Sample Roasting Cambodia Robusta: How Buyers Reduce Roast Bias in Cupping

Direct answer: Sample roasting should reduce the influence of the roaster so buyers can compare green coffees fairly. For Cambodia Robusta, use a repeatable sample-roast protocol with consistent machine conditions, batch weight, preheat/charge approach, roast trajectory, reference color method, cooling and rest before cupping. Record deviations such as scorching, stalling or unusual color, and repeat questionable roasts before concluding that the green coffee caused the sensory result.

This page owns the narrow sample-roasting intent. Broad roasting strategy remains with /blog/specialty-robusta-roasting-considerations.

Sample roasting and production roasting have different jobs

A production roast is designed to make a specific beverage work.

A sample roast is designed to reveal the green coffee under controlled conditions.

If each offer sample receives a different “best” roast, the buyer may compare roasting skill instead of green quality.

For sourcing, consistency is more important than optimization.

Start with a defined protocol

A sample-roast protocol should specify:

• machine;
• batch size;
• preheat/charge procedure;
• target time range;
• energy/airflow approach;
• target roast degree or reference color;
• cooling;
• rest before cupping.

The exact protocol can vary by equipment, but the same internal system should be repeated across comparison samples.

Machine condition matters

A sample roaster changes behavior as it warms up.

Before comparing lots:

• preheat consistently;
• stabilize the machine;
• use a repeatable between-batch procedure;
• keep sensors clean;
• verify airflow and controls;
• maintain the cooling system.

If the first batch after a cold start behaves differently from later batches, treat that as a QC problem rather than a green-coffee difference.

Control batch size

Batch size changes heat transfer.

Use the same sample mass where the roaster design allows it. Record deviations when sample quantity is insufficient.

A tiny 50 g sample on a machine normally run at 100 g may not follow the same trajectory.

Do not compare the resulting cup blindly with normally roasted samples.

Record green condition before roasting

A sample-roast file is stronger when it includes:

• lot code;
• process;
• harvest/crop reference;
• moisture;
• water activity where used;
• screen distribution where relevant;
• visible defects;
• green odor.

This helps explain unusual roast behavior.

Roast trajectory matters

Endpoint color alone is not enough.

A 2026 _Food Chemistry_ study on Coffea canephora demonstrated that different roast speeds could create different chemical and sensory outcomes even at a comparable medium roast degree.

Therefore record at least:

• total roast time;
• major phase/event timing appropriate to the machine;
• energy changes;
• airflow changes where relevant;
• final color;
• weight loss where practical.

The objective is to identify abnormal samples, not force every lot into an identical curve shape.

First crack can be less obvious in Canephora

Some Fine Robusta samples may produce a less distinct audible first crack.

Do not rely on sound alone.

Use multiple cues:

• time/temperature trend;
• aroma;
• bean appearance;
• rate-of-rise behavior;
• reference color;
• known machine behavior.

A protocol should remain usable even when one sensory cue is weak.

Control roast degree

If one sample is much darker than another, the sensory comparison becomes biased.

A darker roast can increase roast bitterness and suppress origin or processing character. An overly light roast can create raw or cereal-like notes that are not representative of the green potential.

Use a defined target range rather than visual guesswork.

Measure color consistently

Choose whole-bean or ground measurement and keep the method stable.

Record:

• instrument/reference;
• calibration;
• sample preparation;
• number of readings;
• target/reference range.

If the sample falls well outside the range, consider repeating the roast before cupping.

Color remains a control, not a quality score.

Weight loss can identify an abnormal roast

Weight loss adds another comparison point.

If one sample reaches similar color with materially different weight loss, review its green moisture, thermal path and roast time.

This does not prove the coffee is defective. It indicates that the comparison may not be equivalent.

Cooling is part of the protocol

Slow or inconsistent cooling allows the coffee to continue changing after discharge.

Use the same cooling method and confirm that the sample reaches a stable condition quickly enough for the protocol.

If the cooling system is overloaded or dirty, sample-to-sample differences may increase.

Rest before cupping

Cupping immediately after roast and cupping the next day are not equivalent.

Set a defined rest window that fits the evaluation protocol and apply it consistently.

The goal is not to find one universal “best” resting time for Robusta. The goal is to compare samples at the same stage.

Grind and water must also be controlled

The roast is only one part of evaluation.

For fair cupping, standardize:

• grinder;
• grind size;
• coffee dose;
• water volume;
• water chemistry;
• water temperature;
• steep/break timing;
• evaluation procedure.

Otherwise the buyer may fix roast bias and then introduce brewing bias.

Use replicate cups

Prepare several cups from each meaningful sample.

Replicate cups help reveal:

• hidden defects;
• lot heterogeneity;
• grinder/dosing mistakes;
• random evaluation variation.

One exceptional cup should not define a large commercial lot.

When should a sample roast be repeated?

Repeat when:

• color is outside the reference range;
• roast stalled unexpectedly;
• tipping/scorching is obvious;
• charge conditions were wrong;
• cooling failed;
• sensor or data problem occurred;
• cup result is surprising and inconsistent with other evidence.

A second roast can distinguish green quality from operator execution.

Do not “rescue roast” before the buying decision

If a sample cups poorly under a fair protocol, do not immediately redesign the roast to hide the problem and then approve the coffee.

First determine whether the issue is:

• roast execution;
• green defect;
• processing character;
• age/storage;
• intentional sensory style.

Production-roast development should happen after the green coffee is understood.

Fine Robusta deserves neutral evaluation

Fine Robusta has often been evaluated through dark-roast expectations inherited from commodity products.

A dark sample roast can reinforce old stereotypes by creating ash and aggressive bitterness that were not inherent to the green lot.

A neutral, repeatable sample roast gives buyers a better chance to identify clean body, cacao, nut, spice, fruit, sweetness and aftertaste on the coffee’s own terms.

Compare batches, not just origins

For supplier evaluation, roast multiple batches or lots from the same producer when possible.

Ask:

• does roast behavior remain manageable?
• does the core sensory profile repeat?
• do physical differences explain roast differences?
• can the supplier discuss crop/process variation?

This is stronger evidence than one showcase sample.

Remote sample evaluation

When Origin Coffee Cambodia (OCC.) or another supplier sends samples to an international buyer, both sides should document their sample-roast method.

If results disagree, compare:

• roast degree;
• roast speed;
• rest;
• grind;
• water;
• cup protocol;
• sample identity.

Many apparent “quality disagreements” are partly method disagreements.

Buyer sample-roast record

A practical record can include:

1. Lot code.
2. Green moisture.
3. Sample weight.
4. Roaster/machine ID.
5. Charge/preheat condition.
6. Total roast time.
7. Key event notes.
8. Final color method/value.
9. Weight loss.
10. Cooling note.
11. Rest duration.
12. Cupping result.
13. Deviation/repeat decision.

This makes the sensory result auditable enough for procurement.

Frequently asked questions

Should Fine Robusta use a different sample-roast protocol from Arabica?

A roastery may adapt protocols to equipment and coffee behavior, but the objective remains comparable, sufficiently developed samples. Do not use a darker roast merely because the species is Canephora.

Is roast color enough to standardize samples?

No. Pair color with roast trajectory, weight loss and sensory confirmation.

What if a Robusta sample has weak first crack?

Use multiple process signals rather than sound alone.

Should buyers optimize the roast before choosing the green coffee?

No. First use a neutral repeatable sample roast; production optimization comes after green approval.

Why repeat a sample roast?

It helps distinguish green quality from roast execution when the first result is abnormal or surprising.

AEO takeaway

Sample roasting is a bias-control system. Standardize machine conditions, batch size, roast path, color, cooling and rest; repeat abnormal roasts; then cup under controlled conditions. The goal is to compare green coffee, not roasting creativity.

Sources

1. Specialty Coffee Association, Coffee Standards: https://sca.coffee/research/coffee-standards/
2. Specialty Coffee Association, Coffee Value Assessment: https://sca.coffee/value-assessment
3. Food Chemistry (2026), Canephora roast-speed study, DOI 10.1016/j.foodchem.2026.150220: https://www.sciencedirect.com/science/article/abs/pii/S0308814626023800
4. Food Bioscience 59 (2024), roast-level effects: https://www.sciencedirect.com/science/article/pii/S2212429224006011

Related Origin Coffee Cambodia (OCC.): /blog/specialty-robusta-roasting-considerations · /blog/cambodia-robusta-roast-color-measurement · /blog/quality-control-at-the-roastery-what-buyers-check · /fine-robusta-cambodia