Rate of Rise Management
Rate of Rise (RoR) is one of the most discussed but easiest-to-overinterpret signals on the Roest. This page explains when RoR is useful, when it is mostly probe behavior, and how to use it as one input alongside taste, color, weight loss, development time, and phase timing. It focuses on practical Roest roasting decisions rather than universal declining-RoR theory.
What RoR Represents on Roest
RoR is derived from bean-temperature data, so its usefulness depends on the quality of the BT signal. If the BT probe is not consistently immersed in the bean mass, RoR can describe probe contact, air contact, steam release, or agitation effects rather than true bean momentum 2 sources.
This is why RoR rules from larger drum roasters do not transfer cleanly to Roest. Rules such as fixed 20–30°C/min targets were described as machine- and probe-dependent, not general laws for all roasting systems source. Roest’s small batches, fast sensors, air movement, probe placement, and low stored system energy make its curves behave differently from large gas drum roasters.
RoR should therefore be treated as a diagnostic signal, not a primary roast objective. Stronger decisions usually come from the combination of taste, weight loss, color, bean appearance, development time, and phase timing source. See Color Reading and Measurement, Weight Loss Targets, and Cupping and Sensory Evaluation for those companion metrics.
When RoR Is Worth Trusting
Batch size is the first filter. At 50–100g, especially with high agitation, BT may read a mix of air and beans, and RoR can become misleading. Multiple experienced users advise ignoring or heavily discounting RoR at 100g and lower; RoR tracking related to bean temperature becomes more useful from roughly 120g upward, but only on Roest models rated for those batch sizes 2 sources. Do not overload an S100/L100 or other 100g-class machine to improve RoR reliability; on 100g-limited machines, treat RoR as less reliable and use inlet/time, color, weight loss, development time, and cupping instead.
At 100g, negative RoR near or after first crack can be normal and may not mean the roast has stopped developing. Roest reportedly told one user that negative RoR is expected with 100g batches, and another explanation was that the BT probe may not remain immersed in the bean pile as beans expand and lighten 2 sources. For small-batch profiling, inlet/time profiles are often preferred over BT/RoR-driven control; one approach is to use inlet/time for 100g or less because IT/BT becomes “too wonky at crack” source.
For machines rated for these loads, the signal improves at 120–150g. One summary placed 120g as “better” and 150g as “a lot more reliable” while still retaining useful ET and exhaust feedback source. Around 185g, on models rated for that batch size, the BT probe can become quite responsive, with a fast turning point and rapid RoR rise as the probe and coffee reach equilibrium source. For more on batch-dependent behavior, see Batch Size Scaling.
Display and Smoothing Settings
RoR display settings change the curve enough that two roasts can look different without tasting different. A 15-second interval gives higher-definition feedback but is more sensitive and can look jumpy; several users report using 15s for high-detail review 2 sources. A 5-second RoR display can look choppy simply because the interval is too short source. For 100g, a 30-second interval has been recommended because the probe is noisy when it is not continuously embedded in the coffee pile source.
A practical default is to use 20–30s smoothing for general interpretation, 15s only when intentionally studying short-term behavior, and avoid making major roast decisions from a 5s curve. If RoR appears wavy or jagged under an ET profile, the curve may be showing heater modulation or sampling effects rather than meaningful bean behavior 2 sources.
Practical RoR Targets
RoR targets should be tied to batch size, process, and roast style. The following values are community starting points, not universal rules.
| Situation | Useful RoR / temperature-rise target | Notes |
|---|---|---|
| Washed coffee entering first crack | About 7–9°C/min | Suggested as an entry target for washed coffees source. |
| Washed coffee finish | Around 55–60s development in one common light-roast approach | This is development guidance tied to post-FC momentum, not a standalone RoR rule source. |
| Naturals entering development | Lower heat and gentler RoR than washed | Naturals tend to speed up into development and generally need less heat near and into development 2 sources. |
| Natural development time | 30–45s max in several light-roast approaches | Often paired with gentle low RoR; some very light naturals were run 31–35s source. |
| Post-FC BT rise, light roasts | About 2–3°C | One user usually aims for ~2°C BT increase after crack; another emphasized 3°C or under after FC 2 sources. |
| Fast natural roasts | 3–4°C increase after FC for 30–40s | Reported as workable for that type of fast natural roast source. |
| End RoR in low-RoR approach | Close to 0 by end | Low RoR from crack and close to 0 at the end was emphasized in one test approach source. |
These targets should be cross-checked against sensory results. A roast can have an ugly RoR curve and cup well, while a smooth curve can still taste baked, underdeveloped, or dull. Development decisions should be made with Development Time and Drop Decisions, not RoR alone.
RoR Management Workflow
Use this workflow as the canonical practical procedure for RoR decisions on Roest.
-
Choose a batch size before interpreting RoR. For 50–100g, treat RoR as advisory at best. For RoR-led analysis, use 120g or more only on Roest models rated for those batch sizes, and expect more stable data around 150g and above on compatible machines 2 sources. Do not overload an S100/L100 or other 100g-class machine to improve RoR reliability; on 100g-limited machines, use inlet/time, color, weight loss, development time, and cupping instead.
-
Set a sane RoR interval. Use 20–30s for general roasting interpretation. Use 15s only if intentionally studying fine movement, and expect more noise. Avoid reacting to every 5s wiggle.
-
Do not chase early-roast RoR. Early RoR is less important than reaching sensible phase landmarks. One common approach is to focus more on visual color change timing than on the first 3–4 minutes of RoR behavior source.
-
Make momentum decisions before first crack. If RoR is too high into first crack, reduce energy before crack rather than waiting for the flick. Small power reductions of 1–2% around 30–40 seconds before FC have been suggested for smoothing entry momentum 2 sources.
-
Avoid air or RPM changes right at first crack unless the profile intentionally requires them. Increasing RPM near FC can create a flick or false RoR increase by changing bean spacing and hot-air contact with the BT probe, and air/RPM changes near FC can distort graph readings source. If airflow is part of the strategy, plan the change before or after the sensitive crack window.
-
Watch ET/inlet direction as a flick predictor. A rising ET or a late inlet increase can predict a flick. If flicks occur, use a larger inlet cut beginning before crack rather than reacting after the RoR has already turned 2 sources.
-
Do not spike heat reactively during a crash. Increasing heat in the crash can make it worse; corrections should usually happen earlier in the profile, by changing the energy path into crack rather than panic-adjusting during the event source.
-
Evaluate the cup before declaring success. If the roast tastes flat, cereal-like, papery, bitter, or roasty, use sensory evidence to decide whether RoR was actually meaningful. A nice-looking graph is not a guarantee of a good roast source.
Symptom → Likely Cause → Adjustment
| Symptom | Likely cause | Adjustment |
|---|---|---|
| RoR goes negative at 100g but cup is not baked | Probe contact / BT artifact from small batch size | Do not chase the curve. Use inlet/time, taste, color, weight loss, and development time instead 2 sources. |
| RoR is extremely noisy or jumpy | Short RoR interval, small batch, high RPM, or poor probe immersion | Increase RoR interval toward 30s for 100g; consider a larger batch if RoR tracking matters source. |
| RoR flicks near or after first crack | Too much momentum, rising ET, late inlet increase, or disruptive airflow/RPM change | Cut inlet earlier or more decisively before FC; avoid increasing fan immediately before FC unless intentionally testing that effect 2 sources. |
| RoR crashes hard at first crack on small batches | Probe exposure, bean expansion, moisture release, or low batch mass | Treat as partly measurement-related at 100g. If moving to 120–150g on a machine rated for those loads removes the behavior, do not solve it with heat changes 2 sources. |
| BT or RoR stalls/plateaus for a long time | Too little carried energy, too much heat cut, or an extended flat phase | Correct the next roast by carrying more energy before the plateau, not by late panic heat. Long slow/flat RoR can lead to baked character source. |
| Outside gets dark while inside tastes underdeveloped | Too high RoR or energy into FC, especially on naturals/anaerobics or low-density coffees | Lower heat into development, reduce FC entry momentum, and shorten development if the coffee runs away 2 sources. |
| Black tips or embryo burning | Too high energy/RoR into FC or too much heat after FC | Enter FC at lower RoR and reduce heat earlier; this was specifically linked to anaerobic/low-density behavior 2 sources. |
| Roast tastes baked, flat, cereal-like, or astringent | Excessive flat/near-zero RoR time or stalled Maillard/development | Avoid holding RoR near zero for long periods; compare against color, weight loss, and sensory notes 2 sources. |
| RoR looks smooth but coffee lacks intensity | Over-prioritizing curve aesthetics | Cup against a slightly faster or higher-energy version; do not optimize for graph beauty alone source. |
| RoR looks ugly but cup is clean and sweet | RoR artifact or nontraditional but valid profile shape | Keep the roast if sensory results are good; record color, weight loss, and development time for repeatability. |
Process-Specific Behavior
Naturals, anaerobics, and heavily processed coffees often take off into development. They may absorb early heat and then accelerate quickly, making them prone to flicking, tipping, black embryo spots, or excessive post-FC temperature gain if heat is not reduced in time 2 sources. These coffees are commonly handled with lower heat near and into development and shorter development windows. See Natural Process Roasting.
Washed coffees can tolerate different energy paths, but they may also crash hard depending on density, batch size, and profile. Washed Ethiopia and Kenya are repeatedly discussed as coffees where RoR behavior can be misleading or strongly bean-dependent. See Washed Process Roasting and Kenya and Ethiopian Coffees.
RoR, Airflow, and RPM
Airflow and RPM changes can alter both the actual roast and the measurement. Changing airflow changes convective heat transfer and can change the RoR reading itself source. Increasing RPM near first crack can create more space between beans, induce more hot-air influence into the BT reading, and show as a flick that may be partly artificial source.
For this reason, RoR work should not be isolated from Airflow and Fan Settings, Drum Speed RPM Settings, and Pressure Management. If airflow or RPM is changed during a test, treat the resulting RoR change as both a roast change and a measurement change.
RoR and Profile Type
Inlet profiles often make more sense than BT-only profiles when RoR is unstable, because RoR inherits BT delay and probe artifacts. One experienced view is that BT is too influenced by batch weight, probe delay, bean shape, volume, and agitation to be a reliable control signal, with inlet or power profiles giving clearer control inputs source. Another user found inlet profiles regulated better and produced straighter RoR than ET profiles source.
BT profiles are still available and have supporters, but Roest BT control can lag because the BT signal is delayed. If the machine cannot hit the requested BT line, the profile may run faster or differently than intended source. For details, see Bean Temperature Profiling, Inlet Temperature Management, and Power Curve Strategies.
Meaningful Conflicts
CONFLICT (Unresolved): Declining RoR as goal vs RoR as secondary data.
Some roasters aim for a smooth declining RoR curve with no acceleration, using temperature, time, and curve together for drop decisions source. Others argue that babying a straight declining RoR on Roest can produce worse cup results and that RoR should be ignored or heavily discounted, especially at small batch sizes 2 sources. The practical resolution is to use RoR as a warning signal, then validate by cup quality and repeatable metrics.
CONFLICT (Unresolved): Low RoR into development vs higher-energy fast approaches.
Several light-roast approaches aim for low RoR after crack, close to zero by drop, and only a small post-FC BT rise. Other fast/high-energy approaches intentionally use higher RoR entering crack, shorter development, and more temperature gain. Both appear in successful community experiments, so the choice depends on batch size, process, target style, and sensory result rather than a single RoR rule.