How to Scale a Fine Robusta Roast From Sample Roaster to Production Roaster

Scaling a coffee roast from a sample roaster to a production machine is not a matter of copying temperatures and multiplying batch size. Different machines transfer heat differently, measure temperature differently, move air differently, and respond differently to changes in gas, power, drum speed, airflow, and batch mass.

For Fine Robusta, the task is especially important because the roaster may be trying to preserve sweetness, body, clean bitterness, process character, and origin distinction while avoiding the assumption that Canephora simply needs more roast development.

The safest approach is to scale the sensory target, not the curve itself.

Start with the purpose of the sample roast

A sample roast answers an early question: what does this green coffee contain, and is it worth further evaluation?

It is usually designed for comparison and sensory access rather than final commercial service.

A production roast answers a different question: how should this coffee be roasted consistently for a specific product on a specific machine?

The production target may be espresso, milk drinks, filter, retail, hospitality, or a blend component.

That means the sample roast is a reference, not a recipe to copy literally.

Write the sensory target before touching the production machine

Record what the best sample roast showed.

Useful notes can include:

• sweetness;
• body;
• bitterness quality;
• acidity;
• aroma;
• flavor clarity;
• process character;
• aftertaste;
• defects or concerns.

Then define what the commercial product needs to preserve or change.

For example, a Fine Robusta sample may show cocoa, dried fruit, dense body, and a clean bittersweet finish. The production espresso may need slightly more body and solubility while preserving the clean finish.

That description is more transferable than a sample-roaster end temperature.

Confirm that the green sample represents the commercial lot

Do not spend time scaling a roast from a sample that cannot be connected to the coffee being purchased.

Check:

• lot or sample code;
• origin;
• process;
• harvest;
• physical condition;
• storage history where available;
• whether the commercial inventory is already assembled.

If the production lot differs materially from the sample, re-evaluate before scaling.

Sample-to-lot integrity is part of roast development.

Measure physical properties that affect roasting

Useful green-coffee observations may include:

• moisture content;
• water activity where used;
• bean size distribution;
• density or bulk-density reference where part of the roastery system;
• defect condition;
• green color and odor;
• visible process variation.

These measurements do not automatically dictate the roast profile. They help explain why the coffee may behave differently from familiar Arabica lots or from another Canephora.

Avoid turning one physical metric into a universal heat rule.

Understand that machine temperatures are machine-specific

A probe reading on a 100-gram sample roaster cannot be assumed to equal the same physical condition on a 15-kilogram drum roaster.

Probe placement, metal mass, airflow, batch size, sensor response, software smoothing, and burner design all affect the displayed number.

Use machine temperatures as internal references for that machine.

Do not treat them as portable truth.

The transferable information is the sequence of coffee changes and the resulting cup.

Begin production scaling with a conservative pilot batch

Do not move directly from a small sample roast to the largest possible production batch.

Use a pilot size that gives the production machine stable control without risking unnecessary green coffee.

The correct pilot size depends on machine design and normal operating range.

A batch that is too small for the roaster may behave abnormally, so “smaller” is not always safer.

The goal is a representative production test.

Match roast intent rather than copying sample time

A sample roast may be fast because the machine is designed for fast standardized evaluation.

A production roast may require a different total duration to achieve the same sensory direction.

Track the major phases, but judge them through cup results.

Possible working references include:

• drying progression;
• color change;
• aroma development;
• rate-of-rise behavior;
• first-crack behavior where detectable;
• post-crack development;
• end color;
• weight loss.

No single phase ratio guarantees quality.

Fine Robusta may challenge Arabica-based visual assumptions

Roasters accustomed to Arabica may interpret bean color, surface appearance, expansion, or crack behavior through familiar references.

Canephora can differ physically.

Do not call the roast defective because the beans do not look identical to an Arabica batch.

Evaluate:

• internal and external color where possible;
• aroma;
• physical uniformity;
• sensory result;
• repeatability.

Appearance is evidence, not the final verdict.

Use roast color as one anchor

Color measurement can help compare production batches more consistently than visual judgment alone.

Record the measurement method and whether the coffee is whole bean or ground.

Color should be interpreted with weight loss, time, machine settings, and sensory result.

Two roasts with similar color can still taste different because heat application and internal development differed.

The value of color is repeatability, not complete diagnosis.

Track weight loss

Roast weight loss gives another production reference.

It can help identify batches that finished unusually differently from the approved profile.

Again, there is no universal Fine Robusta weight-loss target.

Use the roastery's own repeated data and the sensory target.

If a batch suddenly loses materially more or less mass under otherwise similar conditions, investigate before assuming it is acceptable.

Control one variable at a time during trials

When the first production batch misses the target, avoid changing charge condition, gas, airflow, batch size, and end point simultaneously.

Name the sensory problem first.

Examples:

• too roasty;
• too harsh;
• underdeveloped flavor;
• weak sweetness;
• too drying;
• insufficient body;
• process character too dominant;
• flat aftertaste.

Then make one major directional change.

This creates learning rather than a series of unrelated profiles.

Separate roast bitterness from green-coffee bitterness

Fine Robusta can contain a stronger bitterness structure than many Arabicas, but the roaster should distinguish clean, integrated bitterness from roast damage or green defects.

If the sample roast was clean and the production roast becomes smoky or ashy, the production profile is a likely variable.

If harshness appears across multiple controlled roasts, re-evaluate the green sample, process, storage, and sensory condition.

Do not solve every bitter cup by roasting darker or lighter without diagnosis.

Preserve sweetness intentionally

A successful Fine Robusta roast often depends on making sweetness perceptible enough that body and bitterness feel structured rather than aggressive.

Roast development should therefore be evaluated through sweetness in the finished beverage.

Cup the coffee at consistent rest and preparation conditions.

If espresso is the final use, test espresso rather than relying only on cupping.

A production profile can cup well and fail in the intended application.

Test the actual product format

For espresso:

• establish dose and yield;
• test straight shots;
• test milk drinks;
• observe extraction repeatability.

For filter:

• use a controlled brew recipe;
• evaluate clarity, body, bitterness, and cooling behavior.

For blends:

• test the Fine Robusta at the intended component level;
• compare against the control blend;
• record what the component contributes.

Application testing should decide the final roast.

Resting can change the production verdict

Fresh roast behavior can differ significantly from the coffee's normal service state.

Evaluate the same profile across the realistic rest and selling window.

The roastery may find that one profile is harsh on day one but integrates later, while another tastes excellent early and fades quickly.

Do not publish one universal rest period for Fine Robusta.

Build the operating window from the actual roast and packaging.

Compare pilot batches blind where possible

When the team knows which roast was supposed to be “improved,” expectation can influence judgment.

Blind or partially blinded comparison can reduce that bias.

Use coded cups and a consistent evaluation sequence.

The team does not need a competition protocol for every production decision. It needs a repeatable method that separates preference from profile labels.

Scale from pilot to full production carefully

A larger batch changes thermal behavior.

When moving from pilot to normal production size, review:

• recovery after charge;
• heat availability;
• airflow;
• rate-of-rise response;
• phase progression;
• color;
• weight loss;
• sensory result.

Do not assume the pilot profile will scale linearly.

The full production batch must be cupped and application-tested again.

Production consistency requires tolerance ranges

A commercial roast should not depend on hitting one exact timestamp or temperature.

Create acceptable ranges around the variables that actually correlate with sensory quality on that machine.

Examples might include:

• roast color range;
• weight-loss range;
• total-time range;
• end-temperature internal reference;
• sensory acceptance criteria.

The ranges should be built from repeated successful batches, not invented before data exist.

Define sensory acceptance criteria

A Fine Robusta production roast might require:

• no obvious defects;
• clean bitterness;
• target sweetness;
• target body;
• acceptable aftertaste;
• intended process character;
• performance in espresso or milk;
• no significant roast taint.

The exact criteria depend on the product.

A roast can fall inside a numerical profile and still fail the cup.

Keep a reference batch

Once the production profile is approved, keep clear records for the reference batch.

Store:

• green lot;
• batch size;
• machine;
• profile version;
• roast color;
• weight loss;
• roast date;
• sensory notes;
• application recipe;
• approval status.

Future batches can be compared against this reference.

The goal is not to recreate one curve perfectly. It is to recreate the product.

Watch for green-coffee change over time

The same lot can change during storage.

Moisture, physical condition, and sensory expression may shift.

If the production profile begins behaving differently months later, do not assume the machine or operator is the only cause.

Recheck the green coffee and compare with earlier records.

The roast may need controlled adjustment to maintain the product target.

Crop transitions need re-development

A new harvest should not automatically inherit the previous crop's production profile.

Repeat the process:

1. verify sample identity;
2. sample roast;
3. compare sensory profile;
4. run production pilot;
5. application test;
6. set new reference.

The new coffee may be similar enough for a small adjustment or different enough to require a new development cycle.

Treat the transition as quality control, not inconvenience.

Team handoff matters

The roast should be repeatable by more than one operator when the business requires it.

Document:

• profile intent;
• machine controls;
• sensory target;
• acceptable variation;
• what to do when the roast drifts.

Train roasters to diagnose rather than merely follow a line on the screen.

A profile that only one person understands is an operational risk.

Production-roast review checklist

Before approving the scaled Fine Robusta roast, confirm:

1. Commercial lot matches evaluated sample.
2. Green physical data are recorded.
3. Pilot batch used representative machine loading.
4. Roast color is measured consistently.
5. Weight loss is recorded.
6. Major heat and airflow decisions are documented.
7. More than one roast has been tasted.
8. Application testing is complete.
9. Rest behavior is understood.
10. Full production batch has been revalidated.
11. Sensory acceptance criteria exist.
12. Reference batch is stored in records.
13. Crop-transition plan exists.

Common scaling mistakes

Avoid copying sample-roaster temperatures directly, forcing Fine Robusta into an Arabica profile because the curve looks familiar, changing too many variables at once, approving the roast before application testing, judging bean appearance without cupping, or using one excellent batch as proof of production repeatability.

Another common mistake is optimizing the profile for software aesthetics. A smooth curve is useful only when the coffee tastes right and the profile can be reproduced.

Cambodia Fine Robusta context

For OCC, roast development should make Cambodia-origin coffee easier to understand through the cup. The objective is not to impose one national roast style.

A Mondulkiri or other verified Cambodian lot should retain the character that made the lot commercially interesting while being adapted to the intended product.

Any origin or process claim must remain tied to the actual lot.

Bottom line

Scaling Fine Robusta from sample roaster to production roaster is a translation problem. Translate the sensory target, green-coffee evidence, and intended application into the controls of the production machine. Use pilot batches, physical references, controlled changes, application tests, and repeated QC before approving the profile.

The production roast is successful when the customer receives the same coherent product the development team intended—not when the production curve visually resembles the sample-roaster curve.

Explore OCC Roasting Program or Contact OCC for commercial roast development.