How to Use One Roast Curve for Every Batch Size
https://www.rubasseroasters.com/article_d.php?lang=en&tb=7&id=1284
If you roast for a living, you already know the dream – one dialled profile that holds up from sample size to full production. This guide explains the Roast-Master replication approach, how batch-size compensation works, which parameters the system adjusts, and what the tasting data shows when you scale up or down. We keep it neutral and technical so you can apply the ideas on any modern setup.
Hero image: Article header. Short caption, “Roast-Master replication with batch-size compensation.” Rubasse NIR Digital Coffee Roasters
The RM replication idea, match the curve and the context
Hitting the BT line is not enough, you also keep the roast environment under control.
Many replay systems chase temperature and miss the way the environment shifts when the turning point arrives lower than the target. That leads to long, slow power corrections and a wobbly finish. The RM approach references extra signals such as Rate of Rise to correct sooner and hold the roast under steadier conditions. You will see more neutral overall power trends rather than an over or under bias across the back half.
Place image: Profile comparison graphic that shows the different power behaviours. Caption, “Common replay versus RM replication.” Rubasse NIR Digital Coffee Roasters
Batch-size automation, one profile across sizes
Why it matters: Beans do not scale linearly, your controls should.
Batch-size automation lets you roast different weights of the same coffee on the same machine and land very similar flavours, fully automatically. It does this by treating the stored profile as a reference, then compensating several controls for batch mass rather than simply forcing the same BT shape. The result is two profiles that look different on paper yet taste closely aligned in the cup. Rubasse NIR Digital Coffee Roasters
What the system adjusts for you
How to read this: The compensation is per 300 grams, and values are user-tunable where appropriate.
When you load a saved profile, change batch size on the touch panel, then press Auto, the system adjusts:
- Charge temperature and initial heat: Larger batch, higher charge and higher initial power. Rubasse NIR Digital Coffee Roasters
- Drum RPM: Larger batch, higher RPM. Rubasse NIR Digital Coffee Roasters
- Pressure compensation: Larger batch, the controller raises exhaust to maintain a higher differential pressure between the room and the drum, which preserves comparable airflow. Rubasse NIR Digital Coffee Roasters
- Discharge temperature: Larger batch, slightly higher end temperature. Rubasse NIR Digital Coffee Roasters
All compensations are shown and applied on a per-300-gram basis, and most can be adjusted by the user to suit house style. The intention is flexibility without losing the engineering baseline. Rubasse NIR Digital Coffee Roasters
Place image: Control-panel layout with per-300 g notes. Caption, “Compensations expressed per 300 g.” Rubasse NIR Digital Coffee Roasters
Drum speed, aim for consistent first-crack temperature
Shop wisdom: Use RPM to bring first-crack anchor points closer together across sizes.
On this system, the most useful RPM target is a near-identical first-crack temperature for each batch weight of the same coffee. Pay attention at 300 g, which tends to crack a little hotter. A modest RPM increase at 300 g helps. From 600 g upwards, first-crack temperatures line up more naturally, so 600 g is a sensible minimum for highest accuracy. Rubasse NIR Digital Coffee Roasters
Discharge temperature compensation, set for your style
Rule of thumb: Faster profiles that ride a higher RoR into crack often need a larger drop-temperature increment as batch size grows. Converged profiles need less.
Larger batches on the same machine usually need a higher end temperature to taste like the smaller batch. If you roast with a faster run into first crack, you will likely prefer a larger discharge-temperature compensation. If your profiles are more converged and asymptotic, a smaller value tends to replicate better. The example tuning leaned to converged, so the team preferred about 0.15 °C per 300 g rather than the default 0.25 °C per 300 g for their trials. Rubasse NIR Digital Coffee Roasters
Place image: GIF showing identical 1 kg profile applied to 3 kg with two different discharge-comp values. Caption, “Same curve target, different drop compensation.” Rubasse NIR Digital Coffee Roasters
Charge temperature and initial heat, prioritise initial power
Practical guidance: Use charge as a coarse lever, then fine-tune the initial power step.
When you enter a different batch size, the controller by default raises charge about 5 °C per 300 g and adds 3.5 percent initial heat per 300 g. Because the drum is thin stainless steel and does not bank much energy, the recommendation is to leave the charge-temperature compensation at default and do most of your refinement on the initial heat compensation. Note that the initial heat change is applied at roast start rather than shown immediately on edit, which is expected behaviour. Rubasse NIR Digital Coffee Roasters
Pressure compensation, hold airflow by number not by slider
Key idea: Match the differential pressure, not the fan percentage.
The same fan or damper setting does not guarantee the same airflow, since ambient pressure, temperature and even wind can shift things. Air moves because of pressure differences, so holding the room-to-drum differential pressure steady is a truer way to keep airflow constant. Larger batches will usually need more airflow. The default compensation value of 1.5 per 300 g is set for safety, not for flavour. A tasting-led range of roughly 0.5 to 1.5 per 300 g is suggested, and there is a clear caution not to alter the engineering parameter labelled “PID enhancement ratio per 300 g” unless you fully understand the consequences. Rubasse NIR Digital Coffee Roasters
Test design, three real-world scenarios with cupping comparisons
How to copy this: Pick one coffee, fix your reference, then scale up or down with only batch-size compensation enabled.
Coffee: Guatemala Antigua, Bella Carmona, washed, Caturra, Catuai, Bourbon.
User parameters for the trials: discharge-temperature compensation 0.13 °C per 300 g, pressure compensation 0.5 per 300 g, other compensations left at default. Two colleagues cupped all sets blind and recorded qualitative differences. Headline finding, target and replicated batches tasted very similar in all three experiments. Rubasse NIR Digital Coffee Roasters
Place image: Reference profile used and the three scenario traces. Rubasse NIR Digital Coffee Roasters
Scenario 1, 300 g profile to 3 kg production
What to expect: Efficiency without losing your sample intent.
A sample roast profile was promoted directly to a full 3 kg run using batch-size automation. Colour readings were close, and the cupping notes describe small differences in aroma weight and bitterness, with overall flavour alignment. Rubasse NIR Digital Coffee Roasters
Place image: Comparison graph plus colour table. Caption, “300 g to 3 kg, automated.” Rubasse NIR Digital Coffee Roasters
Scenario 2, 1 kg profile to 3 kg production
What to expect: Tripled output with minimal drift.
A well-liked 1 kg profile was replayed at 3 kg. Cuppers found the two cups hard to tell apart, with minor shifts in caramel and fruit emphasis as the bowls cooled. Ground Agtron values were tight, typically in the low seventies. Rubasse NIR Digital Coffee Roasters
Place image: 1 kg to 3 kg trace and colour table. Rubasse NIR Digital Coffee Roasters
Scenario 3, 3 kg profile down to 1 kg
What to expect: Precision when you are finishing the last kilo of a premium lot.
The controller reduced batch size cleanly and preserved profile intent. Tasters noted small swings between caramel weight and fruit lift, yet overall convergence as the cups cooled. Colour readings were essentially matched. Rubasse NIR Digital Coffee Roasters
Place image: 3 kg to 1 kg trace and colour table. Rubasse NIR Digital Coffee Roasters
Practical setup notes that help any shop
Day one checklist: Keep these in mind before you chase the last half-point.
- Treat 600 g as the smallest ideal batch for accurate automation on this platform. Rubasse NIR Digital Coffee Roasters
- Aim for the same first-crack temperature across sizes, using RPM and airflow to align it. Rubasse NIR Digital Coffee Roasters
- For faster styles, expect to use a larger discharge-comp value. For converged styles, try smaller. Validate with blind cupping and consistent bowl weights for grounds and water. Rubasse NIR Digital Coffee Roasters
- Use differential pressure as your airflow truth. Keep the fan percentage as a means, not an end. Rubasse NIR Digital Coffee Roasters
Summary, scale intent not just weight
Bottom line: Batch-size automation is not a gimmick, it is a framework that preserves your reference environment as mass changes.
By combining temperature-aware replication with per-300-gram compensations for charge, initial heat, drum speed, exhaust pressure, and discharge temperature, you can push a favourite profile to different batch sizes and keep flavour on target. Validate with colour and blind cupping, then lock the parameters that serve your house style.
Considered as one of the region’s pioneers in specialty coffee education and Dubai cafe culture development, Ryan Godinho is an Australian entrepreneur who is accredited as the country's first SCAA AST and National Coffee Championships Coordinator. He is a frequent contributor to Forbes and Entrepreneur Magazines and also holds a postgraduate Certificate of Advanced Studies in Coffee Excellence from Zurich University (ZHAW).
