. After pitching, oxygen is the enemy.&lt;br /&gt;
* Fermenting beer runs 1–2 °C above the water bath during the first 48 h. Set the bath ~1.5 °C below the target beer temperature at peak, then raise it to 21.5 °C for the diacetyl rest.&lt;br /&gt;
&lt;br /&gt;
=== Brew day ===&lt;br /&gt;
* &#039;&#039;&#039;Check pre-boil gravity and volume before boiling.&#039;&#039;&#039; Weigh the pot (total − empty pot weight ≈ litres), measure pre-boil gravity, and compute the target end volume = pre-boil points × litres ÷ target points. Boil to that volume, not to a timer. #0009 boiled off 4–5 L (normal is 1–2 L) and hit 1.110.&lt;br /&gt;
* &#039;&#039;&#039;Boil gently, lid off.&#039;&#039;&#039; Hard boiling over-concentrates bitterness and scorches. Excess evaporation turned out to be a big part of the &amp;quot;nasty bitterness&amp;quot; in #0007/#0008.&lt;br /&gt;
* &#039;&#039;&#039;Measure mash pH, not water pH, and be ready to go both ways.&#039;&#039;&#039; Sample 10–15 min after mash-in, cool to 20–25 °C, target 5.2–5.4. Keep both baking soda and phosphoric/lactic acid on hand. This soft, low-alkalinity water swings hard: ~1 g baking soda ≈ +0.67, 1 ml phosphoric ≈ −0.28 — use half steps and re-measure. Calibrate the pH pen every brew day and store the electrode wet.&lt;br /&gt;
* &#039;&#039;&#039;Post-boil pH is a second, separate checkpoint.&#039;&#039;&#039; A long boil concentrates the acidity too: #0011 boiled 14.55 L down to 11.5 L and dropped from 5.37 to 4.83. Correct the cooled wort to 5.2–5.4 before pitching; with a ~0.9 drop in fermentation that lands the beer at ~4.4. (#0010 split the same 2 g of baking soda between mash and post-boil and had no sourness.)&lt;br /&gt;
* &#039;&#039;&#039;Salt: CaCl₂·2H₂O 0.2 g per litre of total brewing water&#039;&#039;&#039; (≈ Ca 54 / Cl 96 ppm), no gypsum. The filtered water has almost no minerals; #0011&#039;s 0.5 g gave ~17 ppm chloride and ~9 ppm calcium, and the beer tasted like &amp;quot;sweet water&amp;quot;.&lt;br /&gt;
* Pale-malt grists (Extra Pale/Pilsner) tend to mash high (~5.7); grists with Munich plus gypsum tend to mash low.&lt;br /&gt;
&lt;br /&gt;
=== Bitterness and aroma ===&lt;br /&gt;
* &#039;&#039;&#039;Fix bitterness at the source, never by cutting hops.&#039;&#039;&#039; The culprits were pH (finished pH 3.83 = throat-biting), gypsum (sulfate sharpens bitterness), over-boiling, hops left in too long, and drinking too young. Bittering hops go in a bag, on time, and come out after the boil; dry-hop bags come out after 2–3 days.&lt;br /&gt;
* &#039;&#039;&#039;Whirlpool and dry hop do different jobs.&#039;&#039;&#039; Whirlpool (75–78 °C) gives juice on the palate; dry hop gives the aroma in the nose. Cutting either one loses something the other can&#039;t replace.&lt;br /&gt;
* Aging softens bitterness but kills hop aroma at the same time — fix bitterness at the source and drink fresh (2–3 weeks) instead.&lt;br /&gt;
* 0.5 kg oats + 1 kg wheat in 4.5 kg of grain is already enough haze; 1 kg of oats (20%) tasted grainy.&lt;br /&gt;
* &#039;&#039;&#039;Lactose: go big or skip it.&#039;&#039;&#039; 50 g in 11.5 L (4.3 g/L) is far below the ~20–25 g/L you can taste — it added nothing.&lt;br /&gt;
* &#039;&#039;&#039;A kit guarantees beer, not good beer.&#039;&#039;&#039; The PB Craft kit has no whirlpool hops, no specialty malt, and says nothing about salt or pH. Brewed exactly as printed (#0011), it gives a pale, thin, low-aroma beer; those four things have to be added.&lt;br /&gt;
&lt;br /&gt;
=== Packaging ===&lt;br /&gt;
* &#039;&#039;&#039;Keg + forced carbonation&#039;&#039;&#039; is now the default (#0009 on). Bottle conditioning kept failing: too much sugar (#0003-1), bottling before fermentation finished (#0003-2, #0004), and #0005 at ~9.5% ABV, where the yeast could not referment.&lt;br /&gt;
* &#039;&#039;&#039;Cold crash is now standard&#039;&#039;&#039; (since #0011): 2–4 °C for 24–48 h before transfer. #0006–#0011 never really cold-crashed, and all six got the same notes — weak aroma, &amp;quot;don&#039;t know what I&#039;m drinking&amp;quot;, grainy. Suspended yeast fits that picture; it is still a hypothesis, to be tested by drinking #0011&#039;s keg (no yeast) next to its bottles (live yeast). For bottle conditioning after a cold crash: under 7% ABV enough yeast remains; above 8%, add 1–2 g of an alcohol-tolerant yeast (e.g. EC-1118) at bottling.&lt;br /&gt;
* &#039;&#039;&#039;Don&#039;t package on a falling gravity with sulfite in the keg.&#039;&#039;&#039; #0011 was packaged at 1.020 while still dropping: no metabisulfite for an early package (keep the yeast alive), and halve the priming sugar for bottles (4.5 g/L — it carbonated without gushing).&lt;br /&gt;
* &#039;&#039;&#039;Never bottle the gravity samples.&#039;&#039;&#039; #0011&#039;s test bottle made from saved samples gushed and measured pH 3.5 — lactic infection from the open sample cup, not the fermenter (the normal bottles were 3.94). Drink samples on the spot; for a control bottle, fill one more from the main beer.&lt;br /&gt;
* Low-oxygen transfer: purge the keg with CO₂ three times, fill from the bottom through the liquid post, minimum headspace.&lt;br /&gt;
* Antioxidant in the keg: &#039;&#039;&#039;ascorbic acid 0.3 g + potassium metabisulfite 0.25 g per 10 L, always as a pair.&#039;&#039;&#039; Ascorbic acid alone makes H₂O₂, which with trace iron/copper (Fenton reaction) oxidises the beer worse than nothing. The chemicals are ~20% of the protection; the purging is the other 80%.&lt;br /&gt;
* &#039;&#039;&#039;Measure the fridge.&#039;&#039;&#039; The keg fridge is actually 7–8 °C, so 2.4–2.5 volumes needs ~15 PSI, not the textbook 11–12 PSI. #0009 and #0010 were both under-carbonated for this reason. A small 3 L keg also absorbs CO₂ slowly — #0011 was only ~70–80% carbonated after 5 days at pressure; lay it down and shake for 2–3 minutes.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_mini_keg_regulator.jpg|Mini keg with CO₂ regulator: carbonation is set by pressure, not sugar&lt;br /&gt;
File:Homebrew_mini_keg_tap.jpg|Mini keg with a wooden tap handle&lt;br /&gt;
File:Homebrew_bottles_box.jpg|A case of bottles filled from the keg&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
=== Method ===&lt;br /&gt;
* Write the plan down before brewing (one line: dry-hop day and gravity, green light, transfer, FG) and measure the error afterwards. The predictions are converging batch by batch.&lt;br /&gt;
* Log everything with a timestamp — gravity with date and time, pH with temperature. Most of the lessons above only became visible once the numbers were next to each other.&lt;br /&gt;
* Change one or two things per version. The recipe history (Crazy Concubine v1→v4.4, Orange Glow v1→v2.6) reads like a commit log: each version cites the batch that justified it.&lt;br /&gt;
&lt;br /&gt;
== Equipment ==&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_first_kit.jpg|Brew day #0001 (Dec 2025): kettle, fermenter, grain and hops&lt;br /&gt;
File:Homebrew_hop_pellets.jpg|Hop pellets, straight from the vacuum pack&lt;br /&gt;
File:Homebrew_mash.jpg|Mashing #0008, thermometer in the lid&lt;br /&gt;
File:Homebrew_ir_thermometer.jpg|IR thermometer check on the kettle&lt;br /&gt;
File:Homebrew_cooling_rig.jpg|Home-built cooling rig: thermoelectric cooler, power supply and thermostat, pumping cold water through the fermenter&lt;br /&gt;
File:Homebrew_o2_co2_cylinders.jpg|CO₂ (left) and pure O₂ (right) cylinders with regulators&lt;br /&gt;
File:Homebrew_fermenter_krausen.jpg|#0008 on Day 3: krausen, fermenting in the water bath&lt;br /&gt;
File:Homebrew_hydrometer.jpg|Gravity sample from #0009&lt;br /&gt;
File:Homebrew_rapt_pill.jpg|RAPT Pill wireless hydrometer (Nov 2026), next to its box&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
* Kettle + fermenter in a temperature-controlled water bath; weigh-the-pot volume estimation&lt;br /&gt;
* Pure O₂ cylinder with diffusion stone; refractometer + hydrometer; pH pen (4.01 / 7.00 buffers)&lt;br /&gt;
* Inkbird IHT-1P instant-read thermometer (from #0010)&lt;br /&gt;
* RAPT Pill wireless hydrometer and thermometer (from Nov 2026): floats in the fermenter and logs gravity and temperature over Bluetooth LE or 2.4 GHz Wi-Fi; range 0.95–1.12, rechargeable 18650 battery (USB-C)&lt;br /&gt;
* 8 L keg on external CO₂; 128 oz mini keg for in-keg conditioning&lt;br /&gt;
* Additives: Fermaid-O (2.5 g at pitch + 2.5 g at ~24 h), Maturex L ALDC 0.2–0.3 ml, CaCl₂&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=860</id>
		<title>Homebrewing</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=860"/>
		<updated>2026-11-20T15:32:15Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​#0012: mini keg photo, both photos in a gallery&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&#039;&#039;&#039;Homebrewing&#039;&#039;&#039; — eleven batches of hazy / New England IPA brewed at home in 10-litre batches, December 2025 to October 2026. Every batch is logged with gravity readings, pH, hop timing and a tasting note, and each recipe is version-controlled in git so every change can be traced back to the batch that caused it. This page summarises what was brewed, what went wrong, and the process rules that came out of it.&lt;br /&gt;
&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_glass_orange_glow.jpg|Orange Glow v2.3 (#0007), five weeks after brew day&lt;br /&gt;
File:Homebrew_glass_in_hand.jpg|A finished hazy IPA, in the glass with the piz brewing logo&lt;br /&gt;
File:Homebrew_keg_and_bottles.jpg|Keg and bottles, with the Strange Haze label in front&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Batches !! Period !! Batch size !! Styles !! Recipes (current version)&lt;br /&gt;
|-&lt;br /&gt;
| 11 (12 brews) || 2025-12 → 2026-10 || 6–11.5 L || Hazy IPA, DDH NEIPA, Milkshake IPA || 妃子起笑 Crazy Concubine (v4.4), Orange Glow (v2.6), Strange Haze (v1.1), Milkshake IPA (v1)&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The data ==&lt;br /&gt;
[[File:Homebrew_og_fg_by_batch.png|900px|center|Original and final gravity of every batch]]&lt;br /&gt;
The first six batches had no written recipe and wild original gravities (up to 1.110 — a 10.5% &amp;quot;IPA&amp;quot;). From #0007 on, every batch follows a versioned recipe and the OG has been pulled back into the 1.050–1.076 range the recipes are designed for.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_abv_attenuation.png|900px|center|ABV and apparent attenuation per batch]]&lt;br /&gt;
Attenuation went from ~71% (early batches, under-pitched, no oxygen) to a consistent 81–88% once pure-oxygen pitching, yeast nutrient and modern yeast strains were adopted. #0011&#039;s 62% is not a stuck fermentation: it is a kit recipe with 50 g of unfermentable lactose, and it landed exactly on the kit&#039;s printed OG 1.052 / FG 1.020.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_gravity_curves.png|900px|center|Fermentation curves of four batches]]&lt;br /&gt;
The single most useful chart. With one pack of Verdant and no oxygen (#0006) fermentation drags on for 12 days. With LalBrew Pomona, or two packs plus pure oxygen, ~90% of the sugar is gone within &#039;&#039;&#039;2 days&#039;&#039;&#039;. Any schedule written as &amp;quot;dry hop on Day 3&amp;quot; is simply wrong for these yeasts.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_plan_vs_actual.png|800px|center|Planned vs actual timing of key events]]&lt;br /&gt;
Since #0009 each batch starts with a one-line written plan (dry-hop day and gravity, green-light day, transfer day, expected FG). The log then measures the error. The first two plans were badly off — fermentation finished 8–12 days early — and each new plan is corrected from the previous error. #0011&#039;s DH1, DH2 and FG all landed within a day of plan (1.0, 0.8 and 0.8 days early).&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_ph.png|900px|center|Mash pH and finished-beer pH]]&lt;br /&gt;
pH turned out to be the hidden cause of &amp;quot;too bitter&amp;quot; — and, in #0011, of &amp;quot;slightly sour&amp;quot;. The same tap water gave a mash pH of 4.76 in one batch and 5.81 in another, so the correction has to go in either direction. #0011&#039;s mash was corrected to 5.37, but the wort measured 4.83 after the boil and was not corrected again; the beer finished at 3.94. In both #0008 (4.75 → 3.83) and #0011 (4.83 → 3.94) pH fell about &#039;&#039;&#039;0.9&#039;&#039;&#039; during fermentation, not the textbook 0.3–0.5 — this soft, low-alkalinity water has almost no buffer.&lt;br /&gt;
&lt;br /&gt;
== Fermentation with the RAPT Pill (#0012) ==&lt;br /&gt;
#0012 (Orange Glow v2.6, brewed 2026-11-10, LalBrew Pomona, OG 1.068) is the first batch logged by the [[#Equipment|RAPT Pill]]: a gravity and temperature reading every 15 minutes instead of a hydrometer sample every 12 h. The three charts below come from that batch and show how the fermentation was steered from the data.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_gravity_vs_days.png|900px|center|#0012: gravity vs days since pitch, with the expected DH1 window and the actual events]]&lt;br /&gt;
The plan is written in gravity, not days: DH1 goes in when gravity is half-way from OG to FG (1.039–1.029 for OG 1.068), and the diacetyl rest starts at ≤ 1.020. Krausen took ~13 h to appear (the wort was pitched at 22.8 °C and then cooled to 17–18 °C), but once going this yeast was faster than #0009 at the same stage. DH1 (56 g Nectaron + Citra, bagged) went in at 1.041 on Day 1.2 — deliberately ~2 points early, because dry-hopping early costs nothing in biotransformation and dry-hopping late is how #0010 missed its window. The bath went up to 21.5 °C for the diacetyl rest at 1.022 on Day 1.5, without waiting for the last 2 points.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_pill_telemetry.png|900px|center|#0012: RAPT Pill raw gravity, offset-corrected gravity, beer temperature and hydrometer checks]]&lt;br /&gt;
* &#039;&#039;&#039;The Pill reads high by a fixed offset.&#039;&#039;&#039; Two degassed hydrometer checks (at pitch and on Day 1) both put it ~13.5 points above the true gravity, so it is a constant offset, not a scaling error. The dashed line is Pill − 13.5. Recalibration is deliberately left until transfer: recalibrating in the middle of fermentation would put a step in the curve, and nobody reading it later could tell that step from fermentation. Alarm thresholds are converted to Pill units instead (DH1 window 1.039–1.029 → Pill 1.053–1.043).&lt;br /&gt;
* &#039;&#039;&#039;Ignore the Pill while krausen sits on it.&#039;&#039;&#039; Around hour 13 the raw reading &#039;&#039;rose&#039;&#039; from 1.0814 to 1.0869. Gravity cannot go up, and warming would make a tilt hydrometer read lower, not higher, so the cause is CO₂ bubbles and krausen pushing the Pill more upright. The bump marks the start of fermentation; it is not data.&lt;br /&gt;
* &#039;&#039;&#039;Degas every hydrometer sample.&#039;&#039;&#039; The Day 1 sample read 1.056 straight from the tap and 1.051 after degassing and temperature correction. What matters is bubbles stuck on the hydrometer, not whether the sample looks fizzy.&lt;br /&gt;
* &#039;&#039;&#039;Take beer temperature from the Pill, not from the sample.&#039;&#039;&#039; A sample warms up in the room within minutes. After the bath was set to 21.5 °C the beer climbed only ~0.3 °C/h, so the 3-day diacetyl rest is counted from when the &#039;&#039;beer&#039;&#039; reaches 21.5 °C, not from when the setpoint was changed.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_pill_rate.png|900px|center|#0012: gravity drop rate in points per day, with the 2 points/day green-light threshold]]&lt;br /&gt;
The rate is the change in Pill gravity smoothed over a ±1.5 h centred window (differences between neighbouring 15-minute readings are almost pure noise), in points per day — the same unit as RAPT&#039;s Gravity Velocity alert. It rose from ~31 points/day at hour 18 to ~73 around the start of the diacetyl rest, and has been falling since. When the line stays under 2 points/day, the gravity has stopped changing. That is one of the two green-light conditions; the other is 3 days at 20–22 °C.&lt;br /&gt;
&lt;br /&gt;
As of 2026-11-12 the raw Pill reading has levelled off at ~1.025, which is ~1.012 true — inside the 1.011–1.014 FG target, for about 7.4% ABV. Next: pull the DH1 bag and take a degassed hydrometer sample on 11/14 (to confirm FG and check whether the offset has drifted), then the green-light check and DH2 (84 g, loose) on 11/15.&lt;br /&gt;
&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_0012_glass.jpg|#0012 Orange Glow v2.6 in the glass, 2026-11-20&lt;br /&gt;
File:Homebrew_0012_keg_glass.jpg|#0012 poured from the mini keg, 2026-11-20&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== Batch log ==&lt;br /&gt;
{| class=&amp;quot;wikitable sortable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! # !! Date !! Beer / version !! Yeast !! OG → FG !! ABV !! Result / key lesson&lt;br /&gt;
|-&lt;br /&gt;
| 0001 || 2025-12-03 || DDH Hazy IPA || LalBrew New England || 1.062 → 1.018 || 5.8% || Bottle-conditioned. Lesson: add priming sugar, don&#039;t refrigerate during bottle conditioning.&lt;br /&gt;
|-&lt;br /&gt;
| 0002 || 2025-12-28 || Hazy IPA || Mauribrew Weiss || 1.040 → ? || – || Poor mash efficiency; dry-hop order wrong; forgot to measure FG before bottling.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-1 || 2026-02-26 || IPA || Mauribrew ALE514 || 1.050 → 1.010 || 5.3% || Flat aroma and taste; too much priming sugar, vegetal smell.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-2 || 2026-02-28 || IPA + 900 g oats || ALE514 || 1.056 → 1.016 || 5.3% || Under-pitched; bottled while airlock still active → gushers on opening.&lt;br /&gt;
|-&lt;br /&gt;
| 0004 || 2026-03-15 || Milkshake IPA || LalBrew New England || 1.110 → 1.030 || 10.5% || Lid not sealed (airlock silent 36 h); bottled while still fermenting; weak carbonation, slightly sour.&lt;br /&gt;
|-&lt;br /&gt;
| 0005 || 2026-04-23 || Milkshake IPA &amp;quot;light&amp;quot; || Verdant IPA || 1.102 → 1.030 || ~9.5% || &#039;&#039;&#039;Failed.&#039;&#039;&#039; Bittering hops 22 g instead of 6 g; bottles never carbonated — at ~9.5% ABV the yeast was too stressed to referment, and the cold crash removed what was left. Most poured away.&lt;br /&gt;
|-&lt;br /&gt;
| 0006 || 2026-05-09 || Hazy IPA || Verdant IPA || 1.062 → 1.008 || 7.1% || Second dry hop at Day 9 triggered &#039;&#039;&#039;hop creep&#039;&#039;&#039; (1.014 → 1.008). Clean, bitterness fine, aroma weak. First keg + forced carbonation.&lt;br /&gt;
|-&lt;br /&gt;
| 0007 || 2026-06-12 || Orange Glow v2.3 || Verdant IPA ||1.080 → 1.015 || 8.5% || First versioned recipe. Not hot, bitterness right after 3 weeks, but aroma lacking and colour unappealing.&lt;br /&gt;
|-&lt;br /&gt;
| 0008 || 2026-06-25 || Orange Glow v2.4 || LalBrew Pomona || 1.090 → 1.016 || 9.7% || Pomona hit 1.016 by Day 3 — dry hop missed its window. Finished pH 3.83 → harsh bitterness drowned the juice.&lt;br /&gt;
|-&lt;br /&gt;
| 0009 || 2026-07-29 || Orange Glow v2.5 &amp;quot;clean&amp;quot; || Pomona || 1.050 → 1.006 || 5.8% || Boiled off 4–5 L, OG shot to 1.110, over-diluted back to 1.050. pH 4.30 — bitterness finally right. Thin body, under-carbonated (keg fridge is 7–8 °C, not 3 °C). Passable daily hazy.&lt;br /&gt;
|-&lt;br /&gt;
| 0010 || 2026-08-26 || Strange Haze v1 (blueberry vanilla) || Verdant ×2 || 1.070 → 1.012 || 7.6% || Pre-boil OG check caught an over-concentrated wort in time. No harshness, no diacetyl — but not enough juice, sweetness, purple colour or vanilla; grainy from 20% oats.&lt;br /&gt;
|-&lt;br /&gt;
| 0011 || 2026-10-08 || 妃子起笑 v4.3 (PB Craft kit) || Weiss || 1.052 → 1.020 || 4.2% || Clean process — no diacetyl, oxidation or harsh bitterness, no bottle bombs — but weak aroma, thin &amp;quot;watery&amp;quot; body, pale (3–4 SRM) and slightly sour (pH 3.94). Every flaw traced to three fixable numbers: post-boil pH not corrected, CaCl₂ only 0.5 g, both dry hops added while still fermenting. Led to v4.4 and to recipe-wide changes.&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The beers ==&lt;br /&gt;
; 妃子起笑 · Crazy Concubine — Hazy Pale Ale (OG 1.052, ~5%)&lt;br /&gt;
: The first recipe. v1&#039;s second dry hop went in after fermentation had finished and hop creep ate the body (FG 1.014 → 1.008, 5.8% instead of 5%). v3 fixed it with zero new cost — one dry hop during active fermentation, cooler mash, pure-oxygen pitch, no cold crash. v4 added a 78 °C whirlpool for front-end juiciness. #0011 brewed the PB Craft kit exactly as printed (v4.3) and hit the kit&#039;s numbers, which showed the kit itself is the limit. v4.4: 10 L instead of 11.5 L (OG back to 1.060), CaCl₂ 3 g, a post-boil pH check, and the dry hops swapped — Idaho Gem first, Citra + Amarillo last. #0002, the same kit, had used that &amp;quot;wrong&amp;quot; order by mistake — and had the lychee aroma #0011 lacked.&lt;br /&gt;
; Orange Glow — DDH NEIPA (OG ~1.066, ~7%)&lt;br /&gt;
: The main development line: v1 → v2.6 across three brews. Swapped glucose for DME, Extra Pale for Golden Promise, Verdant for Pomona, added ALDC and Fermaid-O, moved to a split dry hop. #0009 drifted to a pale 5.8%; v2.6 brings back Munich for the orange-gold colour and uses a Strata / Mosaic / Nelson / Peacharine hop bill.&lt;br /&gt;
; Strange Haze — Imperial Milkshake IPA (blueberry vanilla)&lt;br /&gt;
: Built around a pre-printed label (9% ABV, purple nebula), so the fruit is blueberry for colour. v1 was &amp;quot;neither here nor there&amp;quot;; v1.1 cuts bitterness, adds lactose back, raises mash to 69–70 °C and splits 800 g of blueberry between mid-fermentation and post-fermentation.&lt;br /&gt;
; Milkshake IPA — low residual sugar (6 L)&lt;br /&gt;
: Body from oat/wheat protein instead of lots of lactose; unsweetened mango + passion fruit purée. Not yet brewed under this recipe.&lt;br /&gt;
&lt;br /&gt;
== Label and logo ==&lt;br /&gt;
The beers go out under the name &#039;&#039;&#039;piz brewing&#039;&#039;&#039;. The logo started from our dog, became a pencil sketch of just the ears and nose, and was finished with AI image generation. The labels were made the same way: many drafts generated and refined until one fit the beer.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_logo_dog.jpg|The original&lt;br /&gt;
File:Homebrew_logo_sketch.jpg|Pencil sketch: ears and nose&lt;br /&gt;
File:Homebrew_logo_final.jpg|Final logo&lt;br /&gt;
File:Homebrew_label_strange_haze.jpg|Strange Haze label, made with AI image generation&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== What I learned ==&lt;br /&gt;
=== Fermentation ===&lt;br /&gt;
* &#039;&#039;&#039;Dry hop by gravity, not by the calendar.&#039;&#039;&#039; The first dry hop goes in when gravity has dropped about half-way from OG to FG, while yeast is still active (biotransformation + it mops up hop-creep sugar). With fast yeast that can be Day 1, so start measuring every 12 h from Day 1.&lt;br /&gt;
* &#039;&#039;&#039;Split dry hop.&#039;&#039;&#039; DH1 during active fermentation for biotransformation; DH2 only after fermentation has stopped, so the CO₂ doesn&#039;t scrub the aroma away. After DH2, confirm gravity is stable again (second wave of hop creep). #0011 put both dry hops in while it was still fermenting, and the empty glass smelled only of malt.&lt;br /&gt;
* &#039;&#039;&#039;The last dry hop is the aroma you smell.&#039;&#039;&#039; It soaks latest, is scrubbed least and has the least time to oxidise, so put the signature hop (e.g. Citra) there, loose rather than in a bag. Only bag the dry hop you plan to pull out.&lt;br /&gt;
* &#039;&#039;&#039;Hop creep is real.&#039;&#039;&#039; Hop enzymes break down dextrins, so a late dry hop restarts fermentation: thinner body, higher ABV, and in bottles — gushers or bombs.&lt;br /&gt;
* &#039;&#039;&#039;&amp;quot;Green light&amp;quot; before packaging:&#039;&#039;&#039; two gravity readings ~48 h apart that differ by ≤ 0.001, &#039;&#039;&#039;and&#039;&#039;&#039; at least 3 days of diacetyl rest at a beer temperature of 20–22 °C (water bath 21.5 °C). Since 2026-09 the rest time replaces the forced diacetyl test: below 18 °C the yeast barely cleans up diacetyl, above 24 °C it only adds fusel alcohols, so 3 days at 20–22 °C is the guarantee. After DH2 or fruit, hop creep can make new diacetyl, so hold 20–22 °C for another 48 h and re-check gravity before cold-crashing. #0011 was packaged before creep had stopped and still had no butter flavour. If a beer does taste buttery, warm the keg back to room temperature for 3 days. Never package on a fixed day count — #0003-2, #0004 and #0008 all did, and all had carbonation problems.&lt;br /&gt;
* &#039;&#039;&#039;Pure oxygen 60–90 s before pitching&#039;&#039;&#039; lets a single dry pack ferment a 1.07+ wort; no starter needed. After pitching, oxygen is the enemy.&lt;br /&gt;
* Fermenting beer runs 1–2 °C above the water bath during the first 48 h. Set the bath ~1.5 °C below the target beer temperature at peak, then raise it to 21.5 °C for the diacetyl rest.&lt;br /&gt;
&lt;br /&gt;
=== Brew day ===&lt;br /&gt;
* &#039;&#039;&#039;Check pre-boil gravity and volume before boiling.&#039;&#039;&#039; Weigh the pot (total − empty pot weight ≈ litres), measure pre-boil gravity, and compute the target end volume = pre-boil points × litres ÷ target points. Boil to that volume, not to a timer. #0009 boiled off 4–5 L (normal is 1–2 L) and hit 1.110.&lt;br /&gt;
* &#039;&#039;&#039;Boil gently, lid off.&#039;&#039;&#039; Hard boiling over-concentrates bitterness and scorches. Excess evaporation turned out to be a big part of the &amp;quot;nasty bitterness&amp;quot; in #0007/#0008.&lt;br /&gt;
* &#039;&#039;&#039;Measure mash pH, not water pH, and be ready to go both ways.&#039;&#039;&#039; Sample 10–15 min after mash-in, cool to 20–25 °C, target 5.2–5.4. Keep both baking soda and phosphoric/lactic acid on hand. This soft, low-alkalinity water swings hard: ~1 g baking soda ≈ +0.67, 1 ml phosphoric ≈ −0.28 — use half steps and re-measure. Calibrate the pH pen every brew day and store the electrode wet.&lt;br /&gt;
* &#039;&#039;&#039;Post-boil pH is a second, separate checkpoint.&#039;&#039;&#039; A long boil concentrates the acidity too: #0011 boiled 14.55 L down to 11.5 L and dropped from 5.37 to 4.83. Correct the cooled wort to 5.2–5.4 before pitching; with a ~0.9 drop in fermentation that lands the beer at ~4.4. (#0010 split the same 2 g of baking soda between mash and post-boil and had no sourness.)&lt;br /&gt;
* &#039;&#039;&#039;Salt: CaCl₂·2H₂O 0.2 g per litre of total brewing water&#039;&#039;&#039; (≈ Ca 54 / Cl 96 ppm), no gypsum. The filtered water has almost no minerals; #0011&#039;s 0.5 g gave ~17 ppm chloride and ~9 ppm calcium, and the beer tasted like &amp;quot;sweet water&amp;quot;.&lt;br /&gt;
* Pale-malt grists (Extra Pale/Pilsner) tend to mash high (~5.7); grists with Munich plus gypsum tend to mash low.&lt;br /&gt;
&lt;br /&gt;
=== Bitterness and aroma ===&lt;br /&gt;
* &#039;&#039;&#039;Fix bitterness at the source, never by cutting hops.&#039;&#039;&#039; The culprits were pH (finished pH 3.83 = throat-biting), gypsum (sulfate sharpens bitterness), over-boiling, hops left in too long, and drinking too young. Bittering hops go in a bag, on time, and come out after the boil; dry-hop bags come out after 2–3 days.&lt;br /&gt;
* &#039;&#039;&#039;Whirlpool and dry hop do different jobs.&#039;&#039;&#039; Whirlpool (75–78 °C) gives juice on the palate; dry hop gives the aroma in the nose. Cutting either one loses something the other can&#039;t replace.&lt;br /&gt;
* Aging softens bitterness but kills hop aroma at the same time — fix bitterness at the source and drink fresh (2–3 weeks) instead.&lt;br /&gt;
* 0.5 kg oats + 1 kg wheat in 4.5 kg of grain is already enough haze; 1 kg of oats (20%) tasted grainy.&lt;br /&gt;
* &#039;&#039;&#039;Lactose: go big or skip it.&#039;&#039;&#039; 50 g in 11.5 L (4.3 g/L) is far below the ~20–25 g/L you can taste — it added nothing.&lt;br /&gt;
* &#039;&#039;&#039;A kit guarantees beer, not good beer.&#039;&#039;&#039; The PB Craft kit has no whirlpool hops, no specialty malt, and says nothing about salt or pH. Brewed exactly as printed (#0011), it gives a pale, thin, low-aroma beer; those four things have to be added.&lt;br /&gt;
&lt;br /&gt;
=== Packaging ===&lt;br /&gt;
* &#039;&#039;&#039;Keg + forced carbonation&#039;&#039;&#039; is now the default (#0009 on). Bottle conditioning kept failing: too much sugar (#0003-1), bottling before fermentation finished (#0003-2, #0004), and #0005 at ~9.5% ABV, where the yeast could not referment.&lt;br /&gt;
* &#039;&#039;&#039;Cold crash is now standard&#039;&#039;&#039; (since #0011): 2–4 °C for 24–48 h before transfer. #0006–#0011 never really cold-crashed, and all six got the same notes — weak aroma, &amp;quot;don&#039;t know what I&#039;m drinking&amp;quot;, grainy. Suspended yeast fits that picture; it is still a hypothesis, to be tested by drinking #0011&#039;s keg (no yeast) next to its bottles (live yeast). For bottle conditioning after a cold crash: under 7% ABV enough yeast remains; above 8%, add 1–2 g of an alcohol-tolerant yeast (e.g. EC-1118) at bottling.&lt;br /&gt;
* &#039;&#039;&#039;Don&#039;t package on a falling gravity with sulfite in the keg.&#039;&#039;&#039; #0011 was packaged at 1.020 while still dropping: no metabisulfite for an early package (keep the yeast alive), and halve the priming sugar for bottles (4.5 g/L — it carbonated without gushing).&lt;br /&gt;
* &#039;&#039;&#039;Never bottle the gravity samples.&#039;&#039;&#039; #0011&#039;s test bottle made from saved samples gushed and measured pH 3.5 — lactic infection from the open sample cup, not the fermenter (the normal bottles were 3.94). Drink samples on the spot; for a control bottle, fill one more from the main beer.&lt;br /&gt;
* Low-oxygen transfer: purge the keg with CO₂ three times, fill from the bottom through the liquid post, minimum headspace.&lt;br /&gt;
* Antioxidant in the keg: &#039;&#039;&#039;ascorbic acid 0.3 g + potassium metabisulfite 0.25 g per 10 L, always as a pair.&#039;&#039;&#039; Ascorbic acid alone makes H₂O₂, which with trace iron/copper (Fenton reaction) oxidises the beer worse than nothing. The chemicals are ~20% of the protection; the purging is the other 80%.&lt;br /&gt;
* &#039;&#039;&#039;Measure the fridge.&#039;&#039;&#039; The keg fridge is actually 7–8 °C, so 2.4–2.5 volumes needs ~15 PSI, not the textbook 11–12 PSI. #0009 and #0010 were both under-carbonated for this reason. A small 3 L keg also absorbs CO₂ slowly — #0011 was only ~70–80% carbonated after 5 days at pressure; lay it down and shake for 2–3 minutes.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_mini_keg_regulator.jpg|Mini keg with CO₂ regulator: carbonation is set by pressure, not sugar&lt;br /&gt;
File:Homebrew_mini_keg_tap.jpg|Mini keg with a wooden tap handle&lt;br /&gt;
File:Homebrew_bottles_box.jpg|A case of bottles filled from the keg&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
=== Method ===&lt;br /&gt;
* Write the plan down before brewing (one line: dry-hop day and gravity, green light, transfer, FG) and measure the error afterwards. The predictions are converging batch by batch.&lt;br /&gt;
* Log everything with a timestamp — gravity with date and time, pH with temperature. Most of the lessons above only became visible once the numbers were next to each other.&lt;br /&gt;
* Change one or two things per version. The recipe history (Crazy Concubine v1→v4.4, Orange Glow v1→v2.6) reads like a commit log: each version cites the batch that justified it.&lt;br /&gt;
&lt;br /&gt;
== Equipment ==&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_first_kit.jpg|Brew day #0001 (Dec 2025): kettle, fermenter, grain and hops&lt;br /&gt;
File:Homebrew_hop_pellets.jpg|Hop pellets, straight from the vacuum pack&lt;br /&gt;
File:Homebrew_mash.jpg|Mashing #0008, thermometer in the lid&lt;br /&gt;
File:Homebrew_ir_thermometer.jpg|IR thermometer check on the kettle&lt;br /&gt;
File:Homebrew_cooling_rig.jpg|Home-built cooling rig: thermoelectric cooler, power supply and thermostat, pumping cold water through the fermenter&lt;br /&gt;
File:Homebrew_o2_co2_cylinders.jpg|CO₂ (left) and pure O₂ (right) cylinders with regulators&lt;br /&gt;
File:Homebrew_fermenter_krausen.jpg|#0008 on Day 3: krausen, fermenting in the water bath&lt;br /&gt;
File:Homebrew_hydrometer.jpg|Gravity sample from #0009&lt;br /&gt;
File:Homebrew_rapt_pill.jpg|RAPT Pill wireless hydrometer (Nov 2026), next to its box&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
* Kettle + fermenter in a temperature-controlled water bath; weigh-the-pot volume estimation&lt;br /&gt;
* Pure O₂ cylinder with diffusion stone; refractometer + hydrometer; pH pen (4.01 / 7.00 buffers)&lt;br /&gt;
* Inkbird IHT-1P instant-read thermometer (from #0010)&lt;br /&gt;
* RAPT Pill wireless hydrometer and thermometer (from Nov 2026): floats in the fermenter and logs gravity and temperature over Bluetooth LE or 2.4 GHz Wi-Fi; range 0.95–1.12, rechargeable 18650 battery (USB-C)&lt;br /&gt;
* 8 L keg on external CO₂; 128 oz mini keg for in-keg conditioning&lt;br /&gt;
* Additives: Fermaid-O (2.5 g at pitch + 2.5 g at ~24 h), Maturex L ALDC 0.2–0.3 ml, CaCl₂&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_0012_keg_glass.jpg&amp;diff=859</id>
		<title>檔案:Homebrew 0012 keg glass.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_0012_keg_glass.jpg&amp;diff=859"/>
		<updated>2026-11-20T15:32:15Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
#0012 Orange Glow v2.6 poured from the mini keg (resized, metadata stripped)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
#0012 Orange Glow v2.6 poured from the mini keg (resized, metadata stripped)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=858</id>
		<title>Homebrewing</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=858"/>
		<updated>2026-11-20T14:24:15Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​#0012: photo of the beer in the glass&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&#039;&#039;&#039;Homebrewing&#039;&#039;&#039; — eleven batches of hazy / New England IPA brewed at home in 10-litre batches, December 2025 to October 2026. Every batch is logged with gravity readings, pH, hop timing and a tasting note, and each recipe is version-controlled in git so every change can be traced back to the batch that caused it. This page summarises what was brewed, what went wrong, and the process rules that came out of it.&lt;br /&gt;
&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_glass_orange_glow.jpg|Orange Glow v2.3 (#0007), five weeks after brew day&lt;br /&gt;
File:Homebrew_glass_in_hand.jpg|A finished hazy IPA, in the glass with the piz brewing logo&lt;br /&gt;
File:Homebrew_keg_and_bottles.jpg|Keg and bottles, with the Strange Haze label in front&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Batches !! Period !! Batch size !! Styles !! Recipes (current version)&lt;br /&gt;
|-&lt;br /&gt;
| 11 (12 brews) || 2025-12 → 2026-10 || 6–11.5 L || Hazy IPA, DDH NEIPA, Milkshake IPA || 妃子起笑 Crazy Concubine (v4.4), Orange Glow (v2.6), Strange Haze (v1.1), Milkshake IPA (v1)&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The data ==&lt;br /&gt;
[[File:Homebrew_og_fg_by_batch.png|900px|center|Original and final gravity of every batch]]&lt;br /&gt;
The first six batches had no written recipe and wild original gravities (up to 1.110 — a 10.5% &amp;quot;IPA&amp;quot;). From #0007 on, every batch follows a versioned recipe and the OG has been pulled back into the 1.050–1.076 range the recipes are designed for.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_abv_attenuation.png|900px|center|ABV and apparent attenuation per batch]]&lt;br /&gt;
Attenuation went from ~71% (early batches, under-pitched, no oxygen) to a consistent 81–88% once pure-oxygen pitching, yeast nutrient and modern yeast strains were adopted. #0011&#039;s 62% is not a stuck fermentation: it is a kit recipe with 50 g of unfermentable lactose, and it landed exactly on the kit&#039;s printed OG 1.052 / FG 1.020.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_gravity_curves.png|900px|center|Fermentation curves of four batches]]&lt;br /&gt;
The single most useful chart. With one pack of Verdant and no oxygen (#0006) fermentation drags on for 12 days. With LalBrew Pomona, or two packs plus pure oxygen, ~90% of the sugar is gone within &#039;&#039;&#039;2 days&#039;&#039;&#039;. Any schedule written as &amp;quot;dry hop on Day 3&amp;quot; is simply wrong for these yeasts.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_plan_vs_actual.png|800px|center|Planned vs actual timing of key events]]&lt;br /&gt;
Since #0009 each batch starts with a one-line written plan (dry-hop day and gravity, green-light day, transfer day, expected FG). The log then measures the error. The first two plans were badly off — fermentation finished 8–12 days early — and each new plan is corrected from the previous error. #0011&#039;s DH1, DH2 and FG all landed within a day of plan (1.0, 0.8 and 0.8 days early).&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_ph.png|900px|center|Mash pH and finished-beer pH]]&lt;br /&gt;
pH turned out to be the hidden cause of &amp;quot;too bitter&amp;quot; — and, in #0011, of &amp;quot;slightly sour&amp;quot;. The same tap water gave a mash pH of 4.76 in one batch and 5.81 in another, so the correction has to go in either direction. #0011&#039;s mash was corrected to 5.37, but the wort measured 4.83 after the boil and was not corrected again; the beer finished at 3.94. In both #0008 (4.75 → 3.83) and #0011 (4.83 → 3.94) pH fell about &#039;&#039;&#039;0.9&#039;&#039;&#039; during fermentation, not the textbook 0.3–0.5 — this soft, low-alkalinity water has almost no buffer.&lt;br /&gt;
&lt;br /&gt;
== Fermentation with the RAPT Pill (#0012) ==&lt;br /&gt;
#0012 (Orange Glow v2.6, brewed 2026-11-10, LalBrew Pomona, OG 1.068) is the first batch logged by the [[#Equipment|RAPT Pill]]: a gravity and temperature reading every 15 minutes instead of a hydrometer sample every 12 h. The three charts below come from that batch and show how the fermentation was steered from the data.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_gravity_vs_days.png|900px|center|#0012: gravity vs days since pitch, with the expected DH1 window and the actual events]]&lt;br /&gt;
The plan is written in gravity, not days: DH1 goes in when gravity is half-way from OG to FG (1.039–1.029 for OG 1.068), and the diacetyl rest starts at ≤ 1.020. Krausen took ~13 h to appear (the wort was pitched at 22.8 °C and then cooled to 17–18 °C), but once going this yeast was faster than #0009 at the same stage. DH1 (56 g Nectaron + Citra, bagged) went in at 1.041 on Day 1.2 — deliberately ~2 points early, because dry-hopping early costs nothing in biotransformation and dry-hopping late is how #0010 missed its window. The bath went up to 21.5 °C for the diacetyl rest at 1.022 on Day 1.5, without waiting for the last 2 points.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_pill_telemetry.png|900px|center|#0012: RAPT Pill raw gravity, offset-corrected gravity, beer temperature and hydrometer checks]]&lt;br /&gt;
* &#039;&#039;&#039;The Pill reads high by a fixed offset.&#039;&#039;&#039; Two degassed hydrometer checks (at pitch and on Day 1) both put it ~13.5 points above the true gravity, so it is a constant offset, not a scaling error. The dashed line is Pill − 13.5. Recalibration is deliberately left until transfer: recalibrating in the middle of fermentation would put a step in the curve, and nobody reading it later could tell that step from fermentation. Alarm thresholds are converted to Pill units instead (DH1 window 1.039–1.029 → Pill 1.053–1.043).&lt;br /&gt;
* &#039;&#039;&#039;Ignore the Pill while krausen sits on it.&#039;&#039;&#039; Around hour 13 the raw reading &#039;&#039;rose&#039;&#039; from 1.0814 to 1.0869. Gravity cannot go up, and warming would make a tilt hydrometer read lower, not higher, so the cause is CO₂ bubbles and krausen pushing the Pill more upright. The bump marks the start of fermentation; it is not data.&lt;br /&gt;
* &#039;&#039;&#039;Degas every hydrometer sample.&#039;&#039;&#039; The Day 1 sample read 1.056 straight from the tap and 1.051 after degassing and temperature correction. What matters is bubbles stuck on the hydrometer, not whether the sample looks fizzy.&lt;br /&gt;
* &#039;&#039;&#039;Take beer temperature from the Pill, not from the sample.&#039;&#039;&#039; A sample warms up in the room within minutes. After the bath was set to 21.5 °C the beer climbed only ~0.3 °C/h, so the 3-day diacetyl rest is counted from when the &#039;&#039;beer&#039;&#039; reaches 21.5 °C, not from when the setpoint was changed.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_pill_rate.png|900px|center|#0012: gravity drop rate in points per day, with the 2 points/day green-light threshold]]&lt;br /&gt;
The rate is the change in Pill gravity smoothed over a ±1.5 h centred window (differences between neighbouring 15-minute readings are almost pure noise), in points per day — the same unit as RAPT&#039;s Gravity Velocity alert. It rose from ~31 points/day at hour 18 to ~73 around the start of the diacetyl rest, and has been falling since. When the line stays under 2 points/day, the gravity has stopped changing. That is one of the two green-light conditions; the other is 3 days at 20–22 °C.&lt;br /&gt;
&lt;br /&gt;
As of 2026-11-12 the raw Pill reading has levelled off at ~1.025, which is ~1.012 true — inside the 1.011–1.014 FG target, for about 7.4% ABV. Next: pull the DH1 bag and take a degassed hydrometer sample on 11/14 (to confirm FG and check whether the offset has drifted), then the green-light check and DH2 (84 g, loose) on 11/15.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_glass.jpg|thumb|center|600px|#0012 Orange Glow v2.6 in the glass, 2026-11-20]]&lt;br /&gt;
&lt;br /&gt;
== Batch log ==&lt;br /&gt;
{| class=&amp;quot;wikitable sortable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! # !! Date !! Beer / version !! Yeast !! OG → FG !! ABV !! Result / key lesson&lt;br /&gt;
|-&lt;br /&gt;
| 0001 || 2025-12-03 || DDH Hazy IPA || LalBrew New England || 1.062 → 1.018 || 5.8% || Bottle-conditioned. Lesson: add priming sugar, don&#039;t refrigerate during bottle conditioning.&lt;br /&gt;
|-&lt;br /&gt;
| 0002 || 2025-12-28 || Hazy IPA || Mauribrew Weiss || 1.040 → ? || – || Poor mash efficiency; dry-hop order wrong; forgot to measure FG before bottling.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-1 || 2026-02-26 || IPA || Mauribrew ALE514 || 1.050 → 1.010 || 5.3% || Flat aroma and taste; too much priming sugar, vegetal smell.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-2 || 2026-02-28 || IPA + 900 g oats || ALE514 || 1.056 → 1.016 || 5.3% || Under-pitched; bottled while airlock still active → gushers on opening.&lt;br /&gt;
|-&lt;br /&gt;
| 0004 || 2026-03-15 || Milkshake IPA || LalBrew New England || 1.110 → 1.030 || 10.5% || Lid not sealed (airlock silent 36 h); bottled while still fermenting; weak carbonation, slightly sour.&lt;br /&gt;
|-&lt;br /&gt;
| 0005 || 2026-04-23 || Milkshake IPA &amp;quot;light&amp;quot; || Verdant IPA || 1.102 → 1.030 || ~9.5% || &#039;&#039;&#039;Failed.&#039;&#039;&#039; Bittering hops 22 g instead of 6 g; bottles never carbonated — at ~9.5% ABV the yeast was too stressed to referment, and the cold crash removed what was left. Most poured away.&lt;br /&gt;
|-&lt;br /&gt;
| 0006 || 2026-05-09 || Hazy IPA || Verdant IPA || 1.062 → 1.008 || 7.1% || Second dry hop at Day 9 triggered &#039;&#039;&#039;hop creep&#039;&#039;&#039; (1.014 → 1.008). Clean, bitterness fine, aroma weak. First keg + forced carbonation.&lt;br /&gt;
|-&lt;br /&gt;
| 0007 || 2026-06-12 || Orange Glow v2.3 || Verdant IPA ||1.080 → 1.015 || 8.5% || First versioned recipe. Not hot, bitterness right after 3 weeks, but aroma lacking and colour unappealing.&lt;br /&gt;
|-&lt;br /&gt;
| 0008 || 2026-06-25 || Orange Glow v2.4 || LalBrew Pomona || 1.090 → 1.016 || 9.7% || Pomona hit 1.016 by Day 3 — dry hop missed its window. Finished pH 3.83 → harsh bitterness drowned the juice.&lt;br /&gt;
|-&lt;br /&gt;
| 0009 || 2026-07-29 || Orange Glow v2.5 &amp;quot;clean&amp;quot; || Pomona || 1.050 → 1.006 || 5.8% || Boiled off 4–5 L, OG shot to 1.110, over-diluted back to 1.050. pH 4.30 — bitterness finally right. Thin body, under-carbonated (keg fridge is 7–8 °C, not 3 °C). Passable daily hazy.&lt;br /&gt;
|-&lt;br /&gt;
| 0010 || 2026-08-26 || Strange Haze v1 (blueberry vanilla) || Verdant ×2 || 1.070 → 1.012 || 7.6% || Pre-boil OG check caught an over-concentrated wort in time. No harshness, no diacetyl — but not enough juice, sweetness, purple colour or vanilla; grainy from 20% oats.&lt;br /&gt;
|-&lt;br /&gt;
| 0011 || 2026-10-08 || 妃子起笑 v4.3 (PB Craft kit) || Weiss || 1.052 → 1.020 || 4.2% || Clean process — no diacetyl, oxidation or harsh bitterness, no bottle bombs — but weak aroma, thin &amp;quot;watery&amp;quot; body, pale (3–4 SRM) and slightly sour (pH 3.94). Every flaw traced to three fixable numbers: post-boil pH not corrected, CaCl₂ only 0.5 g, both dry hops added while still fermenting. Led to v4.4 and to recipe-wide changes.&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The beers ==&lt;br /&gt;
; 妃子起笑 · Crazy Concubine — Hazy Pale Ale (OG 1.052, ~5%)&lt;br /&gt;
: The first recipe. v1&#039;s second dry hop went in after fermentation had finished and hop creep ate the body (FG 1.014 → 1.008, 5.8% instead of 5%). v3 fixed it with zero new cost — one dry hop during active fermentation, cooler mash, pure-oxygen pitch, no cold crash. v4 added a 78 °C whirlpool for front-end juiciness. #0011 brewed the PB Craft kit exactly as printed (v4.3) and hit the kit&#039;s numbers, which showed the kit itself is the limit. v4.4: 10 L instead of 11.5 L (OG back to 1.060), CaCl₂ 3 g, a post-boil pH check, and the dry hops swapped — Idaho Gem first, Citra + Amarillo last. #0002, the same kit, had used that &amp;quot;wrong&amp;quot; order by mistake — and had the lychee aroma #0011 lacked.&lt;br /&gt;
; Orange Glow — DDH NEIPA (OG ~1.066, ~7%)&lt;br /&gt;
: The main development line: v1 → v2.6 across three brews. Swapped glucose for DME, Extra Pale for Golden Promise, Verdant for Pomona, added ALDC and Fermaid-O, moved to a split dry hop. #0009 drifted to a pale 5.8%; v2.6 brings back Munich for the orange-gold colour and uses a Strata / Mosaic / Nelson / Peacharine hop bill.&lt;br /&gt;
; Strange Haze — Imperial Milkshake IPA (blueberry vanilla)&lt;br /&gt;
: Built around a pre-printed label (9% ABV, purple nebula), so the fruit is blueberry for colour. v1 was &amp;quot;neither here nor there&amp;quot;; v1.1 cuts bitterness, adds lactose back, raises mash to 69–70 °C and splits 800 g of blueberry between mid-fermentation and post-fermentation.&lt;br /&gt;
; Milkshake IPA — low residual sugar (6 L)&lt;br /&gt;
: Body from oat/wheat protein instead of lots of lactose; unsweetened mango + passion fruit purée. Not yet brewed under this recipe.&lt;br /&gt;
&lt;br /&gt;
== Label and logo ==&lt;br /&gt;
The beers go out under the name &#039;&#039;&#039;piz brewing&#039;&#039;&#039;. The logo started from our dog, became a pencil sketch of just the ears and nose, and was finished with AI image generation. The labels were made the same way: many drafts generated and refined until one fit the beer.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_logo_dog.jpg|The original&lt;br /&gt;
File:Homebrew_logo_sketch.jpg|Pencil sketch: ears and nose&lt;br /&gt;
File:Homebrew_logo_final.jpg|Final logo&lt;br /&gt;
File:Homebrew_label_strange_haze.jpg|Strange Haze label, made with AI image generation&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== What I learned ==&lt;br /&gt;
=== Fermentation ===&lt;br /&gt;
* &#039;&#039;&#039;Dry hop by gravity, not by the calendar.&#039;&#039;&#039; The first dry hop goes in when gravity has dropped about half-way from OG to FG, while yeast is still active (biotransformation + it mops up hop-creep sugar). With fast yeast that can be Day 1, so start measuring every 12 h from Day 1.&lt;br /&gt;
* &#039;&#039;&#039;Split dry hop.&#039;&#039;&#039; DH1 during active fermentation for biotransformation; DH2 only after fermentation has stopped, so the CO₂ doesn&#039;t scrub the aroma away. After DH2, confirm gravity is stable again (second wave of hop creep). #0011 put both dry hops in while it was still fermenting, and the empty glass smelled only of malt.&lt;br /&gt;
* &#039;&#039;&#039;The last dry hop is the aroma you smell.&#039;&#039;&#039; It soaks latest, is scrubbed least and has the least time to oxidise, so put the signature hop (e.g. Citra) there, loose rather than in a bag. Only bag the dry hop you plan to pull out.&lt;br /&gt;
* &#039;&#039;&#039;Hop creep is real.&#039;&#039;&#039; Hop enzymes break down dextrins, so a late dry hop restarts fermentation: thinner body, higher ABV, and in bottles — gushers or bombs.&lt;br /&gt;
* &#039;&#039;&#039;&amp;quot;Green light&amp;quot; before packaging:&#039;&#039;&#039; two gravity readings ~48 h apart that differ by ≤ 0.001, &#039;&#039;&#039;and&#039;&#039;&#039; at least 3 days of diacetyl rest at a beer temperature of 20–22 °C (water bath 21.5 °C). Since 2026-09 the rest time replaces the forced diacetyl test: below 18 °C the yeast barely cleans up diacetyl, above 24 °C it only adds fusel alcohols, so 3 days at 20–22 °C is the guarantee. After DH2 or fruit, hop creep can make new diacetyl, so hold 20–22 °C for another 48 h and re-check gravity before cold-crashing. #0011 was packaged before creep had stopped and still had no butter flavour. If a beer does taste buttery, warm the keg back to room temperature for 3 days. Never package on a fixed day count — #0003-2, #0004 and #0008 all did, and all had carbonation problems.&lt;br /&gt;
* &#039;&#039;&#039;Pure oxygen 60–90 s before pitching&#039;&#039;&#039; lets a single dry pack ferment a 1.07+ wort; no starter needed. After pitching, oxygen is the enemy.&lt;br /&gt;
* Fermenting beer runs 1–2 °C above the water bath during the first 48 h. Set the bath ~1.5 °C below the target beer temperature at peak, then raise it to 21.5 °C for the diacetyl rest.&lt;br /&gt;
&lt;br /&gt;
=== Brew day ===&lt;br /&gt;
* &#039;&#039;&#039;Check pre-boil gravity and volume before boiling.&#039;&#039;&#039; Weigh the pot (total − empty pot weight ≈ litres), measure pre-boil gravity, and compute the target end volume = pre-boil points × litres ÷ target points. Boil to that volume, not to a timer. #0009 boiled off 4–5 L (normal is 1–2 L) and hit 1.110.&lt;br /&gt;
* &#039;&#039;&#039;Boil gently, lid off.&#039;&#039;&#039; Hard boiling over-concentrates bitterness and scorches. Excess evaporation turned out to be a big part of the &amp;quot;nasty bitterness&amp;quot; in #0007/#0008.&lt;br /&gt;
* &#039;&#039;&#039;Measure mash pH, not water pH, and be ready to go both ways.&#039;&#039;&#039; Sample 10–15 min after mash-in, cool to 20–25 °C, target 5.2–5.4. Keep both baking soda and phosphoric/lactic acid on hand. This soft, low-alkalinity water swings hard: ~1 g baking soda ≈ +0.67, 1 ml phosphoric ≈ −0.28 — use half steps and re-measure. Calibrate the pH pen every brew day and store the electrode wet.&lt;br /&gt;
* &#039;&#039;&#039;Post-boil pH is a second, separate checkpoint.&#039;&#039;&#039; A long boil concentrates the acidity too: #0011 boiled 14.55 L down to 11.5 L and dropped from 5.37 to 4.83. Correct the cooled wort to 5.2–5.4 before pitching; with a ~0.9 drop in fermentation that lands the beer at ~4.4. (#0010 split the same 2 g of baking soda between mash and post-boil and had no sourness.)&lt;br /&gt;
* &#039;&#039;&#039;Salt: CaCl₂·2H₂O 0.2 g per litre of total brewing water&#039;&#039;&#039; (≈ Ca 54 / Cl 96 ppm), no gypsum. The filtered water has almost no minerals; #0011&#039;s 0.5 g gave ~17 ppm chloride and ~9 ppm calcium, and the beer tasted like &amp;quot;sweet water&amp;quot;.&lt;br /&gt;
* Pale-malt grists (Extra Pale/Pilsner) tend to mash high (~5.7); grists with Munich plus gypsum tend to mash low.&lt;br /&gt;
&lt;br /&gt;
=== Bitterness and aroma ===&lt;br /&gt;
* &#039;&#039;&#039;Fix bitterness at the source, never by cutting hops.&#039;&#039;&#039; The culprits were pH (finished pH 3.83 = throat-biting), gypsum (sulfate sharpens bitterness), over-boiling, hops left in too long, and drinking too young. Bittering hops go in a bag, on time, and come out after the boil; dry-hop bags come out after 2–3 days.&lt;br /&gt;
* &#039;&#039;&#039;Whirlpool and dry hop do different jobs.&#039;&#039;&#039; Whirlpool (75–78 °C) gives juice on the palate; dry hop gives the aroma in the nose. Cutting either one loses something the other can&#039;t replace.&lt;br /&gt;
* Aging softens bitterness but kills hop aroma at the same time — fix bitterness at the source and drink fresh (2–3 weeks) instead.&lt;br /&gt;
* 0.5 kg oats + 1 kg wheat in 4.5 kg of grain is already enough haze; 1 kg of oats (20%) tasted grainy.&lt;br /&gt;
* &#039;&#039;&#039;Lactose: go big or skip it.&#039;&#039;&#039; 50 g in 11.5 L (4.3 g/L) is far below the ~20–25 g/L you can taste — it added nothing.&lt;br /&gt;
* &#039;&#039;&#039;A kit guarantees beer, not good beer.&#039;&#039;&#039; The PB Craft kit has no whirlpool hops, no specialty malt, and says nothing about salt or pH. Brewed exactly as printed (#0011), it gives a pale, thin, low-aroma beer; those four things have to be added.&lt;br /&gt;
&lt;br /&gt;
=== Packaging ===&lt;br /&gt;
* &#039;&#039;&#039;Keg + forced carbonation&#039;&#039;&#039; is now the default (#0009 on). Bottle conditioning kept failing: too much sugar (#0003-1), bottling before fermentation finished (#0003-2, #0004), and #0005 at ~9.5% ABV, where the yeast could not referment.&lt;br /&gt;
* &#039;&#039;&#039;Cold crash is now standard&#039;&#039;&#039; (since #0011): 2–4 °C for 24–48 h before transfer. #0006–#0011 never really cold-crashed, and all six got the same notes — weak aroma, &amp;quot;don&#039;t know what I&#039;m drinking&amp;quot;, grainy. Suspended yeast fits that picture; it is still a hypothesis, to be tested by drinking #0011&#039;s keg (no yeast) next to its bottles (live yeast). For bottle conditioning after a cold crash: under 7% ABV enough yeast remains; above 8%, add 1–2 g of an alcohol-tolerant yeast (e.g. EC-1118) at bottling.&lt;br /&gt;
* &#039;&#039;&#039;Don&#039;t package on a falling gravity with sulfite in the keg.&#039;&#039;&#039; #0011 was packaged at 1.020 while still dropping: no metabisulfite for an early package (keep the yeast alive), and halve the priming sugar for bottles (4.5 g/L — it carbonated without gushing).&lt;br /&gt;
* &#039;&#039;&#039;Never bottle the gravity samples.&#039;&#039;&#039; #0011&#039;s test bottle made from saved samples gushed and measured pH 3.5 — lactic infection from the open sample cup, not the fermenter (the normal bottles were 3.94). Drink samples on the spot; for a control bottle, fill one more from the main beer.&lt;br /&gt;
* Low-oxygen transfer: purge the keg with CO₂ three times, fill from the bottom through the liquid post, minimum headspace.&lt;br /&gt;
* Antioxidant in the keg: &#039;&#039;&#039;ascorbic acid 0.3 g + potassium metabisulfite 0.25 g per 10 L, always as a pair.&#039;&#039;&#039; Ascorbic acid alone makes H₂O₂, which with trace iron/copper (Fenton reaction) oxidises the beer worse than nothing. The chemicals are ~20% of the protection; the purging is the other 80%.&lt;br /&gt;
* &#039;&#039;&#039;Measure the fridge.&#039;&#039;&#039; The keg fridge is actually 7–8 °C, so 2.4–2.5 volumes needs ~15 PSI, not the textbook 11–12 PSI. #0009 and #0010 were both under-carbonated for this reason. A small 3 L keg also absorbs CO₂ slowly — #0011 was only ~70–80% carbonated after 5 days at pressure; lay it down and shake for 2–3 minutes.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_mini_keg_regulator.jpg|Mini keg with CO₂ regulator: carbonation is set by pressure, not sugar&lt;br /&gt;
File:Homebrew_mini_keg_tap.jpg|Mini keg with a wooden tap handle&lt;br /&gt;
File:Homebrew_bottles_box.jpg|A case of bottles filled from the keg&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
=== Method ===&lt;br /&gt;
* Write the plan down before brewing (one line: dry-hop day and gravity, green light, transfer, FG) and measure the error afterwards. The predictions are converging batch by batch.&lt;br /&gt;
* Log everything with a timestamp — gravity with date and time, pH with temperature. Most of the lessons above only became visible once the numbers were next to each other.&lt;br /&gt;
* Change one or two things per version. The recipe history (Crazy Concubine v1→v4.4, Orange Glow v1→v2.6) reads like a commit log: each version cites the batch that justified it.&lt;br /&gt;
&lt;br /&gt;
== Equipment ==&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_first_kit.jpg|Brew day #0001 (Dec 2025): kettle, fermenter, grain and hops&lt;br /&gt;
File:Homebrew_hop_pellets.jpg|Hop pellets, straight from the vacuum pack&lt;br /&gt;
File:Homebrew_mash.jpg|Mashing #0008, thermometer in the lid&lt;br /&gt;
File:Homebrew_ir_thermometer.jpg|IR thermometer check on the kettle&lt;br /&gt;
File:Homebrew_cooling_rig.jpg|Home-built cooling rig: thermoelectric cooler, power supply and thermostat, pumping cold water through the fermenter&lt;br /&gt;
File:Homebrew_o2_co2_cylinders.jpg|CO₂ (left) and pure O₂ (right) cylinders with regulators&lt;br /&gt;
File:Homebrew_fermenter_krausen.jpg|#0008 on Day 3: krausen, fermenting in the water bath&lt;br /&gt;
File:Homebrew_hydrometer.jpg|Gravity sample from #0009&lt;br /&gt;
File:Homebrew_rapt_pill.jpg|RAPT Pill wireless hydrometer (Nov 2026), next to its box&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
* Kettle + fermenter in a temperature-controlled water bath; weigh-the-pot volume estimation&lt;br /&gt;
* Pure O₂ cylinder with diffusion stone; refractometer + hydrometer; pH pen (4.01 / 7.00 buffers)&lt;br /&gt;
* Inkbird IHT-1P instant-read thermometer (from #0010)&lt;br /&gt;
* RAPT Pill wireless hydrometer and thermometer (from Nov 2026): floats in the fermenter and logs gravity and temperature over Bluetooth LE or 2.4 GHz Wi-Fi; range 0.95–1.12, rechargeable 18650 battery (USB-C)&lt;br /&gt;
* 8 L keg on external CO₂; 128 oz mini keg for in-keg conditioning&lt;br /&gt;
* Additives: Fermaid-O (2.5 g at pitch + 2.5 g at ~24 h), Maturex L ALDC 0.2–0.3 ml, CaCl₂&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_0012_glass.jpg&amp;diff=857</id>
		<title>檔案:Homebrew 0012 glass.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_0012_glass.jpg&amp;diff=857"/>
		<updated>2026-11-20T14:24:14Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
#0012 Orange Glow v2.6 in the glass (resized, metadata stripped)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
#0012 Orange Glow v2.6 in the glass (resized, metadata stripped)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=856</id>
		<title>Homebrewing</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=856"/>
		<updated>2026-11-12T06:55:33Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​New section: fermentation with the RAPT Pill (#0012), 3 charts&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&#039;&#039;&#039;Homebrewing&#039;&#039;&#039; — eleven batches of hazy / New England IPA brewed at home in 10-litre batches, December 2025 to October 2026. Every batch is logged with gravity readings, pH, hop timing and a tasting note, and each recipe is version-controlled in git so every change can be traced back to the batch that caused it. This page summarises what was brewed, what went wrong, and the process rules that came out of it.&lt;br /&gt;
&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_glass_orange_glow.jpg|Orange Glow v2.3 (#0007), five weeks after brew day&lt;br /&gt;
File:Homebrew_glass_in_hand.jpg|A finished hazy IPA, in the glass with the piz brewing logo&lt;br /&gt;
File:Homebrew_keg_and_bottles.jpg|Keg and bottles, with the Strange Haze label in front&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Batches !! Period !! Batch size !! Styles !! Recipes (current version)&lt;br /&gt;
|-&lt;br /&gt;
| 11 (12 brews) || 2025-12 → 2026-10 || 6–11.5 L || Hazy IPA, DDH NEIPA, Milkshake IPA || 妃子起笑 Crazy Concubine (v4.4), Orange Glow (v2.6), Strange Haze (v1.1), Milkshake IPA (v1)&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The data ==&lt;br /&gt;
[[File:Homebrew_og_fg_by_batch.png|900px|center|Original and final gravity of every batch]]&lt;br /&gt;
The first six batches had no written recipe and wild original gravities (up to 1.110 — a 10.5% &amp;quot;IPA&amp;quot;). From #0007 on, every batch follows a versioned recipe and the OG has been pulled back into the 1.050–1.076 range the recipes are designed for.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_abv_attenuation.png|900px|center|ABV and apparent attenuation per batch]]&lt;br /&gt;
Attenuation went from ~71% (early batches, under-pitched, no oxygen) to a consistent 81–88% once pure-oxygen pitching, yeast nutrient and modern yeast strains were adopted. #0011&#039;s 62% is not a stuck fermentation: it is a kit recipe with 50 g of unfermentable lactose, and it landed exactly on the kit&#039;s printed OG 1.052 / FG 1.020.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_gravity_curves.png|900px|center|Fermentation curves of four batches]]&lt;br /&gt;
The single most useful chart. With one pack of Verdant and no oxygen (#0006) fermentation drags on for 12 days. With LalBrew Pomona, or two packs plus pure oxygen, ~90% of the sugar is gone within &#039;&#039;&#039;2 days&#039;&#039;&#039;. Any schedule written as &amp;quot;dry hop on Day 3&amp;quot; is simply wrong for these yeasts.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_plan_vs_actual.png|800px|center|Planned vs actual timing of key events]]&lt;br /&gt;
Since #0009 each batch starts with a one-line written plan (dry-hop day and gravity, green-light day, transfer day, expected FG). The log then measures the error. The first two plans were badly off — fermentation finished 8–12 days early — and each new plan is corrected from the previous error. #0011&#039;s DH1, DH2 and FG all landed within a day of plan (1.0, 0.8 and 0.8 days early).&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_ph.png|900px|center|Mash pH and finished-beer pH]]&lt;br /&gt;
pH turned out to be the hidden cause of &amp;quot;too bitter&amp;quot; — and, in #0011, of &amp;quot;slightly sour&amp;quot;. The same tap water gave a mash pH of 4.76 in one batch and 5.81 in another, so the correction has to go in either direction. #0011&#039;s mash was corrected to 5.37, but the wort measured 4.83 after the boil and was not corrected again; the beer finished at 3.94. In both #0008 (4.75 → 3.83) and #0011 (4.83 → 3.94) pH fell about &#039;&#039;&#039;0.9&#039;&#039;&#039; during fermentation, not the textbook 0.3–0.5 — this soft, low-alkalinity water has almost no buffer.&lt;br /&gt;
&lt;br /&gt;
== Fermentation with the RAPT Pill (#0012) ==&lt;br /&gt;
#0012 (Orange Glow v2.6, brewed 2026-11-10, LalBrew Pomona, OG 1.068) is the first batch logged by the [[#Equipment|RAPT Pill]]: a gravity and temperature reading every 15 minutes instead of a hydrometer sample every 12 h. The three charts below come from that batch and show how the fermentation was steered from the data.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_gravity_vs_days.png|900px|center|#0012: gravity vs days since pitch, with the expected DH1 window and the actual events]]&lt;br /&gt;
The plan is written in gravity, not days: DH1 goes in when gravity is half-way from OG to FG (1.039–1.029 for OG 1.068), and the diacetyl rest starts at ≤ 1.020. Krausen took ~13 h to appear (the wort was pitched at 22.8 °C and then cooled to 17–18 °C), but once going this yeast was faster than #0009 at the same stage. DH1 (56 g Nectaron + Citra, bagged) went in at 1.041 on Day 1.2 — deliberately ~2 points early, because dry-hopping early costs nothing in biotransformation and dry-hopping late is how #0010 missed its window. The bath went up to 21.5 °C for the diacetyl rest at 1.022 on Day 1.5, without waiting for the last 2 points.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_pill_telemetry.png|900px|center|#0012: RAPT Pill raw gravity, offset-corrected gravity, beer temperature and hydrometer checks]]&lt;br /&gt;
* &#039;&#039;&#039;The Pill reads high by a fixed offset.&#039;&#039;&#039; Two degassed hydrometer checks (at pitch and on Day 1) both put it ~13.5 points above the true gravity, so it is a constant offset, not a scaling error. The dashed line is Pill − 13.5. Recalibration is deliberately left until transfer: recalibrating in the middle of fermentation would put a step in the curve, and nobody reading it later could tell that step from fermentation. Alarm thresholds are converted to Pill units instead (DH1 window 1.039–1.029 → Pill 1.053–1.043).&lt;br /&gt;
* &#039;&#039;&#039;Ignore the Pill while krausen sits on it.&#039;&#039;&#039; Around hour 13 the raw reading &#039;&#039;rose&#039;&#039; from 1.0814 to 1.0869. Gravity cannot go up, and warming would make a tilt hydrometer read lower, not higher, so the cause is CO₂ bubbles and krausen pushing the Pill more upright. The bump marks the start of fermentation; it is not data.&lt;br /&gt;
* &#039;&#039;&#039;Degas every hydrometer sample.&#039;&#039;&#039; The Day 1 sample read 1.056 straight from the tap and 1.051 after degassing and temperature correction. What matters is bubbles stuck on the hydrometer, not whether the sample looks fizzy.&lt;br /&gt;
* &#039;&#039;&#039;Take beer temperature from the Pill, not from the sample.&#039;&#039;&#039; A sample warms up in the room within minutes. After the bath was set to 21.5 °C the beer climbed only ~0.3 °C/h, so the 3-day diacetyl rest is counted from when the &#039;&#039;beer&#039;&#039; reaches 21.5 °C, not from when the setpoint was changed.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_0012_pill_rate.png|900px|center|#0012: gravity drop rate in points per day, with the 2 points/day green-light threshold]]&lt;br /&gt;
The rate is the change in Pill gravity smoothed over a ±1.5 h centred window (differences between neighbouring 15-minute readings are almost pure noise), in points per day — the same unit as RAPT&#039;s Gravity Velocity alert. It rose from ~31 points/day at hour 18 to ~73 around the start of the diacetyl rest, and has been falling since. When the line stays under 2 points/day, the gravity has stopped changing. That is one of the two green-light conditions; the other is 3 days at 20–22 °C.&lt;br /&gt;
&lt;br /&gt;
As of 2026-11-12 the raw Pill reading has levelled off at ~1.025, which is ~1.012 true — inside the 1.011–1.014 FG target, for about 7.4% ABV. Next: pull the DH1 bag and take a degassed hydrometer sample on 11/14 (to confirm FG and check whether the offset has drifted), then the green-light check and DH2 (84 g, loose) on 11/15.&lt;br /&gt;
&lt;br /&gt;
== Batch log ==&lt;br /&gt;
{| class=&amp;quot;wikitable sortable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! # !! Date !! Beer / version !! Yeast !! OG → FG !! ABV !! Result / key lesson&lt;br /&gt;
|-&lt;br /&gt;
| 0001 || 2025-12-03 || DDH Hazy IPA || LalBrew New England || 1.062 → 1.018 || 5.8% || Bottle-conditioned. Lesson: add priming sugar, don&#039;t refrigerate during bottle conditioning.&lt;br /&gt;
|-&lt;br /&gt;
| 0002 || 2025-12-28 || Hazy IPA || Mauribrew Weiss || 1.040 → ? || – || Poor mash efficiency; dry-hop order wrong; forgot to measure FG before bottling.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-1 || 2026-02-26 || IPA || Mauribrew ALE514 || 1.050 → 1.010 || 5.3% || Flat aroma and taste; too much priming sugar, vegetal smell.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-2 || 2026-02-28 || IPA + 900 g oats || ALE514 || 1.056 → 1.016 || 5.3% || Under-pitched; bottled while airlock still active → gushers on opening.&lt;br /&gt;
|-&lt;br /&gt;
| 0004 || 2026-03-15 || Milkshake IPA || LalBrew New England || 1.110 → 1.030 || 10.5% || Lid not sealed (airlock silent 36 h); bottled while still fermenting; weak carbonation, slightly sour.&lt;br /&gt;
|-&lt;br /&gt;
| 0005 || 2026-04-23 || Milkshake IPA &amp;quot;light&amp;quot; || Verdant IPA || 1.102 → 1.030 || ~9.5% || &#039;&#039;&#039;Failed.&#039;&#039;&#039; Bittering hops 22 g instead of 6 g; bottles never carbonated — at ~9.5% ABV the yeast was too stressed to referment, and the cold crash removed what was left. Most poured away.&lt;br /&gt;
|-&lt;br /&gt;
| 0006 || 2026-05-09 || Hazy IPA || Verdant IPA || 1.062 → 1.008 || 7.1% || Second dry hop at Day 9 triggered &#039;&#039;&#039;hop creep&#039;&#039;&#039; (1.014 → 1.008). Clean, bitterness fine, aroma weak. First keg + forced carbonation.&lt;br /&gt;
|-&lt;br /&gt;
| 0007 || 2026-06-12 || Orange Glow v2.3 || Verdant IPA ||1.080 → 1.015 || 8.5% || First versioned recipe. Not hot, bitterness right after 3 weeks, but aroma lacking and colour unappealing.&lt;br /&gt;
|-&lt;br /&gt;
| 0008 || 2026-06-25 || Orange Glow v2.4 || LalBrew Pomona || 1.090 → 1.016 || 9.7% || Pomona hit 1.016 by Day 3 — dry hop missed its window. Finished pH 3.83 → harsh bitterness drowned the juice.&lt;br /&gt;
|-&lt;br /&gt;
| 0009 || 2026-07-29 || Orange Glow v2.5 &amp;quot;clean&amp;quot; || Pomona || 1.050 → 1.006 || 5.8% || Boiled off 4–5 L, OG shot to 1.110, over-diluted back to 1.050. pH 4.30 — bitterness finally right. Thin body, under-carbonated (keg fridge is 7–8 °C, not 3 °C). Passable daily hazy.&lt;br /&gt;
|-&lt;br /&gt;
| 0010 || 2026-08-26 || Strange Haze v1 (blueberry vanilla) || Verdant ×2 || 1.070 → 1.012 || 7.6% || Pre-boil OG check caught an over-concentrated wort in time. No harshness, no diacetyl — but not enough juice, sweetness, purple colour or vanilla; grainy from 20% oats.&lt;br /&gt;
|-&lt;br /&gt;
| 0011 || 2026-10-08 || 妃子起笑 v4.3 (PB Craft kit) || Weiss || 1.052 → 1.020 || 4.2% || Clean process — no diacetyl, oxidation or harsh bitterness, no bottle bombs — but weak aroma, thin &amp;quot;watery&amp;quot; body, pale (3–4 SRM) and slightly sour (pH 3.94). Every flaw traced to three fixable numbers: post-boil pH not corrected, CaCl₂ only 0.5 g, both dry hops added while still fermenting. Led to v4.4 and to recipe-wide changes.&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The beers ==&lt;br /&gt;
; 妃子起笑 · Crazy Concubine — Hazy Pale Ale (OG 1.052, ~5%)&lt;br /&gt;
: The first recipe. v1&#039;s second dry hop went in after fermentation had finished and hop creep ate the body (FG 1.014 → 1.008, 5.8% instead of 5%). v3 fixed it with zero new cost — one dry hop during active fermentation, cooler mash, pure-oxygen pitch, no cold crash. v4 added a 78 °C whirlpool for front-end juiciness. #0011 brewed the PB Craft kit exactly as printed (v4.3) and hit the kit&#039;s numbers, which showed the kit itself is the limit. v4.4: 10 L instead of 11.5 L (OG back to 1.060), CaCl₂ 3 g, a post-boil pH check, and the dry hops swapped — Idaho Gem first, Citra + Amarillo last. #0002, the same kit, had used that &amp;quot;wrong&amp;quot; order by mistake — and had the lychee aroma #0011 lacked.&lt;br /&gt;
; Orange Glow — DDH NEIPA (OG ~1.066, ~7%)&lt;br /&gt;
: The main development line: v1 → v2.6 across three brews. Swapped glucose for DME, Extra Pale for Golden Promise, Verdant for Pomona, added ALDC and Fermaid-O, moved to a split dry hop. #0009 drifted to a pale 5.8%; v2.6 brings back Munich for the orange-gold colour and uses a Strata / Mosaic / Nelson / Peacharine hop bill.&lt;br /&gt;
; Strange Haze — Imperial Milkshake IPA (blueberry vanilla)&lt;br /&gt;
: Built around a pre-printed label (9% ABV, purple nebula), so the fruit is blueberry for colour. v1 was &amp;quot;neither here nor there&amp;quot;; v1.1 cuts bitterness, adds lactose back, raises mash to 69–70 °C and splits 800 g of blueberry between mid-fermentation and post-fermentation.&lt;br /&gt;
; Milkshake IPA — low residual sugar (6 L)&lt;br /&gt;
: Body from oat/wheat protein instead of lots of lactose; unsweetened mango + passion fruit purée. Not yet brewed under this recipe.&lt;br /&gt;
&lt;br /&gt;
== Label and logo ==&lt;br /&gt;
The beers go out under the name &#039;&#039;&#039;piz brewing&#039;&#039;&#039;. The logo started from our dog, became a pencil sketch of just the ears and nose, and was finished with AI image generation. The labels were made the same way: many drafts generated and refined until one fit the beer.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_logo_dog.jpg|The original&lt;br /&gt;
File:Homebrew_logo_sketch.jpg|Pencil sketch: ears and nose&lt;br /&gt;
File:Homebrew_logo_final.jpg|Final logo&lt;br /&gt;
File:Homebrew_label_strange_haze.jpg|Strange Haze label, made with AI image generation&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== What I learned ==&lt;br /&gt;
=== Fermentation ===&lt;br /&gt;
* &#039;&#039;&#039;Dry hop by gravity, not by the calendar.&#039;&#039;&#039; The first dry hop goes in when gravity has dropped about half-way from OG to FG, while yeast is still active (biotransformation + it mops up hop-creep sugar). With fast yeast that can be Day 1, so start measuring every 12 h from Day 1.&lt;br /&gt;
* &#039;&#039;&#039;Split dry hop.&#039;&#039;&#039; DH1 during active fermentation for biotransformation; DH2 only after fermentation has stopped, so the CO₂ doesn&#039;t scrub the aroma away. After DH2, confirm gravity is stable again (second wave of hop creep). #0011 put both dry hops in while it was still fermenting, and the empty glass smelled only of malt.&lt;br /&gt;
* &#039;&#039;&#039;The last dry hop is the aroma you smell.&#039;&#039;&#039; It soaks latest, is scrubbed least and has the least time to oxidise, so put the signature hop (e.g. Citra) there, loose rather than in a bag. Only bag the dry hop you plan to pull out.&lt;br /&gt;
* &#039;&#039;&#039;Hop creep is real.&#039;&#039;&#039; Hop enzymes break down dextrins, so a late dry hop restarts fermentation: thinner body, higher ABV, and in bottles — gushers or bombs.&lt;br /&gt;
* &#039;&#039;&#039;&amp;quot;Green light&amp;quot; before packaging:&#039;&#039;&#039; two gravity readings ~48 h apart that differ by ≤ 0.001, &#039;&#039;&#039;and&#039;&#039;&#039; at least 3 days of diacetyl rest at a beer temperature of 20–22 °C (water bath 21.5 °C). Since 2026-09 the rest time replaces the forced diacetyl test: below 18 °C the yeast barely cleans up diacetyl, above 24 °C it only adds fusel alcohols, so 3 days at 20–22 °C is the guarantee. After DH2 or fruit, hop creep can make new diacetyl, so hold 20–22 °C for another 48 h and re-check gravity before cold-crashing. #0011 was packaged before creep had stopped and still had no butter flavour. If a beer does taste buttery, warm the keg back to room temperature for 3 days. Never package on a fixed day count — #0003-2, #0004 and #0008 all did, and all had carbonation problems.&lt;br /&gt;
* &#039;&#039;&#039;Pure oxygen 60–90 s before pitching&#039;&#039;&#039; lets a single dry pack ferment a 1.07+ wort; no starter needed. After pitching, oxygen is the enemy.&lt;br /&gt;
* Fermenting beer runs 1–2 °C above the water bath during the first 48 h. Set the bath ~1.5 °C below the target beer temperature at peak, then raise it to 21.5 °C for the diacetyl rest.&lt;br /&gt;
&lt;br /&gt;
=== Brew day ===&lt;br /&gt;
* &#039;&#039;&#039;Check pre-boil gravity and volume before boiling.&#039;&#039;&#039; Weigh the pot (total − empty pot weight ≈ litres), measure pre-boil gravity, and compute the target end volume = pre-boil points × litres ÷ target points. Boil to that volume, not to a timer. #0009 boiled off 4–5 L (normal is 1–2 L) and hit 1.110.&lt;br /&gt;
* &#039;&#039;&#039;Boil gently, lid off.&#039;&#039;&#039; Hard boiling over-concentrates bitterness and scorches. Excess evaporation turned out to be a big part of the &amp;quot;nasty bitterness&amp;quot; in #0007/#0008.&lt;br /&gt;
* &#039;&#039;&#039;Measure mash pH, not water pH, and be ready to go both ways.&#039;&#039;&#039; Sample 10–15 min after mash-in, cool to 20–25 °C, target 5.2–5.4. Keep both baking soda and phosphoric/lactic acid on hand. This soft, low-alkalinity water swings hard: ~1 g baking soda ≈ +0.67, 1 ml phosphoric ≈ −0.28 — use half steps and re-measure. Calibrate the pH pen every brew day and store the electrode wet.&lt;br /&gt;
* &#039;&#039;&#039;Post-boil pH is a second, separate checkpoint.&#039;&#039;&#039; A long boil concentrates the acidity too: #0011 boiled 14.55 L down to 11.5 L and dropped from 5.37 to 4.83. Correct the cooled wort to 5.2–5.4 before pitching; with a ~0.9 drop in fermentation that lands the beer at ~4.4. (#0010 split the same 2 g of baking soda between mash and post-boil and had no sourness.)&lt;br /&gt;
* &#039;&#039;&#039;Salt: CaCl₂·2H₂O 0.2 g per litre of total brewing water&#039;&#039;&#039; (≈ Ca 54 / Cl 96 ppm), no gypsum. The filtered water has almost no minerals; #0011&#039;s 0.5 g gave ~17 ppm chloride and ~9 ppm calcium, and the beer tasted like &amp;quot;sweet water&amp;quot;.&lt;br /&gt;
* Pale-malt grists (Extra Pale/Pilsner) tend to mash high (~5.7); grists with Munich plus gypsum tend to mash low.&lt;br /&gt;
&lt;br /&gt;
=== Bitterness and aroma ===&lt;br /&gt;
* &#039;&#039;&#039;Fix bitterness at the source, never by cutting hops.&#039;&#039;&#039; The culprits were pH (finished pH 3.83 = throat-biting), gypsum (sulfate sharpens bitterness), over-boiling, hops left in too long, and drinking too young. Bittering hops go in a bag, on time, and come out after the boil; dry-hop bags come out after 2–3 days.&lt;br /&gt;
* &#039;&#039;&#039;Whirlpool and dry hop do different jobs.&#039;&#039;&#039; Whirlpool (75–78 °C) gives juice on the palate; dry hop gives the aroma in the nose. Cutting either one loses something the other can&#039;t replace.&lt;br /&gt;
* Aging softens bitterness but kills hop aroma at the same time — fix bitterness at the source and drink fresh (2–3 weeks) instead.&lt;br /&gt;
* 0.5 kg oats + 1 kg wheat in 4.5 kg of grain is already enough haze; 1 kg of oats (20%) tasted grainy.&lt;br /&gt;
* &#039;&#039;&#039;Lactose: go big or skip it.&#039;&#039;&#039; 50 g in 11.5 L (4.3 g/L) is far below the ~20–25 g/L you can taste — it added nothing.&lt;br /&gt;
* &#039;&#039;&#039;A kit guarantees beer, not good beer.&#039;&#039;&#039; The PB Craft kit has no whirlpool hops, no specialty malt, and says nothing about salt or pH. Brewed exactly as printed (#0011), it gives a pale, thin, low-aroma beer; those four things have to be added.&lt;br /&gt;
&lt;br /&gt;
=== Packaging ===&lt;br /&gt;
* &#039;&#039;&#039;Keg + forced carbonation&#039;&#039;&#039; is now the default (#0009 on). Bottle conditioning kept failing: too much sugar (#0003-1), bottling before fermentation finished (#0003-2, #0004), and #0005 at ~9.5% ABV, where the yeast could not referment.&lt;br /&gt;
* &#039;&#039;&#039;Cold crash is now standard&#039;&#039;&#039; (since #0011): 2–4 °C for 24–48 h before transfer. #0006–#0011 never really cold-crashed, and all six got the same notes — weak aroma, &amp;quot;don&#039;t know what I&#039;m drinking&amp;quot;, grainy. Suspended yeast fits that picture; it is still a hypothesis, to be tested by drinking #0011&#039;s keg (no yeast) next to its bottles (live yeast). For bottle conditioning after a cold crash: under 7% ABV enough yeast remains; above 8%, add 1–2 g of an alcohol-tolerant yeast (e.g. EC-1118) at bottling.&lt;br /&gt;
* &#039;&#039;&#039;Don&#039;t package on a falling gravity with sulfite in the keg.&#039;&#039;&#039; #0011 was packaged at 1.020 while still dropping: no metabisulfite for an early package (keep the yeast alive), and halve the priming sugar for bottles (4.5 g/L — it carbonated without gushing).&lt;br /&gt;
* &#039;&#039;&#039;Never bottle the gravity samples.&#039;&#039;&#039; #0011&#039;s test bottle made from saved samples gushed and measured pH 3.5 — lactic infection from the open sample cup, not the fermenter (the normal bottles were 3.94). Drink samples on the spot; for a control bottle, fill one more from the main beer.&lt;br /&gt;
* Low-oxygen transfer: purge the keg with CO₂ three times, fill from the bottom through the liquid post, minimum headspace.&lt;br /&gt;
* Antioxidant in the keg: &#039;&#039;&#039;ascorbic acid 0.3 g + potassium metabisulfite 0.25 g per 10 L, always as a pair.&#039;&#039;&#039; Ascorbic acid alone makes H₂O₂, which with trace iron/copper (Fenton reaction) oxidises the beer worse than nothing. The chemicals are ~20% of the protection; the purging is the other 80%.&lt;br /&gt;
* &#039;&#039;&#039;Measure the fridge.&#039;&#039;&#039; The keg fridge is actually 7–8 °C, so 2.4–2.5 volumes needs ~15 PSI, not the textbook 11–12 PSI. #0009 and #0010 were both under-carbonated for this reason. A small 3 L keg also absorbs CO₂ slowly — #0011 was only ~70–80% carbonated after 5 days at pressure; lay it down and shake for 2–3 minutes.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_mini_keg_regulator.jpg|Mini keg with CO₂ regulator: carbonation is set by pressure, not sugar&lt;br /&gt;
File:Homebrew_mini_keg_tap.jpg|Mini keg with a wooden tap handle&lt;br /&gt;
File:Homebrew_bottles_box.jpg|A case of bottles filled from the keg&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
=== Method ===&lt;br /&gt;
* Write the plan down before brewing (one line: dry-hop day and gravity, green light, transfer, FG) and measure the error afterwards. The predictions are converging batch by batch.&lt;br /&gt;
* Log everything with a timestamp — gravity with date and time, pH with temperature. Most of the lessons above only became visible once the numbers were next to each other.&lt;br /&gt;
* Change one or two things per version. The recipe history (Crazy Concubine v1→v4.4, Orange Glow v1→v2.6) reads like a commit log: each version cites the batch that justified it.&lt;br /&gt;
&lt;br /&gt;
== Equipment ==&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_first_kit.jpg|Brew day #0001 (Dec 2025): kettle, fermenter, grain and hops&lt;br /&gt;
File:Homebrew_hop_pellets.jpg|Hop pellets, straight from the vacuum pack&lt;br /&gt;
File:Homebrew_mash.jpg|Mashing #0008, thermometer in the lid&lt;br /&gt;
File:Homebrew_ir_thermometer.jpg|IR thermometer check on the kettle&lt;br /&gt;
File:Homebrew_cooling_rig.jpg|Home-built cooling rig: thermoelectric cooler, power supply and thermostat, pumping cold water through the fermenter&lt;br /&gt;
File:Homebrew_o2_co2_cylinders.jpg|CO₂ (left) and pure O₂ (right) cylinders with regulators&lt;br /&gt;
File:Homebrew_fermenter_krausen.jpg|#0008 on Day 3: krausen, fermenting in the water bath&lt;br /&gt;
File:Homebrew_hydrometer.jpg|Gravity sample from #0009&lt;br /&gt;
File:Homebrew_rapt_pill.jpg|RAPT Pill wireless hydrometer (Nov 2026), next to its box&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
* Kettle + fermenter in a temperature-controlled water bath; weigh-the-pot volume estimation&lt;br /&gt;
* Pure O₂ cylinder with diffusion stone; refractometer + hydrometer; pH pen (4.01 / 7.00 buffers)&lt;br /&gt;
* Inkbird IHT-1P instant-read thermometer (from #0010)&lt;br /&gt;
* RAPT Pill wireless hydrometer and thermometer (from Nov 2026): floats in the fermenter and logs gravity and temperature over Bluetooth LE or 2.4 GHz Wi-Fi; range 0.95–1.12, rechargeable 18650 battery (USB-C)&lt;br /&gt;
* 8 L keg on external CO₂; 128 oz mini keg for in-keg conditioning&lt;br /&gt;
* Additives: Fermaid-O (2.5 g at pitch + 2.5 g at ~24 h), Maturex L ALDC 0.2–0.3 ml, CaCl₂&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_0012_gravity_vs_days.png&amp;diff=855</id>
		<title>檔案:Homebrew 0012 gravity vs days.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_0012_gravity_vs_days.png&amp;diff=855"/>
		<updated>2026-11-12T06:55:32Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
Batch #0012 RAPT Pill charts, extracted from the brew log (Dark HUD style)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Batch #0012 RAPT Pill charts, extracted from the brew log (Dark HUD style)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_0012_pill_rate.png&amp;diff=854</id>
		<title>檔案:Homebrew 0012 pill rate.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_0012_pill_rate.png&amp;diff=854"/>
		<updated>2026-11-12T06:55:32Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
Batch #0012 RAPT Pill charts, extracted from the brew log (Dark HUD style)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Batch #0012 RAPT Pill charts, extracted from the brew log (Dark HUD style)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_0012_pill_telemetry.png&amp;diff=853</id>
		<title>檔案:Homebrew 0012 pill telemetry.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_0012_pill_telemetry.png&amp;diff=853"/>
		<updated>2026-11-12T06:55:32Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
Batch #0012 RAPT Pill charts, extracted from the brew log (Dark HUD style)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Batch #0012 RAPT Pill charts, extracted from the brew log (Dark HUD style)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=852</id>
		<title>Homebrewing</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=852"/>
		<updated>2026-11-09T06:22:57Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Equipment: RAPT Pill wireless hydrometer (photo + note)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&#039;&#039;&#039;Homebrewing&#039;&#039;&#039; — eleven batches of hazy / New England IPA brewed at home in 10-litre batches, December 2025 to October 2026. Every batch is logged with gravity readings, pH, hop timing and a tasting note, and each recipe is version-controlled in git so every change can be traced back to the batch that caused it. This page summarises what was brewed, what went wrong, and the process rules that came out of it.&lt;br /&gt;
&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_glass_orange_glow.jpg|Orange Glow v2.3 (#0007), five weeks after brew day&lt;br /&gt;
File:Homebrew_glass_in_hand.jpg|A finished hazy IPA, in the glass with the piz brewing logo&lt;br /&gt;
File:Homebrew_keg_and_bottles.jpg|Keg and bottles, with the Strange Haze label in front&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Batches !! Period !! Batch size !! Styles !! Recipes (current version)&lt;br /&gt;
|-&lt;br /&gt;
| 11 (12 brews) || 2025-12 → 2026-10 || 6–11.5 L || Hazy IPA, DDH NEIPA, Milkshake IPA || 妃子起笑 Crazy Concubine (v4.4), Orange Glow (v2.6), Strange Haze (v1.1), Milkshake IPA (v1)&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The data ==&lt;br /&gt;
[[File:Homebrew_og_fg_by_batch.png|900px|center|Original and final gravity of every batch]]&lt;br /&gt;
The first six batches had no written recipe and wild original gravities (up to 1.110 — a 10.5% &amp;quot;IPA&amp;quot;). From #0007 on, every batch follows a versioned recipe and the OG has been pulled back into the 1.050–1.076 range the recipes are designed for.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_abv_attenuation.png|900px|center|ABV and apparent attenuation per batch]]&lt;br /&gt;
Attenuation went from ~71% (early batches, under-pitched, no oxygen) to a consistent 81–88% once pure-oxygen pitching, yeast nutrient and modern yeast strains were adopted. #0011&#039;s 62% is not a stuck fermentation: it is a kit recipe with 50 g of unfermentable lactose, and it landed exactly on the kit&#039;s printed OG 1.052 / FG 1.020.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_gravity_curves.png|900px|center|Fermentation curves of four batches]]&lt;br /&gt;
The single most useful chart. With one pack of Verdant and no oxygen (#0006) fermentation drags on for 12 days. With LalBrew Pomona, or two packs plus pure oxygen, ~90% of the sugar is gone within &#039;&#039;&#039;2 days&#039;&#039;&#039;. Any schedule written as &amp;quot;dry hop on Day 3&amp;quot; is simply wrong for these yeasts.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_plan_vs_actual.png|800px|center|Planned vs actual timing of key events]]&lt;br /&gt;
Since #0009 each batch starts with a one-line written plan (dry-hop day and gravity, green-light day, transfer day, expected FG). The log then measures the error. The first two plans were badly off — fermentation finished 8–12 days early — and each new plan is corrected from the previous error. #0011&#039;s DH1, DH2 and FG all landed within a day of plan (1.0, 0.8 and 0.8 days early).&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_ph.png|900px|center|Mash pH and finished-beer pH]]&lt;br /&gt;
pH turned out to be the hidden cause of &amp;quot;too bitter&amp;quot; — and, in #0011, of &amp;quot;slightly sour&amp;quot;. The same tap water gave a mash pH of 4.76 in one batch and 5.81 in another, so the correction has to go in either direction. #0011&#039;s mash was corrected to 5.37, but the wort measured 4.83 after the boil and was not corrected again; the beer finished at 3.94. In both #0008 (4.75 → 3.83) and #0011 (4.83 → 3.94) pH fell about &#039;&#039;&#039;0.9&#039;&#039;&#039; during fermentation, not the textbook 0.3–0.5 — this soft, low-alkalinity water has almost no buffer.&lt;br /&gt;
&lt;br /&gt;
== Batch log ==&lt;br /&gt;
{| class=&amp;quot;wikitable sortable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! # !! Date !! Beer / version !! Yeast !! OG → FG !! ABV !! Result / key lesson&lt;br /&gt;
|-&lt;br /&gt;
| 0001 || 2025-12-03 || DDH Hazy IPA || LalBrew New England || 1.062 → 1.018 || 5.8% || Bottle-conditioned. Lesson: add priming sugar, don&#039;t refrigerate during bottle conditioning.&lt;br /&gt;
|-&lt;br /&gt;
| 0002 || 2025-12-28 || Hazy IPA || Mauribrew Weiss || 1.040 → ? || – || Poor mash efficiency; dry-hop order wrong; forgot to measure FG before bottling.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-1 || 2026-02-26 || IPA || Mauribrew ALE514 || 1.050 → 1.010 || 5.3% || Flat aroma and taste; too much priming sugar, vegetal smell.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-2 || 2026-02-28 || IPA + 900 g oats || ALE514 || 1.056 → 1.016 || 5.3% || Under-pitched; bottled while airlock still active → gushers on opening.&lt;br /&gt;
|-&lt;br /&gt;
| 0004 || 2026-03-15 || Milkshake IPA || LalBrew New England || 1.110 → 1.030 || 10.5% || Lid not sealed (airlock silent 36 h); bottled while still fermenting; weak carbonation, slightly sour.&lt;br /&gt;
|-&lt;br /&gt;
| 0005 || 2026-04-23 || Milkshake IPA &amp;quot;light&amp;quot; || Verdant IPA || 1.102 → 1.030 || ~9.5% || &#039;&#039;&#039;Failed.&#039;&#039;&#039; Bittering hops 22 g instead of 6 g; bottles never carbonated — at ~9.5% ABV the yeast was too stressed to referment, and the cold crash removed what was left. Most poured away.&lt;br /&gt;
|-&lt;br /&gt;
| 0006 || 2026-05-09 || Hazy IPA || Verdant IPA || 1.062 → 1.008 || 7.1% || Second dry hop at Day 9 triggered &#039;&#039;&#039;hop creep&#039;&#039;&#039; (1.014 → 1.008). Clean, bitterness fine, aroma weak. First keg + forced carbonation.&lt;br /&gt;
|-&lt;br /&gt;
| 0007 || 2026-06-12 || Orange Glow v2.3 || Verdant IPA ||1.080 → 1.015 || 8.5% || First versioned recipe. Not hot, bitterness right after 3 weeks, but aroma lacking and colour unappealing.&lt;br /&gt;
|-&lt;br /&gt;
| 0008 || 2026-06-25 || Orange Glow v2.4 || LalBrew Pomona || 1.090 → 1.016 || 9.7% || Pomona hit 1.016 by Day 3 — dry hop missed its window. Finished pH 3.83 → harsh bitterness drowned the juice.&lt;br /&gt;
|-&lt;br /&gt;
| 0009 || 2026-07-29 || Orange Glow v2.5 &amp;quot;clean&amp;quot; || Pomona || 1.050 → 1.006 || 5.8% || Boiled off 4–5 L, OG shot to 1.110, over-diluted back to 1.050. pH 4.30 — bitterness finally right. Thin body, under-carbonated (keg fridge is 7–8 °C, not 3 °C). Passable daily hazy.&lt;br /&gt;
|-&lt;br /&gt;
| 0010 || 2026-08-26 || Strange Haze v1 (blueberry vanilla) || Verdant ×2 || 1.070 → 1.012 || 7.6% || Pre-boil OG check caught an over-concentrated wort in time. No harshness, no diacetyl — but not enough juice, sweetness, purple colour or vanilla; grainy from 20% oats.&lt;br /&gt;
|-&lt;br /&gt;
| 0011 || 2026-10-08 || 妃子起笑 v4.3 (PB Craft kit) || Weiss || 1.052 → 1.020 || 4.2% || Clean process — no diacetyl, oxidation or harsh bitterness, no bottle bombs — but weak aroma, thin &amp;quot;watery&amp;quot; body, pale (3–4 SRM) and slightly sour (pH 3.94). Every flaw traced to three fixable numbers: post-boil pH not corrected, CaCl₂ only 0.5 g, both dry hops added while still fermenting. Led to v4.4 and to recipe-wide changes.&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The beers ==&lt;br /&gt;
; 妃子起笑 · Crazy Concubine — Hazy Pale Ale (OG 1.052, ~5%)&lt;br /&gt;
: The first recipe. v1&#039;s second dry hop went in after fermentation had finished and hop creep ate the body (FG 1.014 → 1.008, 5.8% instead of 5%). v3 fixed it with zero new cost — one dry hop during active fermentation, cooler mash, pure-oxygen pitch, no cold crash. v4 added a 78 °C whirlpool for front-end juiciness. #0011 brewed the PB Craft kit exactly as printed (v4.3) and hit the kit&#039;s numbers, which showed the kit itself is the limit. v4.4: 10 L instead of 11.5 L (OG back to 1.060), CaCl₂ 3 g, a post-boil pH check, and the dry hops swapped — Idaho Gem first, Citra + Amarillo last. #0002, the same kit, had used that &amp;quot;wrong&amp;quot; order by mistake — and had the lychee aroma #0011 lacked.&lt;br /&gt;
; Orange Glow — DDH NEIPA (OG ~1.066, ~7%)&lt;br /&gt;
: The main development line: v1 → v2.6 across three brews. Swapped glucose for DME, Extra Pale for Golden Promise, Verdant for Pomona, added ALDC and Fermaid-O, moved to a split dry hop. #0009 drifted to a pale 5.8%; v2.6 brings back Munich for the orange-gold colour and uses a Strata / Mosaic / Nelson / Peacharine hop bill.&lt;br /&gt;
; Strange Haze — Imperial Milkshake IPA (blueberry vanilla)&lt;br /&gt;
: Built around a pre-printed label (9% ABV, purple nebula), so the fruit is blueberry for colour. v1 was &amp;quot;neither here nor there&amp;quot;; v1.1 cuts bitterness, adds lactose back, raises mash to 69–70 °C and splits 800 g of blueberry between mid-fermentation and post-fermentation.&lt;br /&gt;
; Milkshake IPA — low residual sugar (6 L)&lt;br /&gt;
: Body from oat/wheat protein instead of lots of lactose; unsweetened mango + passion fruit purée. Not yet brewed under this recipe.&lt;br /&gt;
&lt;br /&gt;
== Label and logo ==&lt;br /&gt;
The beers go out under the name &#039;&#039;&#039;piz brewing&#039;&#039;&#039;. The logo started from our dog, became a pencil sketch of just the ears and nose, and was finished with AI image generation. The labels were made the same way: many drafts generated and refined until one fit the beer.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_logo_dog.jpg|The original&lt;br /&gt;
File:Homebrew_logo_sketch.jpg|Pencil sketch: ears and nose&lt;br /&gt;
File:Homebrew_logo_final.jpg|Final logo&lt;br /&gt;
File:Homebrew_label_strange_haze.jpg|Strange Haze label, made with AI image generation&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== What I learned ==&lt;br /&gt;
=== Fermentation ===&lt;br /&gt;
* &#039;&#039;&#039;Dry hop by gravity, not by the calendar.&#039;&#039;&#039; The first dry hop goes in when gravity has dropped about half-way from OG to FG, while yeast is still active (biotransformation + it mops up hop-creep sugar). With fast yeast that can be Day 1, so start measuring every 12 h from Day 1.&lt;br /&gt;
* &#039;&#039;&#039;Split dry hop.&#039;&#039;&#039; DH1 during active fermentation for biotransformation; DH2 only after fermentation has stopped, so the CO₂ doesn&#039;t scrub the aroma away. After DH2, confirm gravity is stable again (second wave of hop creep). #0011 put both dry hops in while it was still fermenting, and the empty glass smelled only of malt.&lt;br /&gt;
* &#039;&#039;&#039;The last dry hop is the aroma you smell.&#039;&#039;&#039; It soaks latest, is scrubbed least and has the least time to oxidise, so put the signature hop (e.g. Citra) there, loose rather than in a bag. Only bag the dry hop you plan to pull out.&lt;br /&gt;
* &#039;&#039;&#039;Hop creep is real.&#039;&#039;&#039; Hop enzymes break down dextrins, so a late dry hop restarts fermentation: thinner body, higher ABV, and in bottles — gushers or bombs.&lt;br /&gt;
* &#039;&#039;&#039;&amp;quot;Green light&amp;quot; before packaging:&#039;&#039;&#039; two gravity readings ~48 h apart that differ by ≤ 0.001, &#039;&#039;&#039;and&#039;&#039;&#039; at least 3 days of diacetyl rest at a beer temperature of 20–22 °C (water bath 21.5 °C). Since 2026-09 the rest time replaces the forced diacetyl test: below 18 °C the yeast barely cleans up diacetyl, above 24 °C it only adds fusel alcohols, so 3 days at 20–22 °C is the guarantee. After DH2 or fruit, hop creep can make new diacetyl, so hold 20–22 °C for another 48 h and re-check gravity before cold-crashing. #0011 was packaged before creep had stopped and still had no butter flavour. If a beer does taste buttery, warm the keg back to room temperature for 3 days. Never package on a fixed day count — #0003-2, #0004 and #0008 all did, and all had carbonation problems.&lt;br /&gt;
* &#039;&#039;&#039;Pure oxygen 60–90 s before pitching&#039;&#039;&#039; lets a single dry pack ferment a 1.07+ wort; no starter needed. After pitching, oxygen is the enemy.&lt;br /&gt;
* Fermenting beer runs 1–2 °C above the water bath during the first 48 h. Set the bath ~1.5 °C below the target beer temperature at peak, then raise it to 21.5 °C for the diacetyl rest.&lt;br /&gt;
&lt;br /&gt;
=== Brew day ===&lt;br /&gt;
* &#039;&#039;&#039;Check pre-boil gravity and volume before boiling.&#039;&#039;&#039; Weigh the pot (total − empty pot weight ≈ litres), measure pre-boil gravity, and compute the target end volume = pre-boil points × litres ÷ target points. Boil to that volume, not to a timer. #0009 boiled off 4–5 L (normal is 1–2 L) and hit 1.110.&lt;br /&gt;
* &#039;&#039;&#039;Boil gently, lid off.&#039;&#039;&#039; Hard boiling over-concentrates bitterness and scorches. Excess evaporation turned out to be a big part of the &amp;quot;nasty bitterness&amp;quot; in #0007/#0008.&lt;br /&gt;
* &#039;&#039;&#039;Measure mash pH, not water pH, and be ready to go both ways.&#039;&#039;&#039; Sample 10–15 min after mash-in, cool to 20–25 °C, target 5.2–5.4. Keep both baking soda and phosphoric/lactic acid on hand. This soft, low-alkalinity water swings hard: ~1 g baking soda ≈ +0.67, 1 ml phosphoric ≈ −0.28 — use half steps and re-measure. Calibrate the pH pen every brew day and store the electrode wet.&lt;br /&gt;
* &#039;&#039;&#039;Post-boil pH is a second, separate checkpoint.&#039;&#039;&#039; A long boil concentrates the acidity too: #0011 boiled 14.55 L down to 11.5 L and dropped from 5.37 to 4.83. Correct the cooled wort to 5.2–5.4 before pitching; with a ~0.9 drop in fermentation that lands the beer at ~4.4. (#0010 split the same 2 g of baking soda between mash and post-boil and had no sourness.)&lt;br /&gt;
* &#039;&#039;&#039;Salt: CaCl₂·2H₂O 0.2 g per litre of total brewing water&#039;&#039;&#039; (≈ Ca 54 / Cl 96 ppm), no gypsum. The filtered water has almost no minerals; #0011&#039;s 0.5 g gave ~17 ppm chloride and ~9 ppm calcium, and the beer tasted like &amp;quot;sweet water&amp;quot;.&lt;br /&gt;
* Pale-malt grists (Extra Pale/Pilsner) tend to mash high (~5.7); grists with Munich plus gypsum tend to mash low.&lt;br /&gt;
&lt;br /&gt;
=== Bitterness and aroma ===&lt;br /&gt;
* &#039;&#039;&#039;Fix bitterness at the source, never by cutting hops.&#039;&#039;&#039; The culprits were pH (finished pH 3.83 = throat-biting), gypsum (sulfate sharpens bitterness), over-boiling, hops left in too long, and drinking too young. Bittering hops go in a bag, on time, and come out after the boil; dry-hop bags come out after 2–3 days.&lt;br /&gt;
* &#039;&#039;&#039;Whirlpool and dry hop do different jobs.&#039;&#039;&#039; Whirlpool (75–78 °C) gives juice on the palate; dry hop gives the aroma in the nose. Cutting either one loses something the other can&#039;t replace.&lt;br /&gt;
* Aging softens bitterness but kills hop aroma at the same time — fix bitterness at the source and drink fresh (2–3 weeks) instead.&lt;br /&gt;
* 0.5 kg oats + 1 kg wheat in 4.5 kg of grain is already enough haze; 1 kg of oats (20%) tasted grainy.&lt;br /&gt;
* &#039;&#039;&#039;Lactose: go big or skip it.&#039;&#039;&#039; 50 g in 11.5 L (4.3 g/L) is far below the ~20–25 g/L you can taste — it added nothing.&lt;br /&gt;
* &#039;&#039;&#039;A kit guarantees beer, not good beer.&#039;&#039;&#039; The PB Craft kit has no whirlpool hops, no specialty malt, and says nothing about salt or pH. Brewed exactly as printed (#0011), it gives a pale, thin, low-aroma beer; those four things have to be added.&lt;br /&gt;
&lt;br /&gt;
=== Packaging ===&lt;br /&gt;
* &#039;&#039;&#039;Keg + forced carbonation&#039;&#039;&#039; is now the default (#0009 on). Bottle conditioning kept failing: too much sugar (#0003-1), bottling before fermentation finished (#0003-2, #0004), and #0005 at ~9.5% ABV, where the yeast could not referment.&lt;br /&gt;
* &#039;&#039;&#039;Cold crash is now standard&#039;&#039;&#039; (since #0011): 2–4 °C for 24–48 h before transfer. #0006–#0011 never really cold-crashed, and all six got the same notes — weak aroma, &amp;quot;don&#039;t know what I&#039;m drinking&amp;quot;, grainy. Suspended yeast fits that picture; it is still a hypothesis, to be tested by drinking #0011&#039;s keg (no yeast) next to its bottles (live yeast). For bottle conditioning after a cold crash: under 7% ABV enough yeast remains; above 8%, add 1–2 g of an alcohol-tolerant yeast (e.g. EC-1118) at bottling.&lt;br /&gt;
* &#039;&#039;&#039;Don&#039;t package on a falling gravity with sulfite in the keg.&#039;&#039;&#039; #0011 was packaged at 1.020 while still dropping: no metabisulfite for an early package (keep the yeast alive), and halve the priming sugar for bottles (4.5 g/L — it carbonated without gushing).&lt;br /&gt;
* &#039;&#039;&#039;Never bottle the gravity samples.&#039;&#039;&#039; #0011&#039;s test bottle made from saved samples gushed and measured pH 3.5 — lactic infection from the open sample cup, not the fermenter (the normal bottles were 3.94). Drink samples on the spot; for a control bottle, fill one more from the main beer.&lt;br /&gt;
* Low-oxygen transfer: purge the keg with CO₂ three times, fill from the bottom through the liquid post, minimum headspace.&lt;br /&gt;
* Antioxidant in the keg: &#039;&#039;&#039;ascorbic acid 0.3 g + potassium metabisulfite 0.25 g per 10 L, always as a pair.&#039;&#039;&#039; Ascorbic acid alone makes H₂O₂, which with trace iron/copper (Fenton reaction) oxidises the beer worse than nothing. The chemicals are ~20% of the protection; the purging is the other 80%.&lt;br /&gt;
* &#039;&#039;&#039;Measure the fridge.&#039;&#039;&#039; The keg fridge is actually 7–8 °C, so 2.4–2.5 volumes needs ~15 PSI, not the textbook 11–12 PSI. #0009 and #0010 were both under-carbonated for this reason. A small 3 L keg also absorbs CO₂ slowly — #0011 was only ~70–80% carbonated after 5 days at pressure; lay it down and shake for 2–3 minutes.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_mini_keg_regulator.jpg|Mini keg with CO₂ regulator: carbonation is set by pressure, not sugar&lt;br /&gt;
File:Homebrew_mini_keg_tap.jpg|Mini keg with a wooden tap handle&lt;br /&gt;
File:Homebrew_bottles_box.jpg|A case of bottles filled from the keg&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
=== Method ===&lt;br /&gt;
* Write the plan down before brewing (one line: dry-hop day and gravity, green light, transfer, FG) and measure the error afterwards. The predictions are converging batch by batch.&lt;br /&gt;
* Log everything with a timestamp — gravity with date and time, pH with temperature. Most of the lessons above only became visible once the numbers were next to each other.&lt;br /&gt;
* Change one or two things per version. The recipe history (Crazy Concubine v1→v4.4, Orange Glow v1→v2.6) reads like a commit log: each version cites the batch that justified it.&lt;br /&gt;
&lt;br /&gt;
== Equipment ==&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_first_kit.jpg|Brew day #0001 (Dec 2025): kettle, fermenter, grain and hops&lt;br /&gt;
File:Homebrew_hop_pellets.jpg|Hop pellets, straight from the vacuum pack&lt;br /&gt;
File:Homebrew_mash.jpg|Mashing #0008, thermometer in the lid&lt;br /&gt;
File:Homebrew_ir_thermometer.jpg|IR thermometer check on the kettle&lt;br /&gt;
File:Homebrew_cooling_rig.jpg|Home-built cooling rig: thermoelectric cooler, power supply and thermostat, pumping cold water through the fermenter&lt;br /&gt;
File:Homebrew_o2_co2_cylinders.jpg|CO₂ (left) and pure O₂ (right) cylinders with regulators&lt;br /&gt;
File:Homebrew_fermenter_krausen.jpg|#0008 on Day 3: krausen, fermenting in the water bath&lt;br /&gt;
File:Homebrew_hydrometer.jpg|Gravity sample from #0009&lt;br /&gt;
File:Homebrew_rapt_pill.jpg|RAPT Pill wireless hydrometer (Nov 2026), next to its box&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
* Kettle + fermenter in a temperature-controlled water bath; weigh-the-pot volume estimation&lt;br /&gt;
* Pure O₂ cylinder with diffusion stone; refractometer + hydrometer; pH pen (4.01 / 7.00 buffers)&lt;br /&gt;
* Inkbird IHT-1P instant-read thermometer (from #0010)&lt;br /&gt;
* RAPT Pill wireless hydrometer and thermometer (from Nov 2026): floats in the fermenter and logs gravity and temperature over Bluetooth LE or 2.4 GHz Wi-Fi; range 0.95–1.12, rechargeable 18650 battery (USB-C)&lt;br /&gt;
* 8 L keg on external CO₂; 128 oz mini keg for in-keg conditioning&lt;br /&gt;
* Additives: Fermaid-O (2.5 g at pitch + 2.5 g at ~24 h), Maturex L ALDC 0.2–0.3 ml, CaCl₂&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_rapt_pill.jpg&amp;diff=851</id>
		<title>檔案:Homebrew rapt pill.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_rapt_pill.jpg&amp;diff=851"/>
		<updated>2026-11-09T06:22:56Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
RAPT Pill wireless hydrometer (resized, metadata stripped)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
RAPT Pill wireless hydrometer (resized, metadata stripped)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=MediaWiki:Common.css&amp;diff=850</id>
		<title>MediaWiki:Common.css</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=MediaWiki:Common.css&amp;diff=850"/>
		<updated>2026-10-28T13:45:42Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Sci-fi dark HUD theme for Timeless: dark palette, cyan/amber accents, HUD headings, dark pygments tokens, print stays light&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;/* ============================================================================&lt;br /&gt;
   &amp;quot;flip the world&amp;quot; — DARK HUD sci-fi theme&lt;br /&gt;
   Target skin: Timeless (site default). MediaWiki 1.43.&lt;br /&gt;
   Loaded from MediaWiki:Common.css, so it applies to every skin and every&lt;br /&gt;
   namespace; the structural selectors below are Timeless-specific and are&lt;br /&gt;
   simply inert under other skins.&lt;br /&gt;
&lt;br /&gt;
   Palette:  bg #0a0f1a / panel #0e1524 / text #c8d6e5&lt;br /&gt;
             cyan #35e0ff (structure, links, headings)&lt;br /&gt;
             amber #ffb340 (markers, emphasis, keywords)&lt;br /&gt;
             mint #7cf6c8 (strings, positive)&lt;br /&gt;
   No animation, no scanlines: this wiki is meant to be read.&lt;br /&gt;
   ============================================================================ */&lt;br /&gt;
&lt;br /&gt;
:root {&lt;br /&gt;
	--sf-bg: #0a0f1a;&lt;br /&gt;
	--sf-bg-deep: #060b14;&lt;br /&gt;
	--sf-panel: #0e1524;&lt;br /&gt;
	--sf-panel-2: #111c2e;&lt;br /&gt;
	--sf-panel-3: #16233a;&lt;br /&gt;
	--sf-fg: #c8d6e5;&lt;br /&gt;
	--sf-fg-dim: #8a9cb2;&lt;br /&gt;
	--sf-fg-faint: #63758c;&lt;br /&gt;
	--sf-cyan: #35e0ff;&lt;br /&gt;
	--sf-cyan-2: #7ceaff;&lt;br /&gt;
	--sf-amber: #ffb340;&lt;br /&gt;
	--sf-mint: #7cf6c8;&lt;br /&gt;
	--sf-red: #ff5f6d;&lt;br /&gt;
	--sf-violet: #c9a6ff;&lt;br /&gt;
	--sf-line: rgba( 53, 224, 255, 0.22 );&lt;br /&gt;
	--sf-line-soft: rgba( 53, 224, 255, 0.09 );&lt;br /&gt;
	--sf-glow: 0 0 8px rgba( 53, 224, 255, 0.3 );&lt;br /&gt;
	--sf-mono: &#039;JetBrains Mono&#039;, &#039;SFMono-Regular&#039;, &#039;Menlo&#039;, &#039;Consolas&#039;, &#039;Noto Sans Mono&#039;, &#039;DejaVu Sans Mono&#039;, monospace;&lt;br /&gt;
	--sf-display: &#039;Eurostile&#039;, &#039;Bahnschrift&#039;, &#039;DIN Alternate&#039;, &#039;Segoe UI Semibold&#039;, &#039;Noto Sans&#039;, sans-serif;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Native widgets, scrollbars and form controls follow the theme. */&lt;br /&gt;
html {&lt;br /&gt;
	color-scheme: dark;&lt;br /&gt;
	background-color: var(--sf-bg-deep);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
body {&lt;br /&gt;
	background-color: var(--sf-bg-deep);&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
::selection {&lt;br /&gt;
	background: rgba( 53, 224, 255, 0.3 );&lt;br /&gt;
	color: #fff;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ---------------------------------------------------------------- Header -- */&lt;br /&gt;
&lt;br /&gt;
#mw-header-container {&lt;br /&gt;
	background: linear-gradient( 180deg, #0b1322 0%, var(--sf-bg) 100% );&lt;br /&gt;
	border-bottom: 1px solid var(--sf-line);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#mw-header-container.ts-container {&lt;br /&gt;
	box-shadow: 0 1px 12px rgba( 0, 0, 0, 0.6 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Site title, top left. */&lt;br /&gt;
#p-logo-text a {&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
	font-family: var(--sf-display);&lt;br /&gt;
	font-weight: 600;&lt;br /&gt;
	letter-spacing: 0.07em;&lt;br /&gt;
	text-transform: uppercase;&lt;br /&gt;
	white-space: nowrap;&lt;br /&gt;
	text-shadow: var(--sf-glow);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#p-logo-text a:hover {&lt;br /&gt;
	color: #fff;&lt;br /&gt;
	text-decoration: none;&lt;br /&gt;
	text-shadow: 0 0 10px rgba( 53, 224, 255, 0.7 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* The Maker Faire robot is flat red art on a white square. Inverting it turns&lt;br /&gt;
   the white to black and the red to cyan; `lighten` then drops the black out&lt;br /&gt;
   against the dark header, so it reads as a cyan HUD glyph with no re-upload. */&lt;br /&gt;
#p-logo img,&lt;br /&gt;
#p-logo a.mw-wiki-logo,&lt;br /&gt;
.mw-wiki-logo.timeless-logo {&lt;br /&gt;
	max-width: 100%;&lt;br /&gt;
	max-height: 80px;&lt;br /&gt;
	width: auto;&lt;br /&gt;
	height: auto;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Apply the inversion to the image only — putting it on the &amp;lt;a&amp;gt; as well would&lt;br /&gt;
   invert twice and hand the red robot back. */&lt;br /&gt;
#p-logo img {&lt;br /&gt;
	filter: invert( 1 ) saturate( 1.35 );&lt;br /&gt;
	mix-blend-mode: lighten;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#p-logo {&lt;br /&gt;
	padding: 0.5em 0;&lt;br /&gt;
	overflow: hidden;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Accent stripe under the header: red/blue/green -&amp;gt; cyan/amber HUD bar. */&lt;br /&gt;
.color-left {&lt;br /&gt;
	background: linear-gradient( 90deg, rgba( 53, 224, 255, 0 ) 0%, var(--sf-cyan) 100% );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.color-right {&lt;br /&gt;
	background: linear-gradient( 90deg, var(--sf-amber) 0%, rgba( 255, 179, 64, 0 ) 100% );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.color-middle {&lt;br /&gt;
	background: var(--sf-cyan);&lt;br /&gt;
	box-shadow: 0 0 10px rgba( 53, 224, 255, 0.5 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ---------------------------------------------------------------- Search -- */&lt;br /&gt;
&lt;br /&gt;
#simpleSearch {&lt;br /&gt;
	background: var(--sf-bg-deep);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
	box-shadow: inset 0 0 12px rgba( 53, 224, 255, 0.06 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#simpleSearch:focus-within {&lt;br /&gt;
	border-color: var(--sf-cyan);&lt;br /&gt;
	box-shadow: 0 0 0 1px rgba( 53, 224, 255, 0.35 ), inset 0 0 12px rgba( 53, 224, 255, 0.1 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#searchInput {&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
	font-family: var(--sf-mono);&lt;br /&gt;
	font-size: 0.9em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#searchInput::placeholder {&lt;br /&gt;
	color: var(--sf-fg-faint);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Skin search glyph is a dark grey SVG; lift it onto the dark field. */&lt;br /&gt;
#searchButton,&lt;br /&gt;
#mw-searchButton {&lt;br /&gt;
	filter: invert( 1 ) brightness( 1.1 );&lt;br /&gt;
	opacity: 0.75;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.suggestions .suggestions-results,&lt;br /&gt;
.suggestions .suggestions-special {&lt;br /&gt;
	background: var(--sf-panel-2);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.suggestions .suggestions-result {&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.suggestions .suggestions-result-current,&lt;br /&gt;
.suggestions .suggestions-special.suggestions-result-current {&lt;br /&gt;
	background: var(--sf-panel-3);&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* --------------------------------------------------- Page / content shell -- */&lt;br /&gt;
&lt;br /&gt;
#mw-content-container {&lt;br /&gt;
	background-color: var(--sf-bg);&lt;br /&gt;
	/* Skin ships a decorative cat as a backdrop; replace with a soft glow. */&lt;br /&gt;
	background-image: radial-gradient( 900px 420px at 50% -120px, rgba( 53, 224, 255, 0.08 ), transparent 70% );&lt;br /&gt;
	background-repeat: no-repeat;&lt;br /&gt;
	background-position: center top;&lt;br /&gt;
	border-bottom: 1px solid var(--sf-cyan);&lt;br /&gt;
	box-shadow: inset 0 -1px 14px rgba( 53, 224, 255, 0.18 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* The main reading panel. */&lt;br /&gt;
#mw-content {&lt;br /&gt;
	background: var(--sf-panel);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
	box-shadow: 0 0 0 1px rgba( 0, 0, 0, 0.5 ), 0 8px 30px rgba( 0, 0, 0, 0.45 );&lt;br /&gt;
	position: relative;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Corner bracket, top right — HUD framing without clipping any menus. */&lt;br /&gt;
#mw-content::before {&lt;br /&gt;
	content: &#039;&#039;;&lt;br /&gt;
	position: absolute;&lt;br /&gt;
	top: -1px;&lt;br /&gt;
	right: -1px;&lt;br /&gt;
	width: 22px;&lt;br /&gt;
	height: 22px;&lt;br /&gt;
	border-top: 2px solid var(--sf-cyan);&lt;br /&gt;
	border-right: 2px solid var(--sf-cyan);&lt;br /&gt;
	box-shadow: var(--sf-glow);&lt;br /&gt;
	pointer-events: none;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#mw-content::after {&lt;br /&gt;
	content: &#039;&#039;;&lt;br /&gt;
	position: absolute;&lt;br /&gt;
	bottom: -1px;&lt;br /&gt;
	left: -1px;&lt;br /&gt;
	width: 22px;&lt;br /&gt;
	height: 22px;&lt;br /&gt;
	border-bottom: 2px solid var(--sf-cyan);&lt;br /&gt;
	border-left: 2px solid var(--sf-cyan);&lt;br /&gt;
	box-shadow: var(--sf-glow);&lt;br /&gt;
	pointer-events: none;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ------------------------------------------------------------- Sidebars -- */&lt;br /&gt;
&lt;br /&gt;
#mw-site-navigation .sidebar-chunk,&lt;br /&gt;
#mw-related-navigation .sidebar-chunk {&lt;br /&gt;
	background: var(--sf-panel-2);&lt;br /&gt;
	border: 1px solid var(--sf-line-soft);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
	box-shadow: 0 4px 18px rgba( 0, 0, 0, 0.4 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#mw-site-navigation .sidebar-inner,&lt;br /&gt;
#mw-related-navigation .sidebar-inner {&lt;br /&gt;
	background: transparent;&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.sidebar-chunk .mw-portlet h3,&lt;br /&gt;
#mw-site-navigation h3,&lt;br /&gt;
#mw-related-navigation h3,&lt;br /&gt;
#personal h3 {&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
	font-family: var(--sf-display);&lt;br /&gt;
	font-size: 0.8em;&lt;br /&gt;
	font-weight: 600;&lt;br /&gt;
	letter-spacing: 0.18em;&lt;br /&gt;
	text-transform: uppercase;&lt;br /&gt;
	border-bottom: 1px solid var(--sf-line-soft);&lt;br /&gt;
	padding-bottom: 0.35em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Leading marker on every nav link. */&lt;br /&gt;
.sidebar-chunk .mw-portlet-body li,&lt;br /&gt;
#personal-inner .mw-portlet-body li {&lt;br /&gt;
	list-style: none;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.sidebar-chunk .mw-portlet-body li a::before,&lt;br /&gt;
#personal-inner .mw-portlet-body li a::before {&lt;br /&gt;
	content: &#039;\203A\00a0&#039;;&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
	opacity: 0.55;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.sidebar-chunk .mw-portlet-body li a:hover::before {&lt;br /&gt;
	opacity: 1;&lt;br /&gt;
	color: var(--sf-amber);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Personal tools dropdown (top right). */&lt;br /&gt;
#personal-inner.dropdown,&lt;br /&gt;
#personal .dropdown {&lt;br /&gt;
	background: var(--sf-panel-2);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	box-shadow: 0 6px 24px rgba( 0, 0, 0, 0.6 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ----------------------------------------------------- Page action tabs -- */&lt;br /&gt;
&lt;br /&gt;
#mw-page-header-links .mw-portlet h3,&lt;br /&gt;
#mw-page-header-links h3 {&lt;br /&gt;
	color: var(--sf-fg-faint);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#mw-page-header-links .tools-inline li a {&lt;br /&gt;
	color: var(--sf-fg-dim);&lt;br /&gt;
	font-family: var(--sf-display);&lt;br /&gt;
	font-size: 0.9em;&lt;br /&gt;
	letter-spacing: 0.08em;&lt;br /&gt;
	text-transform: uppercase;&lt;br /&gt;
	border-bottom: 2px solid transparent;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#mw-page-header-links .tools-inline li a:hover {&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
	text-decoration: none;&lt;br /&gt;
	border-bottom-color: var(--sf-line);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#mw-page-header-links .tools-inline li.selected a,&lt;br /&gt;
#mw-page-header-links .tools-inline li.selected a:visited {&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
	border-bottom-color: var(--sf-cyan);&lt;br /&gt;
	text-shadow: var(--sf-glow);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ------------------------------------------------------------ Typography -- */&lt;br /&gt;
&lt;br /&gt;
.mw-body-content {&lt;br /&gt;
	line-height: 1.75;&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#firstHeading,&lt;br /&gt;
.firstHeading {&lt;br /&gt;
	color: #eaf6ff;&lt;br /&gt;
	font-family: var(--sf-display);&lt;br /&gt;
	font-weight: 600;&lt;br /&gt;
	letter-spacing: 0.06em;&lt;br /&gt;
	text-transform: uppercase;&lt;br /&gt;
	border-bottom: 1px solid var(--sf-line);&lt;br /&gt;
	padding-bottom: 0.35em;&lt;br /&gt;
	text-shadow: 0 0 14px rgba( 53, 224, 255, 0.25 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Small callsign above the title. */&lt;br /&gt;
#siteSub {&lt;br /&gt;
	display: block;&lt;br /&gt;
	color: var(--sf-fg-faint);&lt;br /&gt;
	font-family: var(--sf-mono);&lt;br /&gt;
	font-size: 0.72em;&lt;br /&gt;
	letter-spacing: 0.22em;&lt;br /&gt;
	text-transform: uppercase;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content h2,&lt;br /&gt;
.mw-body-content h3,&lt;br /&gt;
.mw-body-content h4 {&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
	font-family: var(--sf-display);&lt;br /&gt;
	font-weight: 600;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content h2 {&lt;br /&gt;
	margin-top: 1.8em;&lt;br /&gt;
	padding: 0 0 0.3em 0.7em;&lt;br /&gt;
	border-bottom: 1px solid var(--sf-line);&lt;br /&gt;
	border-left: 3px solid var(--sf-cyan);&lt;br /&gt;
	letter-spacing: 0.08em;&lt;br /&gt;
	text-transform: uppercase;&lt;br /&gt;
	text-shadow: var(--sf-glow);&lt;br /&gt;
	background: linear-gradient( 90deg, rgba( 53, 224, 255, 0.1 ), transparent 55% );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content h3 {&lt;br /&gt;
	margin-top: 1.4em;&lt;br /&gt;
	padding-left: 0.7em;&lt;br /&gt;
	color: var(--sf-cyan-2);&lt;br /&gt;
	border-left: 2px solid var(--sf-line);&lt;br /&gt;
	letter-spacing: 0.04em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content h4 {&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
	letter-spacing: 0.04em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content b,&lt;br /&gt;
.mw-body-content strong {&lt;br /&gt;
	color: #e6f3ff;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Section [edit] links. */&lt;br /&gt;
.mw-editsection,&lt;br /&gt;
.mw-editsection a {&lt;br /&gt;
	color: var(--sf-fg-faint);&lt;br /&gt;
	font-family: var(--sf-mono);&lt;br /&gt;
	font-size: 0.75em;&lt;br /&gt;
	text-transform: uppercase;&lt;br /&gt;
	letter-spacing: 0.1em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-editsection a:hover {&lt;br /&gt;
	color: var(--sf-amber);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ----------------------------------------------------------------- Links -- */&lt;br /&gt;
&lt;br /&gt;
.mw-body-content a,&lt;br /&gt;
#mw-content a {&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content a:visited,&lt;br /&gt;
#mw-content a:visited {&lt;br /&gt;
	color: #9bb8ff;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content a:hover,&lt;br /&gt;
#mw-content a:hover {&lt;br /&gt;
	color: #fff;&lt;br /&gt;
	text-shadow: var(--sf-glow);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content a.new,&lt;br /&gt;
#mw-content a.new,&lt;br /&gt;
a.new:visited,&lt;br /&gt;
.sidebar-chunk a.new {&lt;br /&gt;
	color: var(--sf-red);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* External links: drop the skin&#039;s blue icon for a HUD arrow. */&lt;br /&gt;
.mw-parser-output a.external {&lt;br /&gt;
	background-image: none;&lt;br /&gt;
	padding-right: 0;&lt;br /&gt;
	border-bottom: 1px dotted var(--sf-line);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-parser-output a.external::after {&lt;br /&gt;
	content: &#039;\2197&#039;;&lt;br /&gt;
	color: var(--sf-amber);&lt;br /&gt;
	font-size: 0.85em;&lt;br /&gt;
	padding-left: 0.15em;&lt;br /&gt;
	vertical-align: super;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Sidebar / footer links. */&lt;br /&gt;
.sidebar-chunk a,&lt;br /&gt;
#personal a,&lt;br /&gt;
#mw-footer a,&lt;br /&gt;
#catlinks a {&lt;br /&gt;
	color: var(--sf-fg-dim);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.sidebar-chunk a:hover,&lt;br /&gt;
#personal a:hover,&lt;br /&gt;
#mw-footer a:hover,&lt;br /&gt;
#catlinks a:hover {&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
	text-decoration: none;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ----------------------------------------------------------------- Lists -- */&lt;br /&gt;
&lt;br /&gt;
.mw-parser-output ul {&lt;br /&gt;
	list-style: none;&lt;br /&gt;
	padding-left: 1.4em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-parser-output ul &amp;gt; li {&lt;br /&gt;
	margin-bottom: 0.4em;&lt;br /&gt;
	position: relative;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-parser-output ul &amp;gt; li::before {&lt;br /&gt;
	content: &#039;\25B8&#039;;&lt;br /&gt;
	position: absolute;&lt;br /&gt;
	left: -1.1em;&lt;br /&gt;
	color: var(--sf-amber);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Nested bullets step down in weight. */&lt;br /&gt;
.mw-parser-output ul ul &amp;gt; li::before {&lt;br /&gt;
	content: &#039;\25AA&#039;;&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
	opacity: 0.7;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-parser-output ol {&lt;br /&gt;
	padding-left: 1.8em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-parser-output ol &amp;gt; li {&lt;br /&gt;
	margin-bottom: 0.4em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-parser-output ol &amp;gt; li::marker {&lt;br /&gt;
	color: var(--sf-amber);&lt;br /&gt;
	font-family: var(--sf-mono);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Galleries and TOC keep their own list semantics — no HUD bullets there. */&lt;br /&gt;
.mw-parser-output ul.gallery &amp;gt; li::before,&lt;br /&gt;
.mw-parser-output #toc ul &amp;gt; li::before,&lt;br /&gt;
.mw-parser-output .toc ul &amp;gt; li::before {&lt;br /&gt;
	content: none;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-parser-output ul.gallery,&lt;br /&gt;
.mw-parser-output #toc ul {&lt;br /&gt;
	padding-left: 0;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ------------------------------------------------------------------ Code -- */&lt;br /&gt;
&lt;br /&gt;
.mw-body-content pre,&lt;br /&gt;
.mw-highlight pre,&lt;br /&gt;
.mw-parser-output pre {&lt;br /&gt;
	background: var(--sf-bg-deep);&lt;br /&gt;
	color: #dbe7f3;&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-left: 3px solid var(--sf-cyan);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
	padding: 0.9em 1.1em;&lt;br /&gt;
	overflow-x: auto;&lt;br /&gt;
	font-family: var(--sf-mono);&lt;br /&gt;
	font-size: 0.9em;&lt;br /&gt;
	line-height: 1.55;&lt;br /&gt;
	box-shadow: inset 0 0 24px rgba( 53, 224, 255, 0.05 );&lt;br /&gt;
	/* Cut corner, bottom right. */&lt;br /&gt;
	clip-path: polygon( 0 0, 100% 0, 100% calc( 100% - 12px ), calc( 100% - 12px ) 100%, 0 100% );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content code,&lt;br /&gt;
.mw-parser-output code,&lt;br /&gt;
.mw-body-content tt {&lt;br /&gt;
	background: var(--sf-bg-deep);&lt;br /&gt;
	color: var(--sf-mint);&lt;br /&gt;
	border: 1px solid var(--sf-line-soft);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
	padding: 0.05em 0.4em;&lt;br /&gt;
	font-family: var(--sf-mono);&lt;br /&gt;
	font-size: 0.9em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content pre code {&lt;br /&gt;
	background: transparent;&lt;br /&gt;
	border: 0;&lt;br /&gt;
	padding: 0;&lt;br /&gt;
	color: inherit;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-highlight,&lt;br /&gt;
.mw-highlight pre {&lt;br /&gt;
	background: var(--sf-bg-deep);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Pygments ships a light theme (dark green keywords, dark red strings) which is&lt;br /&gt;
   unreadable on this background. Remap the tokens the wiki actually uses. */&lt;br /&gt;
.mw-highlight .hll { background-color: rgba( 53, 224, 255, 0.12 ); }&lt;br /&gt;
.mw-highlight .c,&lt;br /&gt;
.mw-highlight .ch,&lt;br /&gt;
.mw-highlight .cm,&lt;br /&gt;
.mw-highlight .c1,&lt;br /&gt;
.mw-highlight .cs,&lt;br /&gt;
.mw-highlight .cpf { color: #5f7f95; font-style: italic; }&lt;br /&gt;
.mw-highlight .cp { color: var(--sf-amber); }&lt;br /&gt;
.mw-highlight .k,&lt;br /&gt;
.mw-highlight .kc,&lt;br /&gt;
.mw-highlight .kd,&lt;br /&gt;
.mw-highlight .kn,&lt;br /&gt;
.mw-highlight .kr { color: var(--sf-amber); font-weight: bold; }&lt;br /&gt;
.mw-highlight .kp { color: var(--sf-amber); font-weight: normal; }&lt;br /&gt;
.mw-highlight .kt { color: #ffd479; font-weight: normal; }&lt;br /&gt;
.mw-highlight .s,&lt;br /&gt;
.mw-highlight .s1,&lt;br /&gt;
.mw-highlight .s2,&lt;br /&gt;
.mw-highlight .sa,&lt;br /&gt;
.mw-highlight .sb,&lt;br /&gt;
.mw-highlight .sc,&lt;br /&gt;
.mw-highlight .dl,&lt;br /&gt;
.mw-highlight .sd,&lt;br /&gt;
.mw-highlight .se,&lt;br /&gt;
.mw-highlight .sh,&lt;br /&gt;
.mw-highlight .si,&lt;br /&gt;
.mw-highlight .sx,&lt;br /&gt;
.mw-highlight .sr,&lt;br /&gt;
.mw-highlight .ss { color: var(--sf-mint); }&lt;br /&gt;
.mw-highlight .m,&lt;br /&gt;
.mw-highlight .mb,&lt;br /&gt;
.mw-highlight .mf,&lt;br /&gt;
.mw-highlight .mh,&lt;br /&gt;
.mw-highlight .mi,&lt;br /&gt;
.mw-highlight .mo,&lt;br /&gt;
.mw-highlight .il { color: #ff9d7a; }&lt;br /&gt;
.mw-highlight .nb,&lt;br /&gt;
.mw-highlight .bp,&lt;br /&gt;
.mw-highlight .na { color: #9fd6ff; }&lt;br /&gt;
.mw-highlight .nc,&lt;br /&gt;
.mw-highlight .nn { color: var(--sf-cyan); font-weight: bold; }&lt;br /&gt;
.mw-highlight .nf,&lt;br /&gt;
.mw-highlight .fm { color: var(--sf-cyan); }&lt;br /&gt;
.mw-highlight .nt { color: var(--sf-amber); }&lt;br /&gt;
.mw-highlight .nv,&lt;br /&gt;
.mw-highlight .vc,&lt;br /&gt;
.mw-highlight .vg,&lt;br /&gt;
.mw-highlight .vi,&lt;br /&gt;
.mw-highlight .vm { color: var(--sf-violet); }&lt;br /&gt;
.mw-highlight .no { color: #ffd479; }&lt;br /&gt;
.mw-highlight .nd,&lt;br /&gt;
.mw-highlight .ne { color: #ff8fa3; }&lt;br /&gt;
.mw-highlight .o,&lt;br /&gt;
.mw-highlight .ow { color: #a9c0d6; }&lt;br /&gt;
.mw-highlight .p { color: var(--sf-fg-dim); }&lt;br /&gt;
.mw-highlight .err { color: var(--sf-red); border: 1px solid rgba( 255, 95, 109, 0.5 ); }&lt;br /&gt;
.mw-highlight .gd { color: var(--sf-red); }&lt;br /&gt;
.mw-highlight .gi { color: var(--sf-mint); }&lt;br /&gt;
.mw-highlight .go { color: var(--sf-fg-dim); }&lt;br /&gt;
.mw-highlight .gr,&lt;br /&gt;
.mw-highlight .gt { color: var(--sf-red); }&lt;br /&gt;
.mw-highlight .gh,&lt;br /&gt;
.mw-highlight .gp,&lt;br /&gt;
.mw-highlight .gu { color: var(--sf-cyan); font-weight: bold; }&lt;br /&gt;
.mw-highlight .linenos,&lt;br /&gt;
.mw-highlight .lineno,&lt;br /&gt;
.mw-highlight .linenodiv pre { color: var(--sf-fg-faint); background: transparent; }&lt;br /&gt;
&lt;br /&gt;
/* Copy button shipped by SyntaxHighlight. */&lt;br /&gt;
.mw-highlight-copy,&lt;br /&gt;
.mw-highlight .mw-highlight-copy {&lt;br /&gt;
	background: var(--sf-panel-3);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ---------------------------------------------------------------- Tables -- */&lt;br /&gt;
&lt;br /&gt;
.mw-body-content table.wikitable,&lt;br /&gt;
.mw-body-content .mw-datatable,&lt;br /&gt;
.mw-body-content table.mw_metadata {&lt;br /&gt;
	background: var(--sf-panel-2);&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-collapse: collapse;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content table.wikitable &amp;gt; tr &amp;gt; th,&lt;br /&gt;
.mw-body-content table.wikitable &amp;gt; * &amp;gt; tr &amp;gt; th,&lt;br /&gt;
.mw-body-content .mw-datatable th,&lt;br /&gt;
.mw-body-content table.mw_metadata th {&lt;br /&gt;
	background: var(--sf-panel-3);&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	font-family: var(--sf-display);&lt;br /&gt;
	letter-spacing: 0.08em;&lt;br /&gt;
	text-transform: uppercase;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content table.wikitable &amp;gt; tr &amp;gt; td,&lt;br /&gt;
.mw-body-content table.wikitable &amp;gt; * &amp;gt; tr &amp;gt; td,&lt;br /&gt;
.mw-body-content .mw-datatable td,&lt;br /&gt;
.mw-body-content table.mw_metadata td {&lt;br /&gt;
	border: 1px solid var(--sf-line-soft);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content table.wikitable &amp;gt; * &amp;gt; tr:hover &amp;gt; td {&lt;br /&gt;
	background: rgba( 53, 224, 255, 0.05 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* The pin-out tables (Raspberry Pi) set cell fills inline. Those fills split&lt;br /&gt;
   into light and dark, so each needs its own text colour — a single blanket&lt;br /&gt;
   rule leaves half the pin numbers invisible. Values enumerated from the&lt;br /&gt;
   wikitext: cyan / #00FF80 / magenta / yellow / orange are light; black, green&lt;br /&gt;
   and red are dark. */&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;#00FFFF&amp;quot; i],&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;#00FF80&amp;quot; i],&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;#FF00FF&amp;quot; i],&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;cyan&amp;quot; i],&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;yellow&amp;quot; i],&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;orange&amp;quot; i],&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;white&amp;quot; i],&lt;br /&gt;
.mw-parser-output th[style*=&amp;quot;#00FFFF&amp;quot; i],&lt;br /&gt;
.mw-parser-output th[style*=&amp;quot;yellow&amp;quot; i] {&lt;br /&gt;
	color: #06121f !important;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;black&amp;quot; i],&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;green&amp;quot; i],&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;red&amp;quot; i],&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;blue&amp;quot; i],&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;navy&amp;quot; i],&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;purple&amp;quot; i] {&lt;br /&gt;
	color: #ffffff !important;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-parser-output td[style*=&amp;quot;background&amp;quot; i] a,&lt;br /&gt;
.mw-parser-output th[style*=&amp;quot;background&amp;quot; i] a {&lt;br /&gt;
	color: inherit !important;&lt;br /&gt;
	text-decoration: underline;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Inline red text (used as a warning marker on a handful of pages) is too dark&lt;br /&gt;
   against this background — lift it without changing the wikitext. */&lt;br /&gt;
.mw-parser-output [style*=&amp;quot;color:red&amp;quot;],&lt;br /&gt;
.mw-parser-output [style*=&amp;quot;color: red&amp;quot;],&lt;br /&gt;
.mw-parser-output [style*=&amp;quot;color:#FF0000&amp;quot;],&lt;br /&gt;
.mw-parser-output [style*=&amp;quot;color:#ff0000&amp;quot;],&lt;br /&gt;
.mw-parser-output [style*=&amp;quot;color: #FF0000&amp;quot;] {&lt;br /&gt;
	color: var(--sf-red) !important;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ---------------------------------------------------- Images &amp;amp; galleries -- */&lt;br /&gt;
&lt;br /&gt;
.mw-body-content .thumbinner,&lt;br /&gt;
.mw-parser-output .thumbinner {&lt;br /&gt;
	background: var(--sf-panel-2);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
	box-shadow: 0 4px 18px rgba( 0, 0, 0, 0.45 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* MediaWiki 1.43 media markup: thumbs are figure[typeof~=&amp;quot;mw:File/Thumb&amp;quot;], which core paints&lt;br /&gt;
   #f9f9f9 with #ccc borders (a white frame on this dark theme). Same HUD frame as .thumbinner&lt;br /&gt;
   above; no inner border, so transparent diagrams blend into the frame. */&lt;br /&gt;
.mw-body-content figure[typeof~=&#039;mw:File/Thumb&#039;],&lt;br /&gt;
.mw-body-content figure[typeof~=&#039;mw:File/Frame&#039;] {&lt;br /&gt;
	background: var(--sf-panel-2);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-bottom: 0;&lt;br /&gt;
	box-shadow: 0 4px 18px rgba( 0, 0, 0, 0.45 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content figure[typeof~=&#039;mw:File/Thumb&#039;] &amp;gt; :not( figcaption ) .mw-file-element,&lt;br /&gt;
.mw-body-content figure[typeof~=&#039;mw:File/Frame&#039;] &amp;gt; :not( figcaption ) .mw-file-element {&lt;br /&gt;
	background: transparent;&lt;br /&gt;
	border: 0;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content figure[typeof~=&#039;mw:File/Thumb&#039;] &amp;gt; figcaption,&lt;br /&gt;
.mw-body-content figure[typeof~=&#039;mw:File/Frame&#039;] &amp;gt; figcaption {&lt;br /&gt;
	background: var(--sf-panel-2);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-top: 0;&lt;br /&gt;
	color: var(--sf-fg-dim);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content .thumbimage,&lt;br /&gt;
.mw-parser-output .thumbimage {&lt;br /&gt;
	background: var(--sf-bg-deep);&lt;br /&gt;
	border: 1px solid var(--sf-line-soft);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content .thumbcaption,&lt;br /&gt;
.mw-parser-output .thumbcaption {&lt;br /&gt;
	color: var(--sf-fg-dim);&lt;br /&gt;
	font-family: var(--sf-mono);&lt;br /&gt;
	font-size: 0.85em;&lt;br /&gt;
	line-height: 1.5;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Enlarge icon is a dark grey SVG. */&lt;br /&gt;
.mw-body-content .magnify a,&lt;br /&gt;
.mw-parser-output .magnify a {&lt;br /&gt;
	filter: invert( 1 ) brightness( 1.1 );&lt;br /&gt;
	opacity: 0.7;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content ul.gallery {&lt;br /&gt;
	margin: 1.4em 0 2em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content .gallerybox div.thumb {&lt;br /&gt;
	background: var(--sf-panel-2);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
	transition: border-color 0.15s ease, box-shadow 0.15s ease;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content .gallerybox:hover div.thumb {&lt;br /&gt;
	border-color: var(--sf-cyan);&lt;br /&gt;
	box-shadow: 0 0 14px rgba( 53, 224, 255, 0.35 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content .gallerybox .thumb img {&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
	box-shadow: none;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-body-content .gallerybox .gallerytext {&lt;br /&gt;
	font-size: 0.88em;&lt;br /&gt;
	color: var(--sf-fg-dim);&lt;br /&gt;
	padding-top: 0.5em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ------------------------------------------------------------------- TOC -- */&lt;br /&gt;
&lt;br /&gt;
#toc,&lt;br /&gt;
.toc,&lt;br /&gt;
.toccolours {&lt;br /&gt;
	background: var(--sf-bg-deep);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
	box-shadow: inset 0 0 20px rgba( 53, 224, 255, 0.05 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#toc .toctitle h2,&lt;br /&gt;
.toc .toctitle h2,&lt;br /&gt;
#mw-toc-heading {&lt;br /&gt;
	color: var(--sf-cyan);&lt;br /&gt;
	font-family: var(--sf-display);&lt;br /&gt;
	font-size: 0.85em;&lt;br /&gt;
	letter-spacing: 0.2em;&lt;br /&gt;
	text-transform: uppercase;&lt;br /&gt;
	border: 0;&lt;br /&gt;
	background: none;&lt;br /&gt;
	padding: 0;&lt;br /&gt;
	text-shadow: none;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#toc .tocnumber,&lt;br /&gt;
.toc .tocnumber {&lt;br /&gt;
	color: var(--sf-amber);&lt;br /&gt;
	font-family: var(--sf-mono);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#toc .toctext,&lt;br /&gt;
.toc .toctext {&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ------------------------------------------------- Categories &amp;amp; footer -- */&lt;br /&gt;
&lt;br /&gt;
#catlinks,&lt;br /&gt;
.catlinks {&lt;br /&gt;
	background: var(--sf-bg-deep);&lt;br /&gt;
	border: 1px solid var(--sf-line-soft);&lt;br /&gt;
	border-left: 3px solid var(--sf-amber);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
	color: var(--sf-fg-dim);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#catlinks .mw-normal-catlinks &amp;gt; a {&lt;br /&gt;
	color: var(--sf-amber);&lt;br /&gt;
	font-family: var(--sf-display);&lt;br /&gt;
	letter-spacing: 0.1em;&lt;br /&gt;
	text-transform: uppercase;&lt;br /&gt;
	font-size: 0.85em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#mw-footer-container,&lt;br /&gt;
#mw-footer {&lt;br /&gt;
	background: var(--sf-bg-deep);&lt;br /&gt;
	color: var(--sf-fg-faint);&lt;br /&gt;
	border-top: 1px solid var(--sf-line);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
#mw-footer li,&lt;br /&gt;
#footer-info li {&lt;br /&gt;
	color: var(--sf-fg-faint);&lt;br /&gt;
	font-family: var(--sf-mono);&lt;br /&gt;
	font-size: 0.82em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* -------------------------------------------------- Forms / editing UI -- */&lt;br /&gt;
&lt;br /&gt;
textarea,&lt;br /&gt;
#wpTextbox1,&lt;br /&gt;
.mw-editform textarea {&lt;br /&gt;
	background: var(--sf-bg-deep);&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
	caret-color: var(--sf-cyan);&lt;br /&gt;
	font-family: var(--sf-mono);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
textarea:focus,&lt;br /&gt;
#wpTextbox1:focus {&lt;br /&gt;
	border-color: var(--sf-cyan);&lt;br /&gt;
	box-shadow: 0 0 0 1px rgba( 53, 224, 255, 0.3 );&lt;br /&gt;
	outline: 0;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
input[type=&amp;quot;text&amp;quot;],&lt;br /&gt;
input[type=&amp;quot;search&amp;quot;],&lt;br /&gt;
input[type=&amp;quot;password&amp;quot;],&lt;br /&gt;
input[type=&amp;quot;number&amp;quot;],&lt;br /&gt;
select {&lt;br /&gt;
	background: var(--sf-bg-deep);&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-editform #wpSummary,&lt;br /&gt;
#wpSummary {&lt;br /&gt;
	background: var(--sf-bg-deep);&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Preview / diff / history and message boxes. */&lt;br /&gt;
table.diff,&lt;br /&gt;
table.diff td,&lt;br /&gt;
.mw-diff-table {&lt;br /&gt;
	background: transparent;&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
td.diff-context,&lt;br /&gt;
.diff-context {&lt;br /&gt;
	background: var(--sf-panel-2);&lt;br /&gt;
	color: var(--sf-fg-dim);&lt;br /&gt;
	border-color: var(--sf-line-soft);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
td.diff-deletedline,&lt;br /&gt;
.diff-deletedline {&lt;br /&gt;
	background: rgba( 255, 95, 109, 0.12 );&lt;br /&gt;
	border-color: rgba( 255, 95, 109, 0.4 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
td.diff-addedline,&lt;br /&gt;
.diff-addedline {&lt;br /&gt;
	background: rgba( 124, 246, 200, 0.12 );&lt;br /&gt;
	border-color: rgba( 124, 246, 200, 0.4 );&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.diff-deletedline .diffchange {&lt;br /&gt;
	background: rgba( 255, 95, 109, 0.35 );&lt;br /&gt;
	color: #fff;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.diff-addedline .diffchange {&lt;br /&gt;
	background: rgba( 124, 246, 200, 0.3 );&lt;br /&gt;
	color: #06121f;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-message-box,&lt;br /&gt;
.mw-message-box-warning,&lt;br /&gt;
.mw-message-box-error,&lt;br /&gt;
.mw-message-box-success,&lt;br /&gt;
.warningbox,&lt;br /&gt;
.errorbox,&lt;br /&gt;
.successbox,&lt;br /&gt;
.mw-warning,&lt;br /&gt;
#mw-content .mw-message-box {&lt;br /&gt;
	background: var(--sf-panel-2);&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
	border: 1px solid var(--sf-line);&lt;br /&gt;
	border-left: 3px solid var(--sf-amber);&lt;br /&gt;
	border-radius: 0;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-message-box-error,&lt;br /&gt;
.errorbox {&lt;br /&gt;
	border-left-color: var(--sf-red);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-message-box-success,&lt;br /&gt;
.successbox {&lt;br /&gt;
	border-left-color: var(--sf-mint);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* Special: pages and search results ride on .mw-body-content, but a few bits&lt;br /&gt;
   carry their own light backgrounds. */&lt;br /&gt;
.mw-search-result-heading,&lt;br /&gt;
.searchresult {&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-search-results li {&lt;br /&gt;
	border-bottom: 1px solid var(--sf-line-soft);&lt;br /&gt;
	padding-bottom: 0.5em;&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
.mw-changeslist-line,&lt;br /&gt;
.mw-changeslist td,&lt;br /&gt;
.mw-changeslist th,&lt;br /&gt;
.mw-history-histlinks {&lt;br /&gt;
	color: var(--sf-fg);&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* ------------------------------------------------- Narrow viewports (Timeless&lt;br /&gt;
   switches to its stacked layout with drop-down menus at 850px) -- */&lt;br /&gt;
&lt;br /&gt;
@media screen and ( max-width: 850px ) {&lt;br /&gt;
	#mw-header-container {&lt;br /&gt;
		background: var(--sf-bg);&lt;br /&gt;
	}&lt;br /&gt;
&lt;br /&gt;
	/* Menu / user / tools toggles are dark grey SVGs on a dark bar. */&lt;br /&gt;
	#user-tools h2,&lt;br /&gt;
	.sidebar-chunk h2,&lt;br /&gt;
	#ca-more span,&lt;br /&gt;
	#ca-languages span,&lt;br /&gt;
	.tools-inline li a {&lt;br /&gt;
		filter: invert( 1 ) brightness( 1.15 );&lt;br /&gt;
	}&lt;br /&gt;
&lt;br /&gt;
	.sidebar-inner,&lt;br /&gt;
	.dropdown {&lt;br /&gt;
		background: var(--sf-panel-2);&lt;br /&gt;
		border: 1px solid var(--sf-line);&lt;br /&gt;
		box-shadow: 0 8px 28px rgba( 0, 0, 0, 0.7 );&lt;br /&gt;
	}&lt;br /&gt;
&lt;br /&gt;
	/* Backdrop behind an open menu was a light wash. */&lt;br /&gt;
	#menus-cover {&lt;br /&gt;
		background: #02060d;&lt;br /&gt;
		opacity: 0.82;&lt;br /&gt;
	}&lt;br /&gt;
&lt;br /&gt;
	#mw-content {&lt;br /&gt;
		border-left: 0;&lt;br /&gt;
		border-right: 0;&lt;br /&gt;
	}&lt;br /&gt;
&lt;br /&gt;
	#mw-content::before,&lt;br /&gt;
	#mw-content::after {&lt;br /&gt;
		display: none;&lt;br /&gt;
	}&lt;br /&gt;
}&lt;br /&gt;
&lt;br /&gt;
/* --------------------------------------------------------------- Printing -- */&lt;br /&gt;
&lt;br /&gt;
@media print {&lt;br /&gt;
	/* Never print the dark theme — it wastes toner and reads badly on paper. */&lt;br /&gt;
	body,&lt;br /&gt;
	#mw-content,&lt;br /&gt;
	#mw-content-container,&lt;br /&gt;
	.mw-body-content,&lt;br /&gt;
	#toc,&lt;br /&gt;
	.toc,&lt;br /&gt;
	#catlinks,&lt;br /&gt;
	.mw-parser-output pre,&lt;br /&gt;
	.mw-body-content pre,&lt;br /&gt;
	.mw-highlight pre,&lt;br /&gt;
	.mw-parser-output code,&lt;br /&gt;
	.thumbinner {&lt;br /&gt;
		background: #fff !important;&lt;br /&gt;
		color: #000 !important;&lt;br /&gt;
		box-shadow: none !important;&lt;br /&gt;
		text-shadow: none !important;&lt;br /&gt;
	}&lt;br /&gt;
&lt;br /&gt;
	.mw-body-content h2,&lt;br /&gt;
	.mw-body-content h3,&lt;br /&gt;
	#firstHeading {&lt;br /&gt;
		color: #000 !important;&lt;br /&gt;
		background: none !important;&lt;br /&gt;
		text-shadow: none !important;&lt;br /&gt;
	}&lt;br /&gt;
&lt;br /&gt;
	#p-logo img {&lt;br /&gt;
		filter: none !important;&lt;br /&gt;
		mix-blend-mode: normal !important;&lt;br /&gt;
	}&lt;br /&gt;
&lt;br /&gt;
	#mw-content::before,&lt;br /&gt;
	#mw-content::after {&lt;br /&gt;
		display: none !important;&lt;br /&gt;
	}&lt;br /&gt;
}&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_abv_attenuation.png&amp;diff=849</id>
		<title>檔案:Homebrew abv attenuation.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_abv_attenuation.png&amp;diff=849"/>
		<updated>2026-10-28T13:38:43Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Homebrew abv attenuation.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Homebrewing charts (generated from brew-log by plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Cicd_golden.png&amp;diff=848</id>
		<title>檔案:Cicd golden.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Cicd_golden.png&amp;diff=848"/>
		<updated>2026-10-28T13:38:43Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Cicd golden.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Firmware CI/CD on GitHub Actions diagram&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Experience_technologies.png&amp;diff=847</id>
		<title>檔案:Experience technologies.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Experience_technologies.png&amp;diff=847"/>
		<updated>2026-10-28T13:38:43Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Experience technologies.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Main page charts (generated by mainpage_plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_og_fg_by_batch.png&amp;diff=846</id>
		<title>檔案:Homebrew og fg by batch.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_og_fg_by_batch.png&amp;diff=846"/>
		<updated>2026-10-28T13:38:43Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Homebrew og fg by batch.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Homebrewing charts (generated from brew-log by plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_gravity_curves.png&amp;diff=845</id>
		<title>檔案:Homebrew gravity curves.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_gravity_curves.png&amp;diff=845"/>
		<updated>2026-10-28T13:38:43Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Homebrew gravity curves.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Homebrewing charts (generated from brew-log by plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Experience_platforms.png&amp;diff=844</id>
		<title>檔案:Experience platforms.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Experience_platforms.png&amp;diff=844"/>
		<updated>2026-10-28T13:38:42Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Experience platforms.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Main page charts (generated by mainpage_plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_states.png&amp;diff=843</id>
		<title>檔案:Agentic states.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_states.png&amp;diff=843"/>
		<updated>2026-10-28T13:38:42Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Agentic states.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Agentic coding diagrams: flow, ticket states, one agent run&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_run.png&amp;diff=842</id>
		<title>檔案:Agentic run.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_run.png&amp;diff=842"/>
		<updated>2026-10-28T13:38:42Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Agentic run.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Agentic coding diagrams: flow, ticket states, one agent run&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_plan_vs_actual.png&amp;diff=841</id>
		<title>檔案:Homebrew plan vs actual.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_plan_vs_actual.png&amp;diff=841"/>
		<updated>2026-10-28T13:38:42Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Homebrew plan vs actual.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Homebrewing charts (generated from brew-log by plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_ph.png&amp;diff=840</id>
		<title>檔案:Homebrew ph.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_ph.png&amp;diff=840"/>
		<updated>2026-10-28T13:38:42Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Homebrew ph.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Homebrewing charts (generated from brew-log by plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_flow.png&amp;diff=839</id>
		<title>檔案:Agentic flow.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_flow.png&amp;diff=839"/>
		<updated>2026-10-28T13:38:42Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Agentic flow.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Agentic coding diagrams: flow, ticket states, one agent run&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E9%A6%96%E9%A0%81&amp;diff=838</id>
		<title>首頁</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E9%A6%96%E9%A0%81&amp;diff=838"/>
		<updated>2026-10-28T13:28:09Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Main page: remove duplicate AI agents subsection (now under CI/CD and AI agents)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;220&amp;quot;&amp;gt;&lt;br /&gt;
File:3D_ripples_01.gif|3D LED cube &amp;amp;ndash; ripple effect&lt;br /&gt;
File:MeArm_stm32f4.jpg|MeArm robotic arm driven by STM32F4&lt;br /&gt;
File:Gps_wifiap.jpg|GPS tracker with Wi-Fi AP&lt;br /&gt;
File:Eye_tracker_alexa.jpg|Eye tracker with Alexa and AWS IoT&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
&amp;lt;div style=&amp;quot;display:flex;flex-wrap:wrap;gap:12px;margin:0 0 18px&amp;quot;&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;{{PAGESINCATEGORY:Boards and SoCs|pages}}&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[#Boards and SoCs|boards &amp;amp;amp; SoCs]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;{{PAGESINCATEGORY:Projects|pages}}&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[#Projects|hands-on projects]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;2010&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[RTD1073|shipping products since]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;11&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[Homebrewing|homebrew batches]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;/div&amp;gt;&lt;br /&gt;
[[File:Experience_platforms.png|frameless|1000px|Hardware I&#039;ve worked on|link=Category:Boards and SoCs]]&lt;br /&gt;
&lt;br /&gt;
[[File:Experience_technologies.png|frameless|1000px|Technologies across my engineering pages|link=Category:Projects]]&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
== About Tommy C. Liu ==&lt;br /&gt;
I have always believed I could flip the world through my own work &amp;amp;mdash; that is where the name of this wiki comes from. For most of my career that meant building things by hand: firmware, boards and products, one line of code at a time.&lt;br /&gt;
&lt;br /&gt;
AI changed that. It carried me past limits I could not step over on my own, and gave me courage, knowledge and speed I never had before. It also brings a new kind of uncertainty: the results are not always what you expect, so judgment matters more than ever.&lt;br /&gt;
&lt;br /&gt;
I have loved craft beer for years. I studied brewing again and again and brewed it in my head many times, but I never had the courage to actually start. With AI I finally had a strong partner to think it through and work on it together &amp;amp;mdash; so I began. Eleven batches of hazy IPA in under a year, every one logged, measured and learned from (see [[Homebrewing]]).&lt;br /&gt;
&lt;br /&gt;
My role has shifted from &#039;&#039;&#039;creator and maintainer&#039;&#039;&#039; to &#039;&#039;&#039;supervisor&#039;&#039;&#039; &amp;amp;mdash; deciding what to build, setting the direction, and reviewing what comes back. That is leadership, and anyone who works with AI is already doing it. AI has not taken our skills away; it has promoted us. I call it the &#039;&#039;&#039;AI promotion&#039;&#039;&#039;.&lt;br /&gt;
&lt;br /&gt;
So the question I ask myself now is: &#039;&#039;where do I want to go next?&#039;&#039;&lt;br /&gt;
&lt;br /&gt;
How about you &amp;amp;mdash; have you been promoted yet?&lt;br /&gt;
&lt;br /&gt;
----&lt;br /&gt;
* Welcome to discuss outsourcing project&lt;br /&gt;
** FW/SW development&lt;br /&gt;
** Training course  &lt;br /&gt;
* E-mail: tommy.using@gmail.com&lt;br /&gt;
* LinkedIn: https://www.linkedin.com/in/chia-chuan-liu-7706b473/&lt;br /&gt;
&lt;br /&gt;
=== Hobbies ===&lt;br /&gt;
* [[Homebrewing]] &amp;amp;mdash; hazy IPA / NEIPA batches, brew log and lessons learned&lt;br /&gt;
* [[LG WE16HW dryer teardown and cleaning]] &amp;amp;mdash; heat-pump dryer taken apart to clear lint from the base and coils, with before/after photos&lt;br /&gt;
&lt;br /&gt;
== Experience ==&lt;br /&gt;
=== Projects ===&lt;br /&gt;
Hands-on builds and application notes &amp;amp;mdash; [[:Category:Projects|all {{PAGESINCATEGORY:Projects|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Boards and SoCs ===&lt;br /&gt;
Dev boards, MCUs and SoCs &amp;amp;mdash; [[:Category:Boards and SoCs|all {{PAGESINCATEGORY:Boards and SoCs|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Software and Tools ===&lt;br /&gt;
OS, build and test tooling &amp;amp;mdash; [[:Category:Software and Tools|all {{PAGESINCATEGORY:Software and Tools|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Sensors and Modules ===&lt;br /&gt;
Sensors and peripheral modules &amp;amp;mdash; [[:Category:Sensors and Modules|all {{PAGESINCATEGORY:Sensors and Modules|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Connectivity ===&lt;br /&gt;
Buses and wireless protocols &amp;amp;mdash; [[:Category:Connectivity|all {{PAGESINCATEGORY:Connectivity|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== CI/CD and AI agents ===&lt;br /&gt;
* [[Firmware CI/CD on GitHub Actions]]&lt;br /&gt;
* [[Agentic coding, ticket to merged code]]&lt;br /&gt;
&lt;br /&gt;
=== Equipment ===&lt;br /&gt;
* Measurement&lt;br /&gt;
** Logic Analyzer&lt;br /&gt;
*** Max Sample Rate 400MHz&lt;br /&gt;
*** 16G stream / 256M buffer&lt;br /&gt;
*** 16 Channels&lt;br /&gt;
** Multimeter&lt;br /&gt;
&lt;br /&gt;
* Computings&lt;br /&gt;
** Linux server - Intel Xeon E3-1231 v3(8 cores)&lt;br /&gt;
*** Intel(R) Xeon(R) CPU E3-1230 v3&lt;br /&gt;
*** 24G RAM&lt;br /&gt;
** Windows desktop&lt;br /&gt;
*** Intel(R) Xeon(R) CPU E3-1231&lt;br /&gt;
** Linux laptop&lt;br /&gt;
*** i7-7600U&lt;br /&gt;
*** 16G RAM&lt;br /&gt;
&lt;br /&gt;
== Reading &amp;amp; sharing ==&lt;br /&gt;
* [[Software version]]&lt;br /&gt;
* [[Software release life cycle]]&lt;br /&gt;
* [[Agent Working Rules]]&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E9%A6%96%E9%A0%81&amp;diff=837</id>
		<title>首頁</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E9%A6%96%E9%A0%81&amp;diff=837"/>
		<updated>2026-10-28T13:26:10Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Main page: CI/CD and AI agents in one subsection; Jenkins link removed (page kept in Software and Tools)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;220&amp;quot;&amp;gt;&lt;br /&gt;
File:3D_ripples_01.gif|3D LED cube &amp;amp;ndash; ripple effect&lt;br /&gt;
File:MeArm_stm32f4.jpg|MeArm robotic arm driven by STM32F4&lt;br /&gt;
File:Gps_wifiap.jpg|GPS tracker with Wi-Fi AP&lt;br /&gt;
File:Eye_tracker_alexa.jpg|Eye tracker with Alexa and AWS IoT&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
&amp;lt;div style=&amp;quot;display:flex;flex-wrap:wrap;gap:12px;margin:0 0 18px&amp;quot;&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;{{PAGESINCATEGORY:Boards and SoCs|pages}}&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[#Boards and SoCs|boards &amp;amp;amp; SoCs]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;{{PAGESINCATEGORY:Projects|pages}}&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[#Projects|hands-on projects]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;2010&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[RTD1073|shipping products since]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;11&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[Homebrewing|homebrew batches]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;/div&amp;gt;&lt;br /&gt;
[[File:Experience_platforms.png|frameless|1000px|Hardware I&#039;ve worked on|link=Category:Boards and SoCs]]&lt;br /&gt;
&lt;br /&gt;
[[File:Experience_technologies.png|frameless|1000px|Technologies across my engineering pages|link=Category:Projects]]&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
== About Tommy C. Liu ==&lt;br /&gt;
I have always believed I could flip the world through my own work &amp;amp;mdash; that is where the name of this wiki comes from. For most of my career that meant building things by hand: firmware, boards and products, one line of code at a time.&lt;br /&gt;
&lt;br /&gt;
AI changed that. It carried me past limits I could not step over on my own, and gave me courage, knowledge and speed I never had before. It also brings a new kind of uncertainty: the results are not always what you expect, so judgment matters more than ever.&lt;br /&gt;
&lt;br /&gt;
I have loved craft beer for years. I studied brewing again and again and brewed it in my head many times, but I never had the courage to actually start. With AI I finally had a strong partner to think it through and work on it together &amp;amp;mdash; so I began. Eleven batches of hazy IPA in under a year, every one logged, measured and learned from (see [[Homebrewing]]).&lt;br /&gt;
&lt;br /&gt;
My role has shifted from &#039;&#039;&#039;creator and maintainer&#039;&#039;&#039; to &#039;&#039;&#039;supervisor&#039;&#039;&#039; &amp;amp;mdash; deciding what to build, setting the direction, and reviewing what comes back. That is leadership, and anyone who works with AI is already doing it. AI has not taken our skills away; it has promoted us. I call it the &#039;&#039;&#039;AI promotion&#039;&#039;&#039;.&lt;br /&gt;
&lt;br /&gt;
So the question I ask myself now is: &#039;&#039;where do I want to go next?&#039;&#039;&lt;br /&gt;
&lt;br /&gt;
How about you &amp;amp;mdash; have you been promoted yet?&lt;br /&gt;
&lt;br /&gt;
----&lt;br /&gt;
* Welcome to discuss outsourcing project&lt;br /&gt;
** FW/SW development&lt;br /&gt;
** Training course  &lt;br /&gt;
* E-mail: tommy.using@gmail.com&lt;br /&gt;
* LinkedIn: https://www.linkedin.com/in/chia-chuan-liu-7706b473/&lt;br /&gt;
&lt;br /&gt;
=== Hobbies ===&lt;br /&gt;
* [[Homebrewing]] &amp;amp;mdash; hazy IPA / NEIPA batches, brew log and lessons learned&lt;br /&gt;
* [[LG WE16HW dryer teardown and cleaning]] &amp;amp;mdash; heat-pump dryer taken apart to clear lint from the base and coils, with before/after photos&lt;br /&gt;
&lt;br /&gt;
== Experience ==&lt;br /&gt;
=== Projects ===&lt;br /&gt;
Hands-on builds and application notes &amp;amp;mdash; [[:Category:Projects|all {{PAGESINCATEGORY:Projects|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Boards and SoCs ===&lt;br /&gt;
Dev boards, MCUs and SoCs &amp;amp;mdash; [[:Category:Boards and SoCs|all {{PAGESINCATEGORY:Boards and SoCs|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Software and Tools ===&lt;br /&gt;
OS, build and test tooling &amp;amp;mdash; [[:Category:Software and Tools|all {{PAGESINCATEGORY:Software and Tools|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Sensors and Modules ===&lt;br /&gt;
Sensors and peripheral modules &amp;amp;mdash; [[:Category:Sensors and Modules|all {{PAGESINCATEGORY:Sensors and Modules|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Connectivity ===&lt;br /&gt;
Buses and wireless protocols &amp;amp;mdash; [[:Category:Connectivity|all {{PAGESINCATEGORY:Connectivity|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== AI agents ===&lt;br /&gt;
* [[Agentic coding, ticket to merged code]]&lt;br /&gt;
&lt;br /&gt;
=== CI/CD and AI agents ===&lt;br /&gt;
* [[Firmware CI/CD on GitHub Actions]]&lt;br /&gt;
* [[Agentic coding, ticket to merged code]]&lt;br /&gt;
&lt;br /&gt;
=== Equipment ===&lt;br /&gt;
* Measurement&lt;br /&gt;
** Logic Analyzer&lt;br /&gt;
*** Max Sample Rate 400MHz&lt;br /&gt;
*** 16G stream / 256M buffer&lt;br /&gt;
*** 16 Channels&lt;br /&gt;
** Multimeter&lt;br /&gt;
&lt;br /&gt;
* Computings&lt;br /&gt;
** Linux server - Intel Xeon E3-1231 v3(8 cores)&lt;br /&gt;
*** Intel(R) Xeon(R) CPU E3-1230 v3&lt;br /&gt;
*** 24G RAM&lt;br /&gt;
** Windows desktop&lt;br /&gt;
*** Intel(R) Xeon(R) CPU E3-1231&lt;br /&gt;
** Linux laptop&lt;br /&gt;
*** i7-7600U&lt;br /&gt;
*** 16G RAM&lt;br /&gt;
&lt;br /&gt;
== Reading &amp;amp; sharing ==&lt;br /&gt;
* [[Software version]]&lt;br /&gt;
* [[Software release life cycle]]&lt;br /&gt;
* [[Agent Working Rules]]&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E9%A6%96%E9%A0%81&amp;diff=836</id>
		<title>首頁</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E9%A6%96%E9%A0%81&amp;diff=836"/>
		<updated>2026-10-28T13:24:22Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Main page: new AI agents subsection under Experience, above CI/CD&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;220&amp;quot;&amp;gt;&lt;br /&gt;
File:3D_ripples_01.gif|3D LED cube &amp;amp;ndash; ripple effect&lt;br /&gt;
File:MeArm_stm32f4.jpg|MeArm robotic arm driven by STM32F4&lt;br /&gt;
File:Gps_wifiap.jpg|GPS tracker with Wi-Fi AP&lt;br /&gt;
File:Eye_tracker_alexa.jpg|Eye tracker with Alexa and AWS IoT&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
&amp;lt;div style=&amp;quot;display:flex;flex-wrap:wrap;gap:12px;margin:0 0 18px&amp;quot;&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;{{PAGESINCATEGORY:Boards and SoCs|pages}}&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[#Boards and SoCs|boards &amp;amp;amp; SoCs]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;{{PAGESINCATEGORY:Projects|pages}}&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[#Projects|hands-on projects]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;2010&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[RTD1073|shipping products since]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;11&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[Homebrewing|homebrew batches]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;/div&amp;gt;&lt;br /&gt;
[[File:Experience_platforms.png|frameless|1000px|Hardware I&#039;ve worked on|link=Category:Boards and SoCs]]&lt;br /&gt;
&lt;br /&gt;
[[File:Experience_technologies.png|frameless|1000px|Technologies across my engineering pages|link=Category:Projects]]&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
== About Tommy C. Liu ==&lt;br /&gt;
I have always believed I could flip the world through my own work &amp;amp;mdash; that is where the name of this wiki comes from. For most of my career that meant building things by hand: firmware, boards and products, one line of code at a time.&lt;br /&gt;
&lt;br /&gt;
AI changed that. It carried me past limits I could not step over on my own, and gave me courage, knowledge and speed I never had before. It also brings a new kind of uncertainty: the results are not always what you expect, so judgment matters more than ever.&lt;br /&gt;
&lt;br /&gt;
I have loved craft beer for years. I studied brewing again and again and brewed it in my head many times, but I never had the courage to actually start. With AI I finally had a strong partner to think it through and work on it together &amp;amp;mdash; so I began. Eleven batches of hazy IPA in under a year, every one logged, measured and learned from (see [[Homebrewing]]).&lt;br /&gt;
&lt;br /&gt;
My role has shifted from &#039;&#039;&#039;creator and maintainer&#039;&#039;&#039; to &#039;&#039;&#039;supervisor&#039;&#039;&#039; &amp;amp;mdash; deciding what to build, setting the direction, and reviewing what comes back. That is leadership, and anyone who works with AI is already doing it. AI has not taken our skills away; it has promoted us. I call it the &#039;&#039;&#039;AI promotion&#039;&#039;&#039;.&lt;br /&gt;
&lt;br /&gt;
So the question I ask myself now is: &#039;&#039;where do I want to go next?&#039;&#039;&lt;br /&gt;
&lt;br /&gt;
How about you &amp;amp;mdash; have you been promoted yet?&lt;br /&gt;
&lt;br /&gt;
----&lt;br /&gt;
* Welcome to discuss outsourcing project&lt;br /&gt;
** FW/SW development&lt;br /&gt;
** Training course  &lt;br /&gt;
* E-mail: tommy.using@gmail.com&lt;br /&gt;
* LinkedIn: https://www.linkedin.com/in/chia-chuan-liu-7706b473/&lt;br /&gt;
&lt;br /&gt;
=== Hobbies ===&lt;br /&gt;
* [[Homebrewing]] &amp;amp;mdash; hazy IPA / NEIPA batches, brew log and lessons learned&lt;br /&gt;
* [[LG WE16HW dryer teardown and cleaning]] &amp;amp;mdash; heat-pump dryer taken apart to clear lint from the base and coils, with before/after photos&lt;br /&gt;
&lt;br /&gt;
== Experience ==&lt;br /&gt;
=== Projects ===&lt;br /&gt;
Hands-on builds and application notes &amp;amp;mdash; [[:Category:Projects|all {{PAGESINCATEGORY:Projects|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Boards and SoCs ===&lt;br /&gt;
Dev boards, MCUs and SoCs &amp;amp;mdash; [[:Category:Boards and SoCs|all {{PAGESINCATEGORY:Boards and SoCs|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Software and Tools ===&lt;br /&gt;
OS, build and test tooling &amp;amp;mdash; [[:Category:Software and Tools|all {{PAGESINCATEGORY:Software and Tools|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Sensors and Modules ===&lt;br /&gt;
Sensors and peripheral modules &amp;amp;mdash; [[:Category:Sensors and Modules|all {{PAGESINCATEGORY:Sensors and Modules|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Connectivity ===&lt;br /&gt;
Buses and wireless protocols &amp;amp;mdash; [[:Category:Connectivity|all {{PAGESINCATEGORY:Connectivity|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== AI agents ===&lt;br /&gt;
* [[Agentic coding, ticket to merged code]]&lt;br /&gt;
&lt;br /&gt;
=== CI/CD ===&lt;br /&gt;
* [[Firmware CI/CD on GitHub Actions]]&lt;br /&gt;
* [[Jenkins]]&lt;br /&gt;
&lt;br /&gt;
=== Equipment ===&lt;br /&gt;
* Measurement&lt;br /&gt;
** Logic Analyzer&lt;br /&gt;
*** Max Sample Rate 400MHz&lt;br /&gt;
*** 16G stream / 256M buffer&lt;br /&gt;
*** 16 Channels&lt;br /&gt;
** Multimeter&lt;br /&gt;
&lt;br /&gt;
* Computings&lt;br /&gt;
** Linux server - Intel Xeon E3-1231 v3(8 cores)&lt;br /&gt;
*** Intel(R) Xeon(R) CPU E3-1230 v3&lt;br /&gt;
*** 24G RAM&lt;br /&gt;
** Windows desktop&lt;br /&gt;
*** Intel(R) Xeon(R) CPU E3-1231&lt;br /&gt;
** Linux laptop&lt;br /&gt;
*** i7-7600U&lt;br /&gt;
*** 16G RAM&lt;br /&gt;
&lt;br /&gt;
== Reading &amp;amp; sharing ==&lt;br /&gt;
* [[Software version]]&lt;br /&gt;
* [[Software release life cycle]]&lt;br /&gt;
* [[Agent Working Rules]]&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=Agentic_coding,_ticket_to_merged_code&amp;diff=835</id>
		<title>Agentic coding, ticket to merged code</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=Agentic_coding,_ticket_to_merged_code&amp;diff=835"/>
		<updated>2026-10-28T13:24:20Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​New page: agentic coding loop, ticket to merged code (flow, ticket states, one run, rules)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&#039;&#039;&#039;Agentic coding, ticket to merged code&#039;&#039;&#039; &amp;amp;mdash; a self-hosted loop where I write the ticket and press merge, and two AI agents do everything in between.&lt;br /&gt;
&lt;br /&gt;
[[File:Agentic_flow.png|1200px|thumb|center|I write the ticket and set it to &#039;&#039;Agent ready&#039;&#039;; a developer agent writes code and tests, checks run, a review agent reviews, and a pull request comes back to me. Blocking review findings and my &amp;lt;code&amp;gt;/agent …&amp;lt;/code&amp;gt; comments send the work back.]]&lt;br /&gt;
&lt;br /&gt;
== What the ticket shows ==&lt;br /&gt;
[[File:Agentic_states.png|1000px|thumb|center|The ticket moves by itself. Green: ready to merge. Purple or red: look closer. A &amp;lt;code&amp;gt;/agent …&amp;lt;/code&amp;gt; comment sends it back to the agent.]]&lt;br /&gt;
&lt;br /&gt;
== Inside one agent run ==&lt;br /&gt;
[[File:Agentic_run.png|1200px|thumb|center|Every &amp;quot;no&amp;quot; sends feedback back to the developer agent. After 3 rounds it stops and hands over.]]&lt;br /&gt;
&lt;br /&gt;
== Rules ==&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
! rule !! how&lt;br /&gt;
|-&lt;br /&gt;
| Only I start it || ticket changes and &amp;lt;code&amp;gt;/agent&amp;lt;/code&amp;gt; comments from anyone else are ignored&lt;br /&gt;
|-&lt;br /&gt;
| Only I merge || agents work on their own branch and can never push to main&lt;br /&gt;
|-&lt;br /&gt;
| Agents can&#039;t cheat || changes to existing tests or to the CI setup are rolled back&lt;br /&gt;
|-&lt;br /&gt;
| It always stops || at most 3 rounds per run; nothing restarts without me&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== How this differs from CI/CD ==&lt;br /&gt;
CI/CD builds, tests and releases code that is already written. This loop &#039;&#039;writes&#039;&#039; the code from a ticket and uses CI as one of its checks. Its review agent runs before the pull request exists; the automated review in [[Firmware CI/CD on GitHub Actions]] reviews pull requests opened by anyone.&lt;br /&gt;
&lt;br /&gt;
== Stack ==&lt;br /&gt;
Plane (tickets) &amp;amp;middot; Gitea (code, pull requests, CI) &amp;amp;middot; Claude Code in headless mode (developer and review agents) &amp;amp;middot; a small webhook receiver and orchestrator script.&lt;br /&gt;
&lt;br /&gt;
[[Category:Software Engineering]]&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_states.png&amp;diff=834</id>
		<title>檔案:Agentic states.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_states.png&amp;diff=834"/>
		<updated>2026-10-28T13:24:20Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
Agentic coding diagrams: flow, ticket states, one agent run&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Agentic coding diagrams: flow, ticket states, one agent run&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_run.png&amp;diff=833</id>
		<title>檔案:Agentic run.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_run.png&amp;diff=833"/>
		<updated>2026-10-28T13:24:20Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
Agentic coding diagrams: flow, ticket states, one agent run&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Agentic coding diagrams: flow, ticket states, one agent run&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_flow.png&amp;diff=832</id>
		<title>檔案:Agentic flow.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Agentic_flow.png&amp;diff=832"/>
		<updated>2026-10-28T13:24:20Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
Agentic coding diagrams: flow, ticket states, one agent run&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Agentic coding diagrams: flow, ticket states, one agent run&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_gravity_curves.png&amp;diff=831</id>
		<title>檔案:Homebrew gravity curves.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_gravity_curves.png&amp;diff=831"/>
		<updated>2026-10-21T11:38:10Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Homebrew gravity curves.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Homebrewing charts (generated from brew-log by plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_plan_vs_actual.png&amp;diff=830</id>
		<title>檔案:Homebrew plan vs actual.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_plan_vs_actual.png&amp;diff=830"/>
		<updated>2026-10-21T11:38:10Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Homebrew plan vs actual.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Homebrewing charts (generated from brew-log by plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=829</id>
		<title>Homebrewing</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=Homebrewing&amp;diff=829"/>
		<updated>2026-10-21T11:20:00Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​#0011 finished: results, Crazy Concubine v4.4, post-boil pH, salt, last dry hop, cold crash now standard, green light by D-rest time&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&#039;&#039;&#039;Homebrewing&#039;&#039;&#039; — eleven batches of hazy / New England IPA brewed at home in 10-litre batches, December 2025 to October 2026. Every batch is logged with gravity readings, pH, hop timing and a tasting note, and each recipe is version-controlled in git so every change can be traced back to the batch that caused it. This page summarises what was brewed, what went wrong, and the process rules that came out of it.&lt;br /&gt;
&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_glass_orange_glow.jpg|Orange Glow v2.3 (#0007), five weeks after brew day&lt;br /&gt;
File:Homebrew_glass_in_hand.jpg|A finished hazy IPA, in the glass with the piz brewing logo&lt;br /&gt;
File:Homebrew_keg_and_bottles.jpg|Keg and bottles, with the Strange Haze label in front&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! Batches !! Period !! Batch size !! Styles !! Recipes (current version)&lt;br /&gt;
|-&lt;br /&gt;
| 11 (12 brews) || 2025-12 → 2026-10 || 6–11.5 L || Hazy IPA, DDH NEIPA, Milkshake IPA || 妃子起笑 Crazy Concubine (v4.4), Orange Glow (v2.6), Strange Haze (v1.1), Milkshake IPA (v1)&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The data ==&lt;br /&gt;
[[File:Homebrew_og_fg_by_batch.png|900px|center|Original and final gravity of every batch]]&lt;br /&gt;
The first six batches had no written recipe and wild original gravities (up to 1.110 — a 10.5% &amp;quot;IPA&amp;quot;). From #0007 on, every batch follows a versioned recipe and the OG has been pulled back into the 1.050–1.076 range the recipes are designed for.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_abv_attenuation.png|900px|center|ABV and apparent attenuation per batch]]&lt;br /&gt;
Attenuation went from ~71% (early batches, under-pitched, no oxygen) to a consistent 81–88% once pure-oxygen pitching, yeast nutrient and modern yeast strains were adopted. #0011&#039;s 62% is not a stuck fermentation: it is a kit recipe with 50 g of unfermentable lactose, and it landed exactly on the kit&#039;s printed OG 1.052 / FG 1.020.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_gravity_curves.png|900px|center|Fermentation curves of four batches]]&lt;br /&gt;
The single most useful chart. With one pack of Verdant and no oxygen (#0006) fermentation drags on for 12 days. With LalBrew Pomona, or two packs plus pure oxygen, ~90% of the sugar is gone within &#039;&#039;&#039;2 days&#039;&#039;&#039;. Any schedule written as &amp;quot;dry hop on Day 3&amp;quot; is simply wrong for these yeasts.&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_plan_vs_actual.png|800px|center|Planned vs actual timing of key events]]&lt;br /&gt;
Since #0009 each batch starts with a one-line written plan (dry-hop day and gravity, green-light day, transfer day, expected FG). The log then measures the error. The first two plans were badly off — fermentation finished 8–12 days early — and each new plan is corrected from the previous error. #0011&#039;s DH1, DH2 and FG all landed within a day of plan (1.0, 0.8 and 0.8 days early).&lt;br /&gt;
&lt;br /&gt;
[[File:Homebrew_ph.png|900px|center|Mash pH and finished-beer pH]]&lt;br /&gt;
pH turned out to be the hidden cause of &amp;quot;too bitter&amp;quot; — and, in #0011, of &amp;quot;slightly sour&amp;quot;. The same tap water gave a mash pH of 4.76 in one batch and 5.81 in another, so the correction has to go in either direction. #0011&#039;s mash was corrected to 5.37, but the wort measured 4.83 after the boil and was not corrected again; the beer finished at 3.94. In both #0008 (4.75 → 3.83) and #0011 (4.83 → 3.94) pH fell about &#039;&#039;&#039;0.9&#039;&#039;&#039; during fermentation, not the textbook 0.3–0.5 — this soft, low-alkalinity water has almost no buffer.&lt;br /&gt;
&lt;br /&gt;
== Batch log ==&lt;br /&gt;
{| class=&amp;quot;wikitable sortable&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
! # !! Date !! Beer / version !! Yeast !! OG → FG !! ABV !! Result / key lesson&lt;br /&gt;
|-&lt;br /&gt;
| 0001 || 2025-12-03 || DDH Hazy IPA || LalBrew New England || 1.062 → 1.018 || 5.8% || Bottle-conditioned. Lesson: add priming sugar, don&#039;t refrigerate during bottle conditioning.&lt;br /&gt;
|-&lt;br /&gt;
| 0002 || 2025-12-28 || Hazy IPA || Mauribrew Weiss || 1.040 → ? || – || Poor mash efficiency; dry-hop order wrong; forgot to measure FG before bottling.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-1 || 2026-02-26 || IPA || Mauribrew ALE514 || 1.050 → 1.010 || 5.3% || Flat aroma and taste; too much priming sugar, vegetal smell.&lt;br /&gt;
|-&lt;br /&gt;
| 0003-2 || 2026-02-28 || IPA + 900 g oats || ALE514 || 1.056 → 1.016 || 5.3% || Under-pitched; bottled while airlock still active → gushers on opening.&lt;br /&gt;
|-&lt;br /&gt;
| 0004 || 2026-03-15 || Milkshake IPA || LalBrew New England || 1.110 → 1.030 || 10.5% || Lid not sealed (airlock silent 36 h); bottled while still fermenting; weak carbonation, slightly sour.&lt;br /&gt;
|-&lt;br /&gt;
| 0005 || 2026-04-23 || Milkshake IPA &amp;quot;light&amp;quot; || Verdant IPA || 1.102 → 1.030 || ~9.5% || &#039;&#039;&#039;Failed.&#039;&#039;&#039; Bittering hops 22 g instead of 6 g; bottles never carbonated — at ~9.5% ABV the yeast was too stressed to referment, and the cold crash removed what was left. Most poured away.&lt;br /&gt;
|-&lt;br /&gt;
| 0006 || 2026-05-09 || Hazy IPA || Verdant IPA || 1.062 → 1.008 || 7.1% || Second dry hop at Day 9 triggered &#039;&#039;&#039;hop creep&#039;&#039;&#039; (1.014 → 1.008). Clean, bitterness fine, aroma weak. First keg + forced carbonation.&lt;br /&gt;
|-&lt;br /&gt;
| 0007 || 2026-06-12 || Orange Glow v2.3 || Verdant IPA ||1.080 → 1.015 || 8.5% || First versioned recipe. Not hot, bitterness right after 3 weeks, but aroma lacking and colour unappealing.&lt;br /&gt;
|-&lt;br /&gt;
| 0008 || 2026-06-25 || Orange Glow v2.4 || LalBrew Pomona || 1.090 → 1.016 || 9.7% || Pomona hit 1.016 by Day 3 — dry hop missed its window. Finished pH 3.83 → harsh bitterness drowned the juice.&lt;br /&gt;
|-&lt;br /&gt;
| 0009 || 2026-07-29 || Orange Glow v2.5 &amp;quot;clean&amp;quot; || Pomona || 1.050 → 1.006 || 5.8% || Boiled off 4–5 L, OG shot to 1.110, over-diluted back to 1.050. pH 4.30 — bitterness finally right. Thin body, under-carbonated (keg fridge is 7–8 °C, not 3 °C). Passable daily hazy.&lt;br /&gt;
|-&lt;br /&gt;
| 0010 || 2026-08-26 || Strange Haze v1 (blueberry vanilla) || Verdant ×2 || 1.070 → 1.012 || 7.6% || Pre-boil OG check caught an over-concentrated wort in time. No harshness, no diacetyl — but not enough juice, sweetness, purple colour or vanilla; grainy from 20% oats.&lt;br /&gt;
|-&lt;br /&gt;
| 0011 || 2026-10-08 || 妃子起笑 v4.3 (PB Craft kit) || Weiss || 1.052 → 1.020 || 4.2% || Clean process — no diacetyl, oxidation or harsh bitterness, no bottle bombs — but weak aroma, thin &amp;quot;watery&amp;quot; body, pale (3–4 SRM) and slightly sour (pH 3.94). Every flaw traced to three fixable numbers: post-boil pH not corrected, CaCl₂ only 0.5 g, both dry hops added while still fermenting. Led to v4.4 and to recipe-wide changes.&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== The beers ==&lt;br /&gt;
; 妃子起笑 · Crazy Concubine — Hazy Pale Ale (OG 1.052, ~5%)&lt;br /&gt;
: The first recipe. v1&#039;s second dry hop went in after fermentation had finished and hop creep ate the body (FG 1.014 → 1.008, 5.8% instead of 5%). v3 fixed it with zero new cost — one dry hop during active fermentation, cooler mash, pure-oxygen pitch, no cold crash. v4 added a 78 °C whirlpool for front-end juiciness. #0011 brewed the PB Craft kit exactly as printed (v4.3) and hit the kit&#039;s numbers, which showed the kit itself is the limit. v4.4: 10 L instead of 11.5 L (OG back to 1.060), CaCl₂ 3 g, a post-boil pH check, and the dry hops swapped — Idaho Gem first, Citra + Amarillo last. #0002, the same kit, had used that &amp;quot;wrong&amp;quot; order by mistake — and had the lychee aroma #0011 lacked.&lt;br /&gt;
; Orange Glow — DDH NEIPA (OG ~1.066, ~7%)&lt;br /&gt;
: The main development line: v1 → v2.6 across three brews. Swapped glucose for DME, Extra Pale for Golden Promise, Verdant for Pomona, added ALDC and Fermaid-O, moved to a split dry hop. #0009 drifted to a pale 5.8%; v2.6 brings back Munich for the orange-gold colour and uses a Strata / Mosaic / Nelson / Peacharine hop bill.&lt;br /&gt;
; Strange Haze — Imperial Milkshake IPA (blueberry vanilla)&lt;br /&gt;
: Built around a pre-printed label (9% ABV, purple nebula), so the fruit is blueberry for colour. v1 was &amp;quot;neither here nor there&amp;quot;; v1.1 cuts bitterness, adds lactose back, raises mash to 69–70 °C and splits 800 g of blueberry between mid-fermentation and post-fermentation.&lt;br /&gt;
; Milkshake IPA — low residual sugar (6 L)&lt;br /&gt;
: Body from oat/wheat protein instead of lots of lactose; unsweetened mango + passion fruit purée. Not yet brewed under this recipe.&lt;br /&gt;
&lt;br /&gt;
== Label and logo ==&lt;br /&gt;
The beers go out under the name &#039;&#039;&#039;piz brewing&#039;&#039;&#039;. The logo started from our dog, became a pencil sketch of just the ears and nose, and was finished with AI image generation. The labels were made the same way: many drafts generated and refined until one fit the beer.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_logo_dog.jpg|The original&lt;br /&gt;
File:Homebrew_logo_sketch.jpg|Pencil sketch: ears and nose&lt;br /&gt;
File:Homebrew_logo_final.jpg|Final logo&lt;br /&gt;
File:Homebrew_label_strange_haze.jpg|Strange Haze label, made with AI image generation&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== What I learned ==&lt;br /&gt;
=== Fermentation ===&lt;br /&gt;
* &#039;&#039;&#039;Dry hop by gravity, not by the calendar.&#039;&#039;&#039; The first dry hop goes in when gravity has dropped about half-way from OG to FG, while yeast is still active (biotransformation + it mops up hop-creep sugar). With fast yeast that can be Day 1, so start measuring every 12 h from Day 1.&lt;br /&gt;
* &#039;&#039;&#039;Split dry hop.&#039;&#039;&#039; DH1 during active fermentation for biotransformation; DH2 only after fermentation has stopped, so the CO₂ doesn&#039;t scrub the aroma away. After DH2, confirm gravity is stable again (second wave of hop creep). #0011 put both dry hops in while it was still fermenting, and the empty glass smelled only of malt.&lt;br /&gt;
* &#039;&#039;&#039;The last dry hop is the aroma you smell.&#039;&#039;&#039; It soaks latest, is scrubbed least and has the least time to oxidise, so put the signature hop (e.g. Citra) there, loose rather than in a bag. Only bag the dry hop you plan to pull out.&lt;br /&gt;
* &#039;&#039;&#039;Hop creep is real.&#039;&#039;&#039; Hop enzymes break down dextrins, so a late dry hop restarts fermentation: thinner body, higher ABV, and in bottles — gushers or bombs.&lt;br /&gt;
* &#039;&#039;&#039;&amp;quot;Green light&amp;quot; before packaging:&#039;&#039;&#039; two gravity readings ~48 h apart that differ by ≤ 0.001, &#039;&#039;&#039;and&#039;&#039;&#039; at least 3 days of diacetyl rest at a beer temperature of 20–22 °C (water bath 21.5 °C). Since 2026-09 the rest time replaces the forced diacetyl test: below 18 °C the yeast barely cleans up diacetyl, above 24 °C it only adds fusel alcohols, so 3 days at 20–22 °C is the guarantee. After DH2 or fruit, hop creep can make new diacetyl, so hold 20–22 °C for another 48 h and re-check gravity before cold-crashing. #0011 was packaged before creep had stopped and still had no butter flavour. If a beer does taste buttery, warm the keg back to room temperature for 3 days. Never package on a fixed day count — #0003-2, #0004 and #0008 all did, and all had carbonation problems.&lt;br /&gt;
* &#039;&#039;&#039;Pure oxygen 60–90 s before pitching&#039;&#039;&#039; lets a single dry pack ferment a 1.07+ wort; no starter needed. After pitching, oxygen is the enemy.&lt;br /&gt;
* Fermenting beer runs 1–2 °C above the water bath during the first 48 h. Set the bath ~1.5 °C below the target beer temperature at peak, then raise it to 21.5 °C for the diacetyl rest.&lt;br /&gt;
&lt;br /&gt;
=== Brew day ===&lt;br /&gt;
* &#039;&#039;&#039;Check pre-boil gravity and volume before boiling.&#039;&#039;&#039; Weigh the pot (total − empty pot weight ≈ litres), measure pre-boil gravity, and compute the target end volume = pre-boil points × litres ÷ target points. Boil to that volume, not to a timer. #0009 boiled off 4–5 L (normal is 1–2 L) and hit 1.110.&lt;br /&gt;
* &#039;&#039;&#039;Boil gently, lid off.&#039;&#039;&#039; Hard boiling over-concentrates bitterness and scorches. Excess evaporation turned out to be a big part of the &amp;quot;nasty bitterness&amp;quot; in #0007/#0008.&lt;br /&gt;
* &#039;&#039;&#039;Measure mash pH, not water pH, and be ready to go both ways.&#039;&#039;&#039; Sample 10–15 min after mash-in, cool to 20–25 °C, target 5.2–5.4. Keep both baking soda and phosphoric/lactic acid on hand. This soft, low-alkalinity water swings hard: ~1 g baking soda ≈ +0.67, 1 ml phosphoric ≈ −0.28 — use half steps and re-measure. Calibrate the pH pen every brew day and store the electrode wet.&lt;br /&gt;
* &#039;&#039;&#039;Post-boil pH is a second, separate checkpoint.&#039;&#039;&#039; A long boil concentrates the acidity too: #0011 boiled 14.55 L down to 11.5 L and dropped from 5.37 to 4.83. Correct the cooled wort to 5.2–5.4 before pitching; with a ~0.9 drop in fermentation that lands the beer at ~4.4. (#0010 split the same 2 g of baking soda between mash and post-boil and had no sourness.)&lt;br /&gt;
* &#039;&#039;&#039;Salt: CaCl₂·2H₂O 0.2 g per litre of total brewing water&#039;&#039;&#039; (≈ Ca 54 / Cl 96 ppm), no gypsum. The filtered water has almost no minerals; #0011&#039;s 0.5 g gave ~17 ppm chloride and ~9 ppm calcium, and the beer tasted like &amp;quot;sweet water&amp;quot;.&lt;br /&gt;
* Pale-malt grists (Extra Pale/Pilsner) tend to mash high (~5.7); grists with Munich plus gypsum tend to mash low.&lt;br /&gt;
&lt;br /&gt;
=== Bitterness and aroma ===&lt;br /&gt;
* &#039;&#039;&#039;Fix bitterness at the source, never by cutting hops.&#039;&#039;&#039; The culprits were pH (finished pH 3.83 = throat-biting), gypsum (sulfate sharpens bitterness), over-boiling, hops left in too long, and drinking too young. Bittering hops go in a bag, on time, and come out after the boil; dry-hop bags come out after 2–3 days.&lt;br /&gt;
* &#039;&#039;&#039;Whirlpool and dry hop do different jobs.&#039;&#039;&#039; Whirlpool (75–78 °C) gives juice on the palate; dry hop gives the aroma in the nose. Cutting either one loses something the other can&#039;t replace.&lt;br /&gt;
* Aging softens bitterness but kills hop aroma at the same time — fix bitterness at the source and drink fresh (2–3 weeks) instead.&lt;br /&gt;
* 0.5 kg oats + 1 kg wheat in 4.5 kg of grain is already enough haze; 1 kg of oats (20%) tasted grainy.&lt;br /&gt;
* &#039;&#039;&#039;Lactose: go big or skip it.&#039;&#039;&#039; 50 g in 11.5 L (4.3 g/L) is far below the ~20–25 g/L you can taste — it added nothing.&lt;br /&gt;
* &#039;&#039;&#039;A kit guarantees beer, not good beer.&#039;&#039;&#039; The PB Craft kit has no whirlpool hops, no specialty malt, and says nothing about salt or pH. Brewed exactly as printed (#0011), it gives a pale, thin, low-aroma beer; those four things have to be added.&lt;br /&gt;
&lt;br /&gt;
=== Packaging ===&lt;br /&gt;
* &#039;&#039;&#039;Keg + forced carbonation&#039;&#039;&#039; is now the default (#0009 on). Bottle conditioning kept failing: too much sugar (#0003-1), bottling before fermentation finished (#0003-2, #0004), and #0005 at ~9.5% ABV, where the yeast could not referment.&lt;br /&gt;
* &#039;&#039;&#039;Cold crash is now standard&#039;&#039;&#039; (since #0011): 2–4 °C for 24–48 h before transfer. #0006–#0011 never really cold-crashed, and all six got the same notes — weak aroma, &amp;quot;don&#039;t know what I&#039;m drinking&amp;quot;, grainy. Suspended yeast fits that picture; it is still a hypothesis, to be tested by drinking #0011&#039;s keg (no yeast) next to its bottles (live yeast). For bottle conditioning after a cold crash: under 7% ABV enough yeast remains; above 8%, add 1–2 g of an alcohol-tolerant yeast (e.g. EC-1118) at bottling.&lt;br /&gt;
* &#039;&#039;&#039;Don&#039;t package on a falling gravity with sulfite in the keg.&#039;&#039;&#039; #0011 was packaged at 1.020 while still dropping: no metabisulfite for an early package (keep the yeast alive), and halve the priming sugar for bottles (4.5 g/L — it carbonated without gushing).&lt;br /&gt;
* &#039;&#039;&#039;Never bottle the gravity samples.&#039;&#039;&#039; #0011&#039;s test bottle made from saved samples gushed and measured pH 3.5 — lactic infection from the open sample cup, not the fermenter (the normal bottles were 3.94). Drink samples on the spot; for a control bottle, fill one more from the main beer.&lt;br /&gt;
* Low-oxygen transfer: purge the keg with CO₂ three times, fill from the bottom through the liquid post, minimum headspace.&lt;br /&gt;
* Antioxidant in the keg: &#039;&#039;&#039;ascorbic acid 0.3 g + potassium metabisulfite 0.25 g per 10 L, always as a pair.&#039;&#039;&#039; Ascorbic acid alone makes H₂O₂, which with trace iron/copper (Fenton reaction) oxidises the beer worse than nothing. The chemicals are ~20% of the protection; the purging is the other 80%.&lt;br /&gt;
* &#039;&#039;&#039;Measure the fridge.&#039;&#039;&#039; The keg fridge is actually 7–8 °C, so 2.4–2.5 volumes needs ~15 PSI, not the textbook 11–12 PSI. #0009 and #0010 were both under-carbonated for this reason. A small 3 L keg also absorbs CO₂ slowly — #0011 was only ~70–80% carbonated after 5 days at pressure; lay it down and shake for 2–3 minutes.&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_mini_keg_regulator.jpg|Mini keg with CO₂ regulator: carbonation is set by pressure, not sugar&lt;br /&gt;
File:Homebrew_mini_keg_tap.jpg|Mini keg with a wooden tap handle&lt;br /&gt;
File:Homebrew_bottles_box.jpg|A case of bottles filled from the keg&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
=== Method ===&lt;br /&gt;
* Write the plan down before brewing (one line: dry-hop day and gravity, green light, transfer, FG) and measure the error afterwards. The predictions are converging batch by batch.&lt;br /&gt;
* Log everything with a timestamp — gravity with date and time, pH with temperature. Most of the lessons above only became visible once the numbers were next to each other.&lt;br /&gt;
* Change one or two things per version. The recipe history (Crazy Concubine v1→v4.4, Orange Glow v1→v2.6) reads like a commit log: each version cites the batch that justified it.&lt;br /&gt;
&lt;br /&gt;
== Equipment ==&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;200&amp;quot;&amp;gt;&lt;br /&gt;
File:Homebrew_first_kit.jpg|Brew day #0001 (Dec 2025): kettle, fermenter, grain and hops&lt;br /&gt;
File:Homebrew_hop_pellets.jpg|Hop pellets, straight from the vacuum pack&lt;br /&gt;
File:Homebrew_mash.jpg|Mashing #0008, thermometer in the lid&lt;br /&gt;
File:Homebrew_ir_thermometer.jpg|IR thermometer check on the kettle&lt;br /&gt;
File:Homebrew_cooling_rig.jpg|Home-built cooling rig: thermoelectric cooler, power supply and thermostat, pumping cold water through the fermenter&lt;br /&gt;
File:Homebrew_o2_co2_cylinders.jpg|CO₂ (left) and pure O₂ (right) cylinders with regulators&lt;br /&gt;
File:Homebrew_fermenter_krausen.jpg|#0008 on Day 3: krausen, fermenting in the water bath&lt;br /&gt;
File:Homebrew_hydrometer.jpg|Gravity sample from #0009&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
* Kettle + fermenter in a temperature-controlled water bath; weigh-the-pot volume estimation&lt;br /&gt;
* Pure O₂ cylinder with diffusion stone; refractometer + hydrometer; pH pen (4.01 / 7.00 buffers)&lt;br /&gt;
* Inkbird IHT-1P instant-read thermometer (from #0010)&lt;br /&gt;
* 8 L keg on external CO₂; 128 oz mini keg for in-keg conditioning&lt;br /&gt;
* Additives: Fermaid-O (2.5 g at pitch + 2.5 g at ~24 h), Maturex L ALDC 0.2–0.3 ml, CaCl₂&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_abv_attenuation.png&amp;diff=828</id>
		<title>檔案:Homebrew abv attenuation.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_abv_attenuation.png&amp;diff=828"/>
		<updated>2026-10-21T11:20:00Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Homebrew abv attenuation.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Homebrewing charts (generated from brew-log by plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_og_fg_by_batch.png&amp;diff=827</id>
		<title>檔案:Homebrew og fg by batch.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_og_fg_by_batch.png&amp;diff=827"/>
		<updated>2026-10-21T11:20:00Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Homebrew og fg by batch.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Homebrewing charts (generated from brew-log by plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_ph.png&amp;diff=826</id>
		<title>檔案:Homebrew ph.png</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:Homebrew_ph.png&amp;diff=826"/>
		<updated>2026-10-21T11:20:00Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Tommy上傳了檔案:Homebrew ph.png的新版本&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
Homebrewing charts (generated from brew-log by plots.py)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=LG_WE16HW_dryer_teardown_and_cleaning&amp;diff=825</id>
		<title>LG WE16HW dryer teardown and cleaning</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=LG_WE16HW_dryer_teardown_and_cleaning&amp;diff=825"/>
		<updated>2026-10-21T09:13:49Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Back frame, not front; motor side only lightly dusty; add reassembly photo and final test video&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&#039;&#039;&#039;LG WE16HW dryer teardown and cleaning&#039;&#039;&#039; — taking an LG WE16HW heat-pump dryer apart down to the base to clear out the lint that the door filter does not catch. The drum comes out, the base duct and the heat-exchanger coils are exposed, and everything is vacuumed and brushed clean before reassembly.&lt;br /&gt;
&lt;br /&gt;
[[File:LG_dryer_before_teardown.jpg|600px|center|The LG WE16HW before teardown]]&lt;br /&gt;
&lt;br /&gt;
== Why open it up ==&lt;br /&gt;
A heat-pump dryer does not vent its air outside. It recirculates it through two coils in the base: the &#039;&#039;&#039;evaporator&#039;&#039;&#039; cools the air so the moisture condenses out, then the &#039;&#039;&#039;condenser&#039;&#039;&#039; heats it again before it goes back into the drum. Any lint that gets past the door filter ends up in that loop. It settles in the base duct and mats onto the face of the evaporator, where it blocks the airflow.&lt;br /&gt;
&lt;br /&gt;
== Teardown ==&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;260&amp;quot;&amp;gt;&lt;br /&gt;
File:LG_dryer_drum_removed.jpg|Drum and side panel out, back frame still standing with the base exposed&lt;br /&gt;
File:LG_dryer_base_overview.jpg|The base from above: compressor, drive motor with belt idler and fan, hoses and the coil housing&lt;br /&gt;
File:LG_dryer_motor_compressor.jpg|Drive motor, idler pulley, fan and compressor, seen from under the drum. Some lint on the floor of the base, but nothing like the evaporator coil&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
* Unplug it first.&lt;br /&gt;
* Photograph every connector before unplugging it. The wiring harness has several similar plugs.&lt;br /&gt;
* The copper refrigerant pipes stay connected the whole time. Do not bend them or lever against them.&lt;br /&gt;
&lt;br /&gt;
== How dirty it was ==&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:LG_dryer_base_lint.jpg|Lint packed into the base duct under the drum, all the way to the grille&lt;br /&gt;
File:LG_dryer_evaporator_before.jpg|The evaporator coil: the bottom half buried in a mat of lint, the rest covered in a lint web&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
The motor and compressor side stayed fairly clean, with only a thin layer of lint on the floor of the base (third teardown photo above). The lint collects where the air hits the evaporator coil.&lt;br /&gt;
&lt;br /&gt;
== Before / after ==&lt;br /&gt;
{| style=&amp;quot;margin:0 auto&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
| [[File:LG_dryer_evaporator_before.jpg|480px|Evaporator coil before cleaning]]&lt;br /&gt;
| [[File:LG_dryer_coils_after.jpg|480px|Evaporator and condenser coils after cleaning]]&lt;br /&gt;
|- style=&amp;quot;text-align:center&amp;quot;&lt;br /&gt;
| &#039;&#039;&#039;Before&#039;&#039;&#039; — lint mat on the evaporator&lt;br /&gt;
| &#039;&#039;&#039;After&#039;&#039;&#039; — both coils clean, fins open&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
[[File:LG_dryer_motor_after.jpg|700px|center|Motor area and coil inlet after cleaning]]&lt;br /&gt;
&lt;br /&gt;
== Reassembly and final test ==&lt;br /&gt;
[[File:LG_dryer_before_assembly.jpg|700px|center|Drum back in, ready for the top cover. The control board sits on the left]]&lt;br /&gt;
&lt;br /&gt;
&#039;&#039;&#039;[[Media:LG_dryer_final_test.mp4|▶ Final test run after reassembly (video, 8 s)]]&#039;&#039;&#039;&lt;br /&gt;
&lt;br /&gt;
== Cleaning notes ==&lt;br /&gt;
* Coil fins are thin aluminium and bend easily. Use a soft brush and a vacuum with a brush nozzle, and always move along the fins, never across them.&lt;br /&gt;
* Get the thick mat off by hand first. It comes away in sheets, and the fine lint underneath is then easy to vacuum.&lt;br /&gt;
* Vacuum the base duct, the grille and the drum seals before putting the drum back.&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_final_test.mp4&amp;diff=824</id>
		<title>檔案:LG dryer final test.mp4</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_final_test.mp4&amp;diff=824"/>
		<updated>2026-10-21T09:13:48Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
LG WE16HW dryer: before reassembly photo (resized, metadata stripped) and final test video&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
LG WE16HW dryer: before reassembly photo (resized, metadata stripped) and final test video&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_before_assembly.jpg&amp;diff=823</id>
		<title>檔案:LG dryer before assembly.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_before_assembly.jpg&amp;diff=823"/>
		<updated>2026-10-21T09:13:48Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
LG WE16HW dryer: before reassembly photo (resized, metadata stripped) and final test video&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
LG WE16HW dryer: before reassembly photo (resized, metadata stripped) and final test video&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E9%A6%96%E9%A0%81&amp;diff=822</id>
		<title>首頁</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E9%A6%96%E9%A0%81&amp;diff=822"/>
		<updated>2026-10-21T09:01:25Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​Main page: add LG dryer cleaning under Hobbies&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;220&amp;quot;&amp;gt;&lt;br /&gt;
File:3D_ripples_01.gif|3D LED cube &amp;amp;ndash; ripple effect&lt;br /&gt;
File:MeArm_stm32f4.jpg|MeArm robotic arm driven by STM32F4&lt;br /&gt;
File:Gps_wifiap.jpg|GPS tracker with Wi-Fi AP&lt;br /&gt;
File:Eye_tracker_alexa.jpg|Eye tracker with Alexa and AWS IoT&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
&amp;lt;div style=&amp;quot;display:flex;flex-wrap:wrap;gap:12px;margin:0 0 18px&amp;quot;&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;{{PAGESINCATEGORY:Boards and SoCs|pages}}&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[#Boards and SoCs|boards &amp;amp;amp; SoCs]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;{{PAGESINCATEGORY:Projects|pages}}&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[#Projects|hands-on projects]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;2010&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[RTD1073|shipping products since]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;div style=&amp;quot;flex:1 1 150px;padding:12px 16px;border:1px solid var(--sf-line,#2a3a55);border-radius:8px;background:var(--sf-panel-2,#111c2e)&amp;quot;&amp;gt;&amp;lt;div style=&amp;quot;font-size:2em;font-weight:bold;line-height:1.1;color:var(--sf-cyan,#35e0ff)&amp;quot;&amp;gt;11&amp;lt;/div&amp;gt;&amp;lt;div style=&amp;quot;color:var(--sf-fg-dim,#8a9cb2)&amp;quot;&amp;gt;[[Homebrewing|homebrew batches]]&amp;lt;/div&amp;gt;&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;lt;/div&amp;gt;&lt;br /&gt;
[[File:Experience_platforms.png|frameless|1000px|Hardware I&#039;ve worked on|link=Category:Boards and SoCs]]&lt;br /&gt;
&lt;br /&gt;
[[File:Experience_technologies.png|frameless|1000px|Technologies across my engineering pages|link=Category:Projects]]&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
== About Tommy C. Liu ==&lt;br /&gt;
I have always believed I could flip the world through my own work &amp;amp;mdash; that is where the name of this wiki comes from. For most of my career that meant building things by hand: firmware, boards and products, one line of code at a time.&lt;br /&gt;
&lt;br /&gt;
AI changed that. It carried me past limits I could not step over on my own, and gave me courage, knowledge and speed I never had before. It also brings a new kind of uncertainty: the results are not always what you expect, so judgment matters more than ever.&lt;br /&gt;
&lt;br /&gt;
I have loved craft beer for years. I studied brewing again and again and brewed it in my head many times, but I never had the courage to actually start. With AI I finally had a strong partner to think it through and work on it together &amp;amp;mdash; so I began. Eleven batches of hazy IPA in under a year, every one logged, measured and learned from (see [[Homebrewing]]).&lt;br /&gt;
&lt;br /&gt;
My role has shifted from &#039;&#039;&#039;creator and maintainer&#039;&#039;&#039; to &#039;&#039;&#039;supervisor&#039;&#039;&#039; &amp;amp;mdash; deciding what to build, setting the direction, and reviewing what comes back. That is leadership, and anyone who works with AI is already doing it. AI has not taken our skills away; it has promoted us. I call it the &#039;&#039;&#039;AI promotion&#039;&#039;&#039;.&lt;br /&gt;
&lt;br /&gt;
So the question I ask myself now is: &#039;&#039;where do I want to go next?&#039;&#039;&lt;br /&gt;
&lt;br /&gt;
How about you &amp;amp;mdash; have you been promoted yet?&lt;br /&gt;
&lt;br /&gt;
----&lt;br /&gt;
* Welcome to discuss outsourcing project&lt;br /&gt;
** FW/SW development&lt;br /&gt;
** Training course  &lt;br /&gt;
* E-mail: tommy.using@gmail.com&lt;br /&gt;
* LinkedIn: https://www.linkedin.com/in/chia-chuan-liu-7706b473/&lt;br /&gt;
&lt;br /&gt;
=== Hobbies ===&lt;br /&gt;
* [[Homebrewing]] &amp;amp;mdash; hazy IPA / NEIPA batches, brew log and lessons learned&lt;br /&gt;
* [[LG WE16HW dryer teardown and cleaning]] &amp;amp;mdash; heat-pump dryer taken apart to clear lint from the base and coils, with before/after photos&lt;br /&gt;
&lt;br /&gt;
== Experience ==&lt;br /&gt;
=== Projects ===&lt;br /&gt;
Hands-on builds and application notes &amp;amp;mdash; [[:Category:Projects|all {{PAGESINCATEGORY:Projects|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Boards and SoCs ===&lt;br /&gt;
Dev boards, MCUs and SoCs &amp;amp;mdash; [[:Category:Boards and SoCs|all {{PAGESINCATEGORY:Boards and SoCs|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Software and Tools ===&lt;br /&gt;
OS, build and test tooling &amp;amp;mdash; [[:Category:Software and Tools|all {{PAGESINCATEGORY:Software and Tools|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Sensors and Modules ===&lt;br /&gt;
Sensors and peripheral modules &amp;amp;mdash; [[:Category:Sensors and Modules|all {{PAGESINCATEGORY:Sensors and Modules|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== Connectivity ===&lt;br /&gt;
Buses and wireless protocols &amp;amp;mdash; [[:Category:Connectivity|all {{PAGESINCATEGORY:Connectivity|pages}} pages &amp;amp;rarr;]]&lt;br /&gt;
&lt;br /&gt;
=== CI/CD ===&lt;br /&gt;
* [[Firmware CI/CD on GitHub Actions]]&lt;br /&gt;
* [[Jenkins]]&lt;br /&gt;
&lt;br /&gt;
=== Equipment ===&lt;br /&gt;
* Measurement&lt;br /&gt;
** Logic Analyzer&lt;br /&gt;
*** Max Sample Rate 400MHz&lt;br /&gt;
*** 16G stream / 256M buffer&lt;br /&gt;
*** 16 Channels&lt;br /&gt;
** Multimeter&lt;br /&gt;
&lt;br /&gt;
* Computings&lt;br /&gt;
** Linux server - Intel Xeon E3-1231 v3(8 cores)&lt;br /&gt;
*** Intel(R) Xeon(R) CPU E3-1230 v3&lt;br /&gt;
*** 24G RAM&lt;br /&gt;
** Windows desktop&lt;br /&gt;
*** Intel(R) Xeon(R) CPU E3-1231&lt;br /&gt;
** Linux laptop&lt;br /&gt;
*** i7-7600U&lt;br /&gt;
*** 16G RAM&lt;br /&gt;
&lt;br /&gt;
== Reading &amp;amp; sharing ==&lt;br /&gt;
* [[Software version]]&lt;br /&gt;
* [[Software release life cycle]]&lt;br /&gt;
* [[Agent Working Rules]]&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=LG_WE16HW_dryer_teardown_and_cleaning&amp;diff=821</id>
		<title>LG WE16HW dryer teardown and cleaning</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=LG_WE16HW_dryer_teardown_and_cleaning&amp;diff=821"/>
		<updated>2026-10-21T09:01:23Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​New page: LG WE16HW heat-pump dryer teardown and lint cleaning, before/after photos&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;&#039;&#039;&#039;LG WE16HW dryer teardown and cleaning&#039;&#039;&#039; — taking an LG WE16HW heat-pump dryer apart down to the base to clear out the lint that the door filter does not catch. The drum comes out, the base duct and the heat-exchanger coils are exposed, and everything is vacuumed and brushed clean before reassembly.&lt;br /&gt;
&lt;br /&gt;
[[File:LG_dryer_before_teardown.jpg|600px|center|The LG WE16HW before teardown]]&lt;br /&gt;
&lt;br /&gt;
== Why open it up ==&lt;br /&gt;
A heat-pump dryer does not vent its air outside. It recirculates it through two coils in the base: the &#039;&#039;&#039;evaporator&#039;&#039;&#039; cools the air so the moisture condenses out, then the &#039;&#039;&#039;condenser&#039;&#039;&#039; heats it again before it goes back into the drum. Any lint that gets past the door filter ends up in that loop. It settles in the base duct and mats onto the face of the evaporator, where it blocks the airflow.&lt;br /&gt;
&lt;br /&gt;
== Teardown ==&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;260&amp;quot;&amp;gt;&lt;br /&gt;
File:LG_dryer_drum_removed.jpg|Drum and side panel out, front frame still standing with the base exposed&lt;br /&gt;
File:LG_dryer_base_overview.jpg|The base from above: compressor, drive motor with belt idler and fan, hoses and the coil housing&lt;br /&gt;
File:LG_dryer_motor_compressor.jpg|Drive motor, idler pulley, fan and compressor, seen from under the drum&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
* Unplug it first.&lt;br /&gt;
* Photograph every connector before unplugging it. The front wiring harness has several similar plugs.&lt;br /&gt;
* The copper refrigerant pipes stay connected the whole time. Do not bend them or lever against them.&lt;br /&gt;
&lt;br /&gt;
== How dirty it was ==&lt;br /&gt;
&amp;lt;gallery mode=&amp;quot;packed&amp;quot; heights=&amp;quot;300&amp;quot;&amp;gt;&lt;br /&gt;
File:LG_dryer_base_lint.jpg|Lint packed into the base duct under the drum, all the way to the front grille&lt;br /&gt;
File:LG_dryer_evaporator_before.jpg|The evaporator coil: the bottom half buried in a mat of lint, the rest covered in a lint web&lt;br /&gt;
&amp;lt;/gallery&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== Before / after ==&lt;br /&gt;
{| style=&amp;quot;margin:0 auto&amp;quot;&lt;br /&gt;
|-&lt;br /&gt;
| [[File:LG_dryer_evaporator_before.jpg|480px|Evaporator coil before cleaning]]&lt;br /&gt;
| [[File:LG_dryer_coils_after.jpg|480px|Evaporator and condenser coils after cleaning]]&lt;br /&gt;
|- style=&amp;quot;text-align:center&amp;quot;&lt;br /&gt;
| &#039;&#039;&#039;Before&#039;&#039;&#039; — lint mat on the evaporator&lt;br /&gt;
| &#039;&#039;&#039;After&#039;&#039;&#039; — both coils clean, fins open&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
[[File:LG_dryer_motor_after.jpg|700px|center|Motor area and coil inlet after cleaning]]&lt;br /&gt;
&lt;br /&gt;
== Cleaning notes ==&lt;br /&gt;
* Coil fins are thin aluminium and bend easily. Use a soft brush and a vacuum with a brush nozzle, and always move along the fins, never across them.&lt;br /&gt;
* Get the thick mat off by hand first. It comes away in sheets, and the fine lint underneath is then easy to vacuum.&lt;br /&gt;
* Vacuum the base duct, the front grille and the drum seals before putting the drum back.&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_coils_after.jpg&amp;diff=820</id>
		<title>檔案:LG dryer coils after.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_coils_after.jpg&amp;diff=820"/>
		<updated>2026-10-21T09:01:22Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_before_teardown.jpg&amp;diff=819</id>
		<title>檔案:LG dryer before teardown.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_before_teardown.jpg&amp;diff=819"/>
		<updated>2026-10-21T09:01:22Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_drum_removed.jpg&amp;diff=818</id>
		<title>檔案:LG dryer drum removed.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_drum_removed.jpg&amp;diff=818"/>
		<updated>2026-10-21T09:01:22Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_evaporator_before.jpg&amp;diff=817</id>
		<title>檔案:LG dryer evaporator before.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_evaporator_before.jpg&amp;diff=817"/>
		<updated>2026-10-21T09:01:22Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_base_lint.jpg&amp;diff=816</id>
		<title>檔案:LG dryer base lint.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_base_lint.jpg&amp;diff=816"/>
		<updated>2026-10-21T09:01:22Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_motor_compressor.jpg&amp;diff=815</id>
		<title>檔案:LG dryer motor compressor.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_motor_compressor.jpg&amp;diff=815"/>
		<updated>2026-10-21T09:01:22Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_motor_after.jpg&amp;diff=814</id>
		<title>檔案:LG dryer motor after.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_motor_after.jpg&amp;diff=814"/>
		<updated>2026-10-21T09:01:22Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
	<entry>
		<id>https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_base_overview.jpg&amp;diff=813</id>
		<title>檔案:LG dryer base overview.jpg</title>
		<link rel="alternate" type="text/html" href="https://tommy0804.ddns.net/index.php?title=%E6%AA%94%E6%A1%88:LG_dryer_base_overview.jpg&amp;diff=813"/>
		<updated>2026-10-21T09:01:22Z</updated>

		<summary type="html">&lt;p&gt;Tommy：​== 摘要 ==
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/p&gt;
&lt;hr /&gt;
&lt;div&gt;== 摘要 ==&lt;br /&gt;
LG WE16HW dryer cleaning photos (resized, metadata stripped)&lt;/div&gt;</summary>
		<author><name>Tommy</name></author>
	</entry>
</feed>