001 — Aged-wine depth¶
WB-ING-2026-001 · status: designed · Disclaimer
Related: Maillard preparations · Skill user guide
Safety
Experimental home procedure. Read the disclaimer first.
Accent archetype, one 4.5 L carboy. ×4.44 for 20 L. Arithmetic from
scripts/prep_calc.py on design-001.json.
1. Request summary¶
What was asked. A beer fermented with WLP645 (B. claussenii), WLP648 (B. bruxellensis Trois Vrai) and Holy Goat bottle dregs — described as neutral and Brett-like — was blended with a ~1.5 L Philly Sour ferment of 1 L red grape juice, 400 g sable grapes, 50 g aronia and 20 g dried elderberry, topped up with ~3 L of the base beer. Two weeks later a further 300 g of sable grapes and 2 g of medium-toast oak chips went in. The brewer's assessment: it lacks the depth and complexity of a fine wine or a good gueuze, and he asked for a simple prepared ingredient to get it there.
Mode used: 3 (base-beer-first), run through mode 2 (target-first). Intake was taken before designing: ~5 L in the carboy and willing to rack off fruit to make space; near final gravity but still slowly fermenting; added sugar acceptable but ABV increase to be limited; pH unmeasured and the beer does not taste sour (Philly Sour underperformed); 2 g of medium-toast chips in for three weeks; no headspace; more acidity and brightness wanted; no bottling deadline.
Two things pushed back on at intake, and where they landed. First, two weeks is too early to judge a beer whose Brett has not begun to express — the brewer's own read on the trajectory was accepted as the governing input. Second, the repeated sable-grape additions were agreed to be the wrong lever: seedless table grapes and clarified juice carry sugar, water and anthocyanin and almost no proanthocyanidin or bound aroma precursor, so each addition diluted the beer without moving either axis the target needs.
Build constraints that followed from the larder. Dried rosehip and dried aronia in place of fresh (none found foraging); dried whole sloe, cracked, in place of bitter apricot kernel; medium-toast chips in place of cubes.
The thesis, in one paragraph¶
Three things separate an aged red wine or a good gueuze from this beer, and none of them is a
flavour you can add. (1) A norisoprenoid pool — β-damascenone, β-ionone, TDN, vitispirane — built
from carotenoid and bound-glycoside precursors that release slowly under acid over months to years.
Table grapes, grape juice, aronia and elderberry are all carotenoid-poor (ingredients.md: Car = L
across the board), so this beer has essentially no precursor pool at all. (2) A furanone and
melanoidin mid-palate — furaneol, maltol, cyclotene — which reads as depth and sweetness without
sugar. Nothing in the beer has been near a Maillard reaction. (3) Polymeric phenolic structure,
which this method only partly reaches; see the honest limit at the end.
So the prep loads precursors, not flavours, and lets the low pH and the open-ended timeline do the conversion. That is the mechanism actual wine uses, and the no-deadline constraint is what makes it the right choice here.
Binding constraint¶
Not nitrogen, and not Brix. It is potassium and beer pH. The brief asks for more acidity. A
conventional Stage B jar based to pH 6.15 would carry ~30 meq/L of potassium malate into a 4.5 L beer
— a +0.8 to +1.7 pH-unit lift (chemistry.md §18) — the exact opposite of what was asked, arriving
alongside a fresh sugar and FAN load.
Everything downstream follows:
- Stage B runs as a furanone jar at pH 4.4, not the pyrazine jar at 6.15. Base drops from ~7 g to 1.17 g. Furanones are also the right compounds for a wine target — cocoa and roast would read as chocolate cherry stout.
- The inversion step is skipped. 4.5 g of sucrose in the jar; dosing phosphoric to 2.9 then basing back to 4.4 would spend potassium to undo acid. Count RS only, count sucrose as fermentable. Deliberate deviation from the default workflow.
- Stage C is unbased and carries 51.8 meq of free fruit acid in.
- Hibiscus is dropped. With sloe flesh in Stage C alongside the aronia and rosehip, the unbased jar carried ~76 meq — about 17 meq/L of free acid, which at a 20–40 meq/L buffer is a 0.4–0.85 pH-unit fall. From a Brett beer likely near 3.4 that lands between harsh and slow-Brett. Sloe supplies the acid hibiscus was there for and supplies VH polyphenol, which hibiscus does not and which is this design's weakest axis. Without it: 11.5 meq/L, more brightness than the hibiscus build delivered, still controlled. Cost: some HCA loading, partly covered by aronia and sloe.
- Net beer-side: +2.6 meq/L of base against 11.5 meq/L of free acid. The beer gets brighter.
Second constraint, softer: fermentable load. The sloe flesh and second rosehip added ~13 g of
Stage C reducing sugar, and Stage C sugar barely reacts. The first pass of this build tripped three
warnings — FAN 33 mg N/L, worst-case fermentable 19.8 g/L, and the sizing anchor. Per
process-inputs.md, glycine comes down before sugar when a design is over the FAN threshold, so
glycine went 1.4 → 1.0 g and date 55 → 45 g. All three clear.
2. Formula, by stage and jar — one 4.5 L carboy, accent archetype¶
Stage B — furanone / norisoprenoid jar¶
The mid-palate and the precursor pool: furaneol, maltol, cyclotene, furfural, β-ionone, β-damascenone, phenylacetaldehyde → 2-phenylethanol. Caramel-jam depth, dried apricot, violet, honey, rose.
| Ingredient | g | Role |
|---|---|---|
| Date, Medjool-type, stoned | 45 | RS backbone; lowest-asparagine choice (§22); near-zero TA |
| Dried apricot, unsulphited | 45 | carotenoid VH + glycoside H |
| Dried rosehip shells | 48 | carotenoid extreme; pentose for furfural/furaneol |
| Rehydration water | 58 mL nominal — start at 30 | see section 6 |
| Glycine | 1.0 | bulk reactive nitrogen |
| L-phenylalanine | 0.8 | honey/rose lever |
| Potassium bicarbonate | 1.17 (dose 0.93 first) | to pH 4.4 |
196 g in → reduce to ~95–110 g of paste (refractometer decides).
Stage C — structure and acid jar¶
The grip, the brightness and the funk loading: catechin/epicatechin from acid-catalysed proanthocyanidin depolymerisation, protocatechuic acid, 4-vinylguaiacol / 4-vinylphenol, oak lactone, vanillin, eugenol — plus the second, acid-catalysed norisoprenoid route.
| Ingredient | g | Role |
|---|---|---|
| Dried aronia, unsweetened | 27 | polyphenol VH; depolymerised, not raw |
| Dried sloe flesh, stones removed | ~23 | polyphenol VH + high acid; replaces hibiscus |
| Dried rosehip shells | 25 | carotenoid in an unbased acid jar — the β-damascenone route |
| Toasted oak, medium, chips | 4.5 | vanillin, oak lactone, furfural, cyclotene, eugenol |
| Rehydration water | 173 mL | restores fresh-equivalent mass |
252 g in. Unbased, native pH ~3.46.
Stage A2 — kernel hydrolysis¶
Marzipan and kirsch — the note that reads as old kriek and wine cellar. Zero sugar cost.
| Ingredient | g |
|---|---|
| Cracked sloe stones (from the 30 g of whole sloe) | ~7 |
| Water | 56 mL (8× stone mass) |
Stage D¶
Vented, strained A2 liquor into the beer. No fruit — the grapes have been added twice already.
Stages A1 and T¶
Both skipped. A1: the pentose comes from the hob reduction; a Lallzyme hold buys little. T: the target is depth, not florals. Reconsider T at bottling only if the nose reads flat.
3. Nitrogen¶
| Glycine | 1.0 g (13.30 mmol) |
| L-phenylalanine | 0.8 g (4.84 mmol) |
| Collagen | 0 g — off by default (v4), no reason to reinstate |
| Protease step | not run |
| Reactive amine | 18.14 mmol |
| Sugar : amine | 16.6 : 1 molar |
| 10:1 rule | does not apply (glycine-only jar, no THP constraint) |
16.6:1 is above the 10–14 preference band, deliberately. The extra protein from sloe, aronia and rosehip pushed added FAN to 33 mg N/L at the original glycine dose — over the warn. Cutting glycine is the sanctioned lever and brings FAN to 27. The cost is some furanone depth; it is the right trade in a live secondary.
4. Base dose¶
| Stage B | Stage C | |
|---|---|---|
| Mixture native pH | 3.5 (estimated from ingredient pH + TA, Low) | 3.46 (est., Low) |
| Target pH, cold | 4.4 | unbased, at native pH |
| Fruit TA in jar | 40.6 meq | 51.8 meq |
| Full dose | 1.17 g KHCO₃ (11.7 meq) | 0 |
| 80 % first pass | 0.93 g | — |
| Effective factor vs fruit TA | 0.29 | — |
Procedure. Add 0.93 g in two or three portions (it foams), stir, measure cold on a thin smear with the flat/spear probe, titrate the remainder to 4.4. Record predicted vs measured — first entry in the pH calibration table.
Two caveats the solve flags itself: date's native pH 5.8 sits near malic pKa₂ so the anchored solve falls back for that term (its TA is 1.35 meq — immaterial); and every TA figure is proxied, so the dose already carries the 0.9 discount. The unanchored v3 method would have said 1.85 g, 58 % high.
Stage C is unbased, and here that is a design objective, not a default: its 51.8 meq of free acid is the brightness lever.
5. Ingredients list — shop and check inventory¶
To buy or confirm in stock¶
| Item | Need | Buy | Notes |
|---|---|---|---|
| Dried rosehip shells | 73 g | 100 g | 48 g B + 25 g C. Strain the hairs. If the shells look brown rather than deep orange-red, carotenoid is depleted — drop to 60 g total |
| Dried sloe, whole, stone-in | 30 g | 50 g | Stone fraction ~23 % is Med — buy margin and weigh after cracking |
| Dried aronia | 27 g | 50 g | Check the label: unsweetened only. Sugar-infused dried berries break the whole sugar accounting |
| Dried apricot | 45 g | 100 g | Unsulphited only — sulphite inhibits both Maillard and Brett |
| Date, Medjool-type | 45 g | 100 g | Stoned weight. Prefer Medjool over Deglet (less sucrose) |
| Toasted oak chips, medium | 4.5 g | — | In stock |
| Glycine | 1.0 g | — | In stock |
| L-phenylalanine | 0.8 g | 50 g | Agreed in design 004; buy if not yet held |
| Potassium bicarbonate | 1.17 g | — | In stock |
| Lallzyme Beta | 225 mg | — | In stock. Beer-side, see the closing note |
| Lactic acid 88 % | to taste | — | In stock. Contingency only, see section 13 |
Not needed for this build: hibiscus, bitter apricot kernel, sweet almond, calendula, shiitake, phosphoric acid, collagen, papain, neutral spirit.
Water additions¶
| Where | Volume | Purpose |
|---|---|---|
| Stage B rehydration | 58 mL nominal — start at 30 mL | soften apricot and rosehip enough to blend. Most comes back off on the hob, so add as little as will work |
| Stage C rehydration | 173 mL | restore dried aronia, sloe flesh and rosehip to fresh-equivalent mass. Not reduced — Stage C has no Brix target |
| Stage A2 kernel | 56 mL | 8× the stone mass, per §17's 5–10× band |
| Total | ~287 mL |
Kit¶
1 mg scale · kitchen scale · jam refractometer 58–90 °Brix · pH meter with flat or spear probe, calibrated at 4.0 and 7.0 · stovetop pressure cooker · 2 Kilner/Mason jars · small jar or pan + thermometer for the A2 bath · ice bath · mallet or pestle for the sloe stones · sieve for the rosehip hairs.
6. Cook¶
| Step | Detail |
|---|---|
| Crack the sloe first | Mallet the 30 g of dried whole sloe. Separate the stones from the flesh and weigh both. Expect ~7 g stone, ~23 g flesh. If your stone fraction differs, scale section 10 |
| A2 hold | Cracked stones + 56 mL water, 40–50 °C, covered (not sealed), 2 h. pH ~5.5, low-acid → the §21 2 h cap applies, and it is a cap not a target |
| A2 check | The water should smell strongly of bitter almond, the drained stone much less. If the stone still smells strong, hydrolysis was incomplete — retention is worse than 0.58 |
| A2 vent | Uncovered, 10 min at 90–95 °C, stirring. Then strain: keep the liquor, discard the solids |
| Inversion | Not run. 4.5 g of sucrose in the jar; dosing phosphoric to 2.9 and basing back to 4.4 would spend potassium to undo acid. No phosphoric acid in this design |
| Stage B rehydrate | Soak apricot + rosehip warm in the minimum water, ≤2 h, then blend to a paste |
| Stage B reduce | Read the refractometer before reducing anything. Reduce on the hob only if under 60 °Brix. Stop at 60–64, not 65–68: the glycine, Phe and KHCO₃ go in cold afterwards and add 2–3 points, and §5 notes amino acids and potassium salts depress a_w more efficiently per Brix point than sugar |
| Stage B base | After reduction, cold, per section 4. Never seal a based, unreduced jar |
| Stage C | Rehydrate aronia, sloe flesh and rosehip in 173 mL, add the oak chips, jar. No reduction, no base |
| Headspace | 10–15 % (fill to 85–90 %), lids finger-tight |
| Cook | 121 °C, time from full pressure. Stage C 25 min, then depressurise and remove. Reseal and run Stage B 100 min. Jars standing in 3–5 cm of water; never paste on the base |
| Cooling | Stage C: ice bath, fast, then ≤3 °C, use within 5 days. Stage B: slow, sealed |
| Hot vent | Both jars, 5–10 min open and stirring while hot, before blending. Non-optional — section 15 |
The prep_calc.py figures for Stage B: 196 g in, ~86 g of water present, lower-bound Brix 31, target
mass ~95 g. The stated "remove up to 100 g" exceeds the water that exists — that is the upper
bound announcing that the jar starts near band, not an instruction. The refractometer is the authority.
7. Predicted compound inventory¶
Furanone / caramel — Stage B (pH 4.4 is the furanone band) HDMF (furaneol) caramelised strawberry, jam; homofuraneol; mesifurane; maltol caramel-malt; cyclotene maple, burnt sugar; DDMP. Furfural and 5-methylfurfural almond-bready from rosehip pentose and ascorbate browning. (High on the band; Med on amounts.)
Norisoprenoid — Stage B and Stage C, the design's reason for existing β-ionone violet, tea (threshold ~7 ng/L); β-damascenone honey, stewed apple, rose (~50 ng/L); dihydroactinidiolide tea, hay; 3-hydroxy-β-ionone; TDN and vitispirane precursors kerosene-petrol, aged-Riesling at long age. Two routes, deliberately: thermal cleavage over 100 min in Stage B, and acid-catalysed release in the unbased Stage C at pH 3.46, which then continues in the beer for months (§8). (High on mechanism; Low on concentration and on which compound dominates.)
Honey / rose — Stage B Phenylacetaldehyde honey, hyacinth in the jar → 2-phenylethanol 34–69 mg/L rose, honey added.
Lactone — Stage B γ- and δ-decalactone from apricot peach skin, creamy. (Med.)
Structure and phenolic — Stage C Catechin, epicatechin and new oligomers drying grip, mid-palate persistence from acid-catalysed proanthocyanidin depolymerisation — astringency down, depth up. Protocatechuic acid and phloroglucinaldehyde mild smoky-phenolic from anthocyanin thermolysis. 4-vinylguaiacol clove and 4-vinylphenol medicinal-phenolic, pre-formed, awaiting Brett.
Oak — Stage C cis/trans-oak lactone coconut, woody, vanillin vanilla, eugenol clove, furfural, cyclotene. Chip extraction skews toward the toast-layer compounds and away from lactone — see the tweak.
Kernel — Stage D Benzaldehyde 1.66–2.07 mg/L marzipan, kirsch, stone fruit.
Deliberately absent: alkylpyrazines. pH 4.4 suppresses them, and cocoa/roast would read as chocolate cherry stout, not wine. Shiitake was considered and left out for the same reason.
8. Survives vs destroyed¶
Survives to 18 months and beyond: furanones — furaneol partly reduced or methylated over a long secondary (Med) — maltol caramel, cyclotene; β-damascenone honey, stewed apple, β-ionone violet, TDN kerosene at age; melanoidins dark caramel, body; catechin/epicatechin and the new oligomers grip; vanillin, eugenol clove, oak lactone coconut; phenolic acids.
Transformed, usefully: phenylacetaldehyde honey → 2-phenylethanol rose — acceptable, often preferable. 4-VG and 4-VP → 4-ethylguaiacol smoky, spicy and 4-ethylphenol barnyard, horse-blanket by Brett VPR, skipping the PAD step. That is the gueuze axis, arriving over 6–12 months.
Lost: furfural and HMF → furfuryl alcohol (neutral). Benzaldehyde is the one at risk — Brett ADH reduces it to benzyl alcohol and the beer is still slowly fermenting. See tweak 2.
What the beer should taste of at 18 months. Nose: stewed red fruit and dried apricot, honey moving to rose, violet and tea with age, a clove-and-smoke edge as the ethylphenols arrive, a faint marzipan behind it. Palate: caramel-jam mid-palate without residual sweetness, then a real drying grip through the finish that the beer does not currently have. Acidity firmer and brighter than now, tart rather than sour. Finish long, dried-fruit and woody, with the sloe tannin holding it.
9. Fusel, methionol, 2-PE, FAN and THP budget¶
| Added mg/L | Threshold / band | Verdict | |
|---|---|---|---|
| Isoamyl alcohol | 1.4 – 2.8 | warn >25 | clear |
| All BCAA fusels | 3.4 – 6.8 | — | clear |
| Methionol | 0.16 – 0.33 | warn >0.5 at midpoint | clear |
| 2-Phenylethanol | 34.4 – 68.8 | useful 40–100; soapy >100 | wanted; lands in band |
| Added FAN | 27 mg N/L | warn >30 | under — was 33 before the glycine cut |
| Lys + Arg carried | 117 mg | — | fruit, stone and seed protein only |
Collagen 0 g against a 4 g cap. Other protein 6.5 g — sloe stone (~20 %), aronia, rosehip and apricot all carry it. Exactly the v4 correction: v3 would have counted it as zero.
10. Kernel calculation — cracked sloe stone¶
| Cracked sloe stone | 7 g, A2 in water, vented |
| Amygdalin, planning | ~0.8 % of stone |
| A2 completeness assumed | 0.70 |
| Effective HCN retention | 0.58 (not the 0.2 best case) |
| Route | Stage D, post-primary |
| Benzaldehyde potential | 2.96 mg/L → 1.66–2.07 mg/L in beer after the 0.70 vent survival |
| Flavour verdict | in range |
| HCN per 375 mL bottle | 0.169 mg |
| % ARfD, one bottle | 12.1 % |
| % ARfD, two bottles | 24.2 % |
| 50 % ARfD warn point | 29 g of stone — this design is at a quarter of it |
Matches the bitter-apricot-kernel build it replaces almost exactly (2.1 mg/L benzaldehyde, 12.4 %/24.9 % ARfD), at 7 g of stone instead of 1.5 g of kernel.
The one thing this build must not repeat from batch 001. There, cracked stones went straight into the hot cook. The emulsin was denatured within minutes, so most of the amygdalin arrived in the beer intact for Brett to hydrolyse in the bottle, with nothing to vent — which is why 001's HCN is still an open item in the log. Separate the stones, A2 them in water, vent, and only then use the liquor. Raw or hot-cooked stone into Stage C or D is an exclusion.
Ethyl carbamate: residual cyanide plus ethanol over months, light-accelerated. Store these bottles dark.
11. Colour and body¶
Stage B at pH 4.4 makes less melanoidin than a 6.15 jar — the trade for the furanones — but Stage C thermolyses the anthocyanin in 27 g of aronia and 23 g of sloe flesh (Anth H), on top of 25 g of rosehip.
Direction: from purple-red toward garnet then tawny-red, roughly +3 to +5 SRM, and it is not reversible. For a wine target that is right, and the beer-side Lallzyme below will accelerate it.
Body comes from melanoidins and the Stage C oligomers, not residual sugar. The sloe tannin adds perceived weight on the finish that the current beer has none of.
12. Fermentable load¶
| Planned (B 25 %, C 10 % reacted) | 67 g → 14.8 g/L |
| Worst case (0 % reacted) | 83 g → 18.4 g/L |
| ΔABV planned | +0.96 % |
| ΔABV worst case | +1.19 % |
| Sucrose included | 4.5 g, uninverted, 100 % fermentable |
Plan on the worst case. The 25 % conversion figure is Med at best and its error runs toward more sugar.
Warning raised by the calculation: at 18.4 g/L this goes in early in secondary and the gravity gate is absolute — stable gravity read twice, 4–6 weeks apart, before bottling.
Carve-out: the Stage D kernel liquor carries no sugar. No tincture. Nothing sits outside the gate.
Against the stated ABV constraint: +1.19 % worst case. Up from the fresh-fruit build's +1.09, and the one number that drifted the wrong way across revisions — the sloe flesh and the extra rosehip did it, and Stage C sugar barely reacts. If it is too much, pull tweak 3.
Sizing check. Stage B concentrate 21 g/L (accent anchor 10–30) — comfortable. Total fruit reports as 100 g/L against the ≤100 anchor, but that counts the 231 mL of modelled rehydration water as fruit mass. The real fruit dry load is 218 g, or 48 g/L — less than half the anchor. The check is an artefact of representing water as an ingredient, not a breach; noted rather than designed around.
13. Beer-side acid, potassium and FAN¶
| Base carried in | 11.7 meq → 2.6 meq/L (warn >10) |
| Expected beer pH rise from base | +0.06 to +0.13 units |
| Free fruit acid from unbased Stage C | 11.5 meq/L, pulling the other way |
| Net acid balance | +8.9 meq/L in favour of acidity |
| Expected net pH fall | roughly 0.2–0.45 units (buffer 20–40 meq/L per unit, Med) |
| Potassium added | 101 mg/L (warn >500) |
| FAN added | 27 mg N/L (warn >30) |
| MLF substrate | malate from aronia, sloe, apricot and rosehip; quinic and citric are not convertible |
The design doing the job asked of it: the potassium and pH load that normally dominates a Maillard prep is gone, and the net movement on brightness is clearly positive.
Measure the beer pH before adding anything, and again about four weeks after. A 0.2–0.45 fall from an unmeasured start is the one place this could overshoot. If the beer is already at 3.2 or below, hold back a third of the Stage C liquor and taste before committing it.
If it is still not bright enough at three months, lactic acid 88 % beer-side to taste is the blunt control. Do that after tasting the prep in, not alongside, or you will not know which lever moved it.
14. §21 compliance¶
| Hold | Duration | pH class | Verdict |
|---|---|---|---|
| A2, cracked sloe stone in water, 40–50 °C | 2 h | ~5.5, low-acid | at the cap, compliant |
| Stage B rehydration soak, warm | ≤2 h | ~3.5, acidic | compliant (3 h limit applies) |
| A1 / protease | not run | — | — |
- Nothing held warm enters the beer uncooked: the A2 liquor is vented, a sub-boil pasteurisation, and the Stage B soak goes to 121 °C.
- No raw fraction is pre-held warm. Stage D is kernel liquor only — no raw fruit.
- Stage B Brix confirmed on the refractometer before sealing. This is the water-activity safety hurdle, not a kinetic preference.
- Stage C unbased at native pH 3.46, well below 4.6, so the low-acid process does not apply. Ice bath, then ≤3 °C, ≤5 days.
15. §22 — acrylamide and furan¶
Acrylamide levers, three of four pulled:
- pH 4.4, not 6.15. The largest lever available, taken for flavour reasons with the contaminant benefit as a genuine bonus.
- Glycine present (1.0 g), competing for the dicarbonyl pool. Smaller than in the earlier build — the cut was forced by the FAN budget, and it slightly weakens this lever. Stated, not hidden.
- Date backbone, the lowest-asparagine RS donor.
- Not pulled: apricot and rosehip are asparagine-bearing. Accepted — they are the carotenoid donors.
Furan: this design sits squarely in the formation regime. Sealed jars at 121 °C with sugar, amino acid and ascorbate-rich rosehip is the textbook condition, and there is rosehip in both jars. The hot vent is mandatory — 5–10 min open and stirring while hot, on both jars. Furan boils at 31 °C and leaves readily; furanones, norisoprenoids and melanoidins are orders of magnitude less volatile and stay. The vent also strips residual HCN from the A2 liquor. (High that the condition applies; Low on concentration at this scale.)
HMF: not a concern — no inversion, no low-pH hob phase.
Confidence tags¶
| Claim | Confidence |
|---|---|
| Beer has essentially no carotenoid / norisoprenoid precursor pool | High |
| pH 4.4 favours furanones over alkylpyrazines | High |
| β-damascenone release is acid-catalysed and continues in beer over months | High |
| Base solve arithmetic | High; the absolute pH it targets, Med |
| Kernel benzaldehyde and HCN arithmetic | High on arithmetic, Med on stone amygdalin |
| Acid-catalysed proanthocyanidin depolymerisation reduces astringency, raises depth | Med-high |
| Sloe stone fraction ~23 % of dried whole sloe | Med — weigh yours |
| Stage B / C native pH estimates (3.5 / 3.46) | Low — measure both |
| Dried rosehip carotenoid retention, and the 1.5× dose compensation | Med |
| Absolute norisoprenoid concentration delivered | Low |
| Expected net beer pH fall of 0.2–0.45 units | Med — buffer capacity is a range, not a number |
| 25 % Stage B sugar conversion | Med at best — plan on worst case |
| Oak lactone yield from 25 min hot extraction of 4.5 g chips | Low — blend to taste |
| Benzaldehyde survival through a still-fermenting secondary | Med |
Tweak points¶
1. Keep the tannin. Sloe 30 g → 45 g whole. More flesh into Stage C (VH polyphenol) and ~10 g of stone, which lifts benzaldehyde to ~3 mg/L and HCN to ~17 % ARfD per bottle — still well inside. Costs ~5 g more Stage C sugar and ~8 meq more acid. Pull this if the beer tastes thin rather than dull.
2. Keep the marzipan. Hold the vented A2 liquor refrigerated and dose it at bottling rather than now, so Brett ADH does not reduce the benzaldehyde to benzyl alcohol over the next year. Highest value on the list and it costs nothing.
3. Cut the sugar load. Date 45 g → 30 g and drop the Stage C rosehip to 15 g. Worst case falls to roughly 15.5 g/L and ΔABV to ~+1.0 %. Costs furanone depth and some of the second norisoprenoid route. Pull this if +1.19 % is more ABV than you want.
4. Protect the oak. Swap the 4.5 g of chips for 8 g of cubes if you can get them, or hold the chips out of the hot jar entirely and put them in the beer with the blend. Hot chip extraction is fast and skews toward toast-layer compounds; cubes give the oak lactone the wine profile actually wants.
Two things outside the prep¶
Dose Lallzyme Beta beer-side, 225 mg into 4.5 L (50 mg/L), once the blend is in. Glucose is gone, which is the condition the enzyme needs. It releases bound aroma from the grapes, aronia and elderberry already in the beer and from the glycosides that survive Stage C — the cheapest complexity in the design. Its pectinase side activity degrades anthocyanin-stabilising structures, so it will accelerate the colour shift in section 11. Given the target is tawny-garnet, that works with you.
The honest limit. This build reaches the norisoprenoid pool and the furanone mid-palate, and with the sloe it reaches further into polymeric phenolic structure than the earlier version did. It still does not fully replicate the condensed seed-and-skin tannin of a fine red. If that exact texture is the goal, an oenological grape tannin at 10–30 mg/L is the direct answer and this method has no better substitute.
Numbers to record¶
- Stone and flesh yield from cracking the 30 g of dried sloe.
- Stage B native pH, measured cold before anything is added (predicted 3.5, Low).
- Stage B cold pH after the 0.93 g first pass — the calibration delta.
- Final KHCO₃ to reach 4.4.
- °Brix at seal.
- Stage C native pH (predicted 3.46).
- Beer pH before the addition and ~4 weeks after — this build is designed to move it down 0.2–0.45 units and that prediction needs testing.
- Whether the A2 stone still smelled of bitter almond after the hold.
Say the word and I will write the mode-6 batch-log block.
Log¶
2026-09-16 · Designed¶
Design intaken from design-001-aged-wine-depth-dried-larder.md.
2026-09-22 · Renumbered¶
Cross-reference updated after the Maillard designs were renumbered to match their IDs: the L-phenylalanine agreement is in design 004 (was numbered 003). This design keeps its number, 001.