Skip to main content

Proposed: A Sorghum Malt Evaluation Standard (GFB-SME, draft v0.3)

We know of no good standard for evaluating sorghum malt in modern brewing — and we have looked. Labs fell back on South Africa's SABS 235 — an opaque-beer spec — or barley procedures, or Taylor's methods, or, in our own correspondence over the years, they asked us. So here is our draft — sorghum malt as a character ingredient, not a conversion engine — published for the field to adopt, adapt, or tear apart. If a better standard exists, showing it to us improves this page.

Status: proposed, draft v0.3, never ratified by anyone. This is a discussion document in the same spirit as The Maltose Mash: a specific, criticizable design beats an empty field. If you run a lab, a malting program, or a breeding program and this draft is wrong for your work, we want to know exactly where.

What this proposal answers

The full case file lives in the Evaluation Methods survey — every methodology the field has used, each reviewed on its own page. The one-paragraph version: the only real standard, SABS/SANS 235, serves South Africa's opaque-beer industry where the malt is the conversion engine, and it became everyone else's stopgap because it was the only sorghum-shaped standard on the shelf. Everything else was borrowed — barley methods that measure sorghum's deficits, institute methods that fragmented the units, research conventions built for different questions, and commercial panels built for other grains entirely.

Our proposal is for the product class none of the borrowed tools were built for: clear, stable, packaged gluten-free beer brewed with external enzymes, where the malt's job is character.

The core problem in the borrowed methods

Outside the opaque-beer world, sorghum malt evaluation was barley evaluation with the labels changed. The reference methods — congress mash, ASBC/EBC malt analysis — were built for a grain whose malt converts itself. Run them on sorghum and the numbers mostly measure sorghum's enzyme deficit: low DP, low alpha, incomplete conversion, and an extract figure gated by the enzymes rather than the malt's actual contribution. Decades of screening data (including ours) scored cultivars DP-first because that's what the borrowed methods measure best.

Our position, built at commercial scale: that's measuring the wrong thing. In a working gluten-free process, external enzymes own conversion. The malt's job is flavor, color, body, clean wort, and yeast nutrition. An evaluation standard should measure the job, not the handicap — and as an aside we'll defend anywhere: further research into maximizing native sorghum enzymes is effort the brewing side of the field doesn't need.

Principle 1 — Evaluate in an external-enzyme reference mash

Replace the congress mash with a standardized enzyme-added mini-mash: fixed heat-stable alpha-amylase and saccharification doses (specified per gram of grist, products and activities stated), a fixed temperature program modeled on the 190°F liquefaction baseline, fixed water ratio in L/kg. Conversion is held constant by design — so every difference between samples is the malt, not the enzymes.

From that one mash, the measures that matter:

MeasureWhy it's first-class
Extract (% d.b., method stated)The malt's material contribution, ungated by native enzymes
Wort turbidity + clarityThe practical problem that decides whether sorghum beer is drinkable — method and units specified below (v0.3), as the 2002–2004 program measured but did not specify
FAN (ppm)Sorghum's real deficit; drives yeast performance downstream
Wort colorCharacter contribution, kiln response
Wort pH and viscosityProcess-fit signals
Defined sensory panelMalt is for taste — wort and hot-steep aroma/flavor scored against a published lexicon, not "seemed fine"

Native enzyme measures (DP, alpha-amylase) become optional descriptive data — reported if measured, never a ranking criterion, and always with the unit system named.

The reference mash, specified

Added in v0.2 so a laboratory can actually run this — doses derived from our documented production practice, offered for criticism, not validated as a lab method:

Laboratory Run Sheet

GFB-SME Reference Mash

Protocol
GFB-SME
Version
Draft v0.2
Effective
August 2026

The proposed standard's enzyme-added mini-mash: conversion held constant by design, so every difference between malt samples is the malt, not the enzymes. Doses are GFB's documented production doses scaled per kilogram of grist — a starting point offered for criticism, not a validated laboratory method.

Draft status: Proposed, draft v0.2, never ratified by anyone. Product names churn — the binding specification is the enzyme class plus the declared activity on the label, and every report must state the exact product, declared activity, lot, and dose. If your laboratory substitutes equivalent-activity products, say so: that friction report is exactly what this draft needs.

Water ratio3.0 L/kgThin end of the documented 2.2–3.0 production band, for lab handling
Liquefy88°C (190°F), 60 minThe documented production baseline
Convert63°C (145°F), 45 minGlucoamylase — fixed, maximal endpoint so extract is ungated

Declare before the first sample

Format
1
Dose

Enzyme into the full-volume water

Full water volume
Do
Add heat-stable bacterial alpha-amylase at 4.4 mL/kg of grist (Termamyl SC DS class) to the full-volume water.
Add
  • Heat-stable alpha-amylase, 4.4 mL/kg (state product, activity, lot)
Advance when
Dosed and mixing.
Record
Product, declared activity, lot, dose
More detail for step 1
2
Mash in

Grist in gradually, hot

88°C (190°F) at 3.0 L/kg
Do
With the water at 88°C, add the grist gradually with continuous stirring.
Add
  • Grist, gradually
Advance when
Grist fully in, mash even.
Record
Grist mass (d.b.) · Actual mash-in temperature
More detail for step 2
3
Liquefy

The 60-minute hold

88°C60 min
Do
Hold 88°C for 60 minutes with continuous stirring.
Add
None
Advance when
Hold complete.
Record
Temperature log · Viscosity observations
More detail for step 3
4
Convert

Glucoamylase to a fixed endpoint

63°C (145°F)45 min
Do
Cool to 63°C. Add glucoamylase at 1.65 mL/kg (Amylase AG 300L class). Hold 45 minutes.
Add
  • Glucoamylase, 1.65 mL/kg (state product, activity, lot)
Advance when
Hold complete.
Record
Product, activity, lot, dose · Cooling time
More detail for step 4
5
Mash out

End enzyme work

82°C (180°F)10 min
Do
Raise to 82°C and hold 10 minutes.
Add
None
Advance when
Hold complete.
Record
Time to temperature
More detail for step 5
6
Measure

The measures that matter

Filtered wort
Do
Filter. Measure extract (% d.b., method stated), FAN (ppm), turbidity, wort clarity, wort color, pH, and viscosity. Run the sensory panel per Principle 4.
Add
None
Advance when
All declared measures recorded with units.
Record
Every measure with its method and units · Sensory scores against the published lexicon
More detail for step 6
StepActionTargetAddAdvance when
1. Dose: more detailAdd heat-stable bacterial alpha-amylase at 4.4 mL/kg of grist (Termamyl SC DS class) to the full-volume water.Full water volume
  • Heat-stable alpha-amylase, 4.4 mL/kg (state product, activity, lot)
Dosed and mixing.
2. Mash in: more detailWith the water at 88°C, add the grist gradually with continuous stirring.88°C (190°F) at 3.0 L/kg
  • Grist, gradually
Grist fully in, mash even.
3. Liquefy: more detailHold 88°C for 60 minutes with continuous stirring.88°C
60 min
NoneHold complete.
4. Convert: more detailCool to 63°C. Add glucoamylase at 1.65 mL/kg (Amylase AG 300L class). Hold 45 minutes.63°C (145°F)
45 min
  • Glucoamylase, 1.65 mL/kg (state product, activity, lot)
Hold complete.
5. Mash out: more detailRaise to 82°C and hold 10 minutes.82°C (180°F)
10 min
NoneHold complete.
6. Measure: more detailFilter. Measure extract (% d.b., method stated), FAN (ppm), turbidity, wort clarity, wort color, pH, and viscosity. Run the sensory panel per Principle 4.Filtered wort
NoneAll declared measures recorded with units.
Authority, limitations, and review control

GFB-SME draft v0.2 — proposed by Gluten Free Brewer, ratified by nobody, published for the field to adopt, adapt, or tear apart. Doses derive from documented production practice scaled per kilogram; the temperature program is the documented 190°F baseline plus a fixed glucoamylase endpoint. Running this sheet and reporting friction is Research Ask #5.

Review when: laboratories report friction, breeding programs compare it against their own protocols, or criticism lands — the version history on this page records every change.

Research Ask #5 — run the proposed standard

Level 2 · Detailed Procedure

Run each step with its checks and context

Use the run sheet during a routine mash. Use these expanded steps when setting up the process, training an operator, documenting a deviation, or diagnosing unexpected behavior.

Enzyme into the full-volume water

Objective

Active liquefaction enzyme before grist — the same rule the production process lives by.

Operating instruction

Add heat-stable bacterial alpha-amylase at 4.4 mL/kg of grist (Termamyl SC DS class) to the full-volume water.

Add
  • Heat-stable alpha-amylase, 4.4 mL/kg (state product, activity, lot)

What good looks like

Dosed and mixing.

Watch for

No additional step-specific warning.

Record

Product, declared activity, lot, dose

Grist in gradually, hot

Objective

Even wetting at liquefaction temperature without clumping.

Operating instruction

With the water at 88°C, add the grist gradually with continuous stirring.

Add
  • Grist, gradually

What good looks like

Grist fully in, mash even.

Watch for

No additional step-specific warning.

Record

Grist mass (d.b.) · Actual mash-in temperature

The 60-minute hold

Objective

Full liquefaction on the documented production baseline — the fixed half of holding conversion constant.

Operating instruction

Hold 88°C for 60 minutes with continuous stirring.

What good looks like

Hold complete.

Watch for

No additional step-specific warning.

Record

Temperature log · Viscosity observations

Glucoamylase to a fixed endpoint

Objective

Drive conversion to a fixed, maximal endpoint so extract measures the malt, not the conversion chemistry.

Operating instruction

Cool to 63°C. Add glucoamylase at 1.65 mL/kg (Amylase AG 300L class). Hold 45 minutes.

Add
  • Glucoamylase, 1.65 mL/kg (state product, activity, lot)

What good looks like

Hold complete.

Watch for

No additional step-specific warning.

Record

Product, activity, lot, dose · Cooling time

End enzyme work

Objective

A clean stop before filtration.

Operating instruction

Raise to 82°C and hold 10 minutes.

What good looks like

Hold complete.

Watch for

No additional step-specific warning.

Record

Time to temperature

The measures that matter

Objective

The standard ranks for the beer, not the assay: clean wort and real extract gate; flavor differentiates; FAN sets the nutrition plan.

Operating instruction

Filter. Measure extract (% d.b., method stated), FAN (ppm), turbidity, wort clarity, wort color, pH, and viscosity. Run the sensory panel per Principle 4.

What good looks like

All declared measures recorded with units.

Watch for

Reporting DP or alpha-amylase as a ranking criterion — they are optional descriptive data here, never rank.

Record

Every measure with its method and units · Sensory scores against the published lexicon

Design notes, criticize freely: glucoamylase fills the saccharification slot because it drives conversion to a fixed, maximal endpoint — extract is then ungated by conversion chemistry and every difference between samples is the malt. The 3.0 L/kg ratio sits at the thin end of the documented production band (2.2–3.0 L/kg) for laboratory handling. Products churn names — our own records track the same enzyme under multiple labels — so the binding specification is the class plus the declared activity on the label, and every report must state the exact product, declared activity, lot, and dose per Principle 3. If your laboratory substitutes equivalent-activity products, say so; that report is exactly the friction data this draft needs.

Turbidity, specified

A standard that promotes turbidity to a first-class measure and then leaves it methodless would be failing its own Principle 3. So, added in v0.3:

Turbidity. Reported in formazin units (FTU/EBC formazin; NTU acceptable where the instrument reports it), with the unit and instrument type declared — nephelometric or absorptiometric, calibrated against a formazin standard. Two values per wort sample: as-run (wort as the reference mash produced it, settled 10 minutes, no centrifuge) and clarified (after a declared spin or filtration step). The pair is the measure: as-run tells the brewer what the lauter and kettle will see; the difference tells them how much of the haze is removable solids versus stable colloid. A single undeclared number is non-conforming.

The two-value form encodes documented production practice — our own 2014–2017 QA log recorded turbidity in FTU at two spin levels, batch by batch. The 10-minute settle is a proposed convention, exactly the kind of detail an adopting lab should pressure-test.

Principle 2 — The reference malting schedule

Screening comparisons need a shared malting method. We propose the Madison bench schedule as the reference, because it's documented and we've published its output: 26-hour steep (8h wet / 1h air / 8h wet / 1h air / 8h wet) at 24°C; germination at 24°C with spray-watering to 41% moisture (wet basis); endpoint at 1:1 acrospire-to-grain length (~72 hours bench scale); five-stage kiln finishing at 85°C. The 85°C finish is deliberate doctrine — character over enzyme survival.

Principle 3 — Mandatory declarations

Every archived screwup in our own records traces to an unstated basis. So the standard makes these declarations mandatory on every report:

  1. Moisture basis — wet or dry, on every moisture and extract figure.
  2. Units on every enzymatic measure — °Lintner, °WK, ASBC, or SDU, never bare numbers. (Our own 2002–2004 dataset carries an unlabeled DP column; it's published with a warning because the units died with the worksheet. Don't be that worksheet.)
  3. Ratios in L/kg, with any local units in parentheses.
  4. Germination length and scale — bench and production timelines differ by days.
  5. Cultivar identity — named, or explicitly declared anonymous. Anonymous grain is a different, lesser class of result.

Principle 4 — Rank for the beer, not the assay

A scoring framework, not a single index: clean wort and real extract are gate criteria; flavor panel differentiates; FAN sets the nutrition plan; color maps the malt to styles. A cultivar with modest numbers and a clean, flavorful wort outranks a hazy DP champion — because the first one makes better beer with external enzymes, and the second one makes a mess with slightly less enzyme.

What would make this a real standard

Adoption, criticism, and iteration — in that order. Specifically: labs running the reference mash and reporting friction; breeding programs (the University of Nevada, Reno program is building exactly the methodology this draft addresses) comparing it against their own protocols; and anyone publishing sorghum malt data adopting the mandatory declarations even if they reject everything else. The declarations alone would fix half the field's comparability problem.

Version history

  • v0.3 — 2026-08-23. Turbidity method and units specified (formazin units, declared instrument, as-run + clarified pair) — the standard's own units discipline applied to its headline measure. Approved by Craig from the drafting record.
  • v0.2 — 2026-08-09. Reference mash specified with products-by-class, doses, temperature program, and ratio; "proven at commercial scale" softened to "built." Prompted by academic-readiness review: a standard nobody can execute from the page is a position paper.
  • v0.1 — 2026-08-08. Initial draft.
Source and Validation Notes

This proposal synthesizes Bard's commercial malting and brewing experience, the USDA-ARS screening program's documented methods, and the site's published doctrine. It is a draft authored to be criticized; nothing here has been ratified by any standards body, and the reference-mash enzyme specification (products, doses, activities) is deliberately left at v0.1 as the first thing adopters should pressure-test with us.

Research needed