Power Curve Strategies
Power curves define how much heater power the Roest applies through the roast, either directly in a power profile or indirectly while following inlet, ET, BT, or hybrid targets. This page explains how to choose a power-curve shape, how batch size changes that shape, and how to make practical adjustments without chasing misleading curves. It focuses on Roest-style small-batch roasting; charge temperature, airflow, pressure, and process-specific handling are covered in linked pages where those topics require more detail.
What the Power Curve Actually Represents
The displayed power graph is only the power sent to the heating elements, not total machine power or fan power source. This matters because fan changes alter the heat-transfer environment around the same power number. On temperature-based profiles, increasing airflow increases power draw because the heater must heat more air; on power profiles or manual roasting, increasing airflow tends to make temperatures drop because the same heater output is spread across more air source.
Power profiles are attractive because they are simple, responsive, and useful for manual experimentation. Several experienced users prefer them for learning a coffee because the heat application is visible and easy to alter in steps. They are also sensitive to voltage, ambient conditions, and machine variation, so a shared “75% power” curve should be treated as a shape, not a recipe. Profiles often need adjustment by unit, beans, voltage, and airflow setup rather than being copied exactly 2 sources.
For the related temperature-control side of the same problem, see Inlet Temperature Management, Bean Temperature Profiling, and Calibration and Environment.
The Core Rule: Batch Size Changes the Curve Shape
Power curves do not scale linearly with batch size. Changing from 50g to 100g, 150g, 185g, or 200g changes bean mass, air-to-bean ratio, bean movement, probe contact, and heat storage, so the same curve can behave like a different roast altogether 2 sources. A 150g profile is not a 100g profile plus a few percent power, and a 200g profile is not simply a 150g profile stretched longer.
A useful rule of thumb from community experiments is that peak power moves later as batch size gets smaller and earlier as batch size gets larger. For 190-200g, peak power tends to sit left of center; for 130-150g, closer to the center/right; and for 50-100g, later again because the bean mass holds less heat source. This is one reason 150-185g profiles can often use more heat early and then step down, while 50-100g profiles often need a slower ramp and smaller reductions.
Batch size also affects what the sensors appear to show. Low-mass batches can make BT and RoR less trustworthy because the probe is not always embedded in the bean pile, while larger batches tend to provide better readings. See Batch Size Scaling and Rate of Rise Management for sensor interpretation.
Starting Power-Curve Templates
These templates are starting points, not universal profiles. They consolidate recurring community patterns for Roest-style power profiling and should be adjusted by taste, first-crack timing, weight loss, color, and defect symptoms.
Use only batch sizes within your model’s rated capacity. The 150-200g examples are for L200/Ultra-class machines or other Roest models explicitly rated for those loads. S100/L100 users should stay within their manual’s maximum batch size; P3000 users should use model-specific guidance rather than copying small-roaster percentages directly. curated High heater percentages assume normal airflow/exhaust and a clear chaff path; do not test high power with the fan/exhaust effectively closed, a blocked filter, or an overfull chaff collector, and stop or vent the roast if you see chaff ignition, abnormal smoke, or over-temperature warnings.
| Use case | Curve shape | Starting values | Adjustment target |
|---|---|---|---|
| 50g sample roast | Later, gentler peak with very small reductions | Reduce power only about 3-4% over the roast; use this primarily for quick sample evaluation rather than consumption roasting source. | Keep the curve smooth and avoid aggressive early heat; BT/RoR may not be reliable at this size. |
| 100g sample or learning roast | Slow rise toward yellow, then minimal decline | A slow rise up to around yellow with only a small decline can work well because 100g loses heat quickly source. For 50-100g, a slow ramp helps avoid a burned-onion or metallic taste source. | Make small changes, usually 1-2% or a few degrees equivalent. Avoid large late power cuts unless the roast has too much momentum. |
| 120-130g bridge profile | Moderate power with enough mass for better readings | One suggested simple profile was 56% power, 40% air, 55 rpm, 130g load, with FC marked manually and about 60 seconds development source. | Use this range when 100g feels too noisy but 150g is too much green to spend. |
| 150g washed/high-density | Early energy, then step-downs | A common 150g power pattern is 85-75-65: charge around 180°C, 45s soak with no power, 85% to 4:10, 75% to 5:00, then 65% source. Another recurring pattern is 80-70-60 toward FC, then lower to 55 or 50 depending on the roast source. | Aim for a controlled FC around the intended window; avoid end power that falls so low the coffee tastes under. |
| 150-180g general power profile | More early heat, then reductions | For 150-180g, one practical plan is 20-30% power for up to 45s, then 85-90%, reach about 155°C BT around 3:30, and drop power in 5-8% increments roughly every 30s if needed source. | If crack lands around 5-6 minutes in that style, the roast is in range; if it runs away, lower peak or reduce earlier. |
| 180-185g consumption roast | Early push, controlled decline | Fill to 180-185g, start pushing 80-90% power from about 00:35, then decline by 5-10% steps to arrive around 50-55% power just before FC source. Another suggested 185g setup used 225°C charge, 55 rpm, fan 45-60%, and 65-75% peak power between about 3:30 and 5:00 source. | Use the larger batch mass to carry momentum; do not copy low-batch late-peak curves unchanged. |
| 200g exploratory roast | Earlier peak and larger late reductions | On 200g roasts, one reported approach reduced from about 75% to 50% power, while 100g roasts only needed 5-7% total reduction source. Another 200g example reached FC after 7:30 with decreasing power from 74% source. | Treat 200g as machine- and bean-dependent; see Batch Size Scaling before using it as a routine setting. |
Choosing a Curve Shape
A power curve can be understood as a tradeoff between heat applied early and heat left available near first crack. More heat early tends to increase intensity, body, and weight loss, while slower early heat can preserve delicacy but risks underdevelopment or flatness if the roast never builds enough momentum source. For larger batches, early heat is often needed because the beans absorb more energy, but after drying the stored momentum means power must usually come down more than it would on a small batch source.
For manual roasting, stepped reductions are often easier to control than a perfectly smooth decline. One experienced user found that manual step-downs worked better than a smooth power ramp down on Roest source. Another summarized the logic as accumulating enough energy early so it can be reduced at the finish source.
Large power pulses at the wrong time can create defects. Two big pulses right at yellow and into development were associated with baked, mellow, burned-bread, savory character in a side-by-side comparison source. Conversely, reducing power too early can leave the roast short of energy: a profile that reduced power too fast could be corrected either by lowering dose by 20g or increasing the end power point by about 5-10% source.
Process and Density Adjustments
Washed, high-density coffees can often take more energy and faster heat application than naturals or heavily processed coffees. One recommended high-density washed approach was to load 120g, use high power for the first 30-60 seconds, then drop to 68% and keep the RPM high source. Another broad recommendation was to use only 5-7% power cuts for washed coffees but 10-15% cuts for naturals when adapting a high-power dry phase into a stepped-down profile source.
Naturals and processed coffees are less consistent. Fast 4-5 minute roasts on natural or anaerobic coffees did not cup well in one set of trials, while 6:30-7:30 roasts produced much better results source. Naturals may also need shorter development once FC is reached, because excess heat near and after FC can push surface burn, roasty notes, or blackened tips. Detailed process-specific handling belongs in Natural Process Roasting and Washed Process Roasting.
Bean density should influence the first curve chosen. Higher-density beans were described as needing higher-energy curves, while bigger and less-dense beans tend to need lower-energy curves source. If a coffee tastes woody, flat, or under even with more development time, the problem may be earlier heat application rather than post-FC time.
Diagnosing Power-Curve Problems
| Symptom | Likely power-curve issue | Practical fix |
|---|---|---|
| ET and BT are too close together, or the roast stalls before FC | Too little heat input for the bean and batch | Add power earlier or raise the middle/end power; ET/BT being too close was described as too low power input source. |
| Underdeveloped taste persists even after extending development | Early or mid-roast energy is insufficient | Do not rely only on longer development; under taste was described as not fixable by simply extending development source. |
| Beans smell under and end inlet/power is low | End power is too low | Avoid going below about 60-55% power for 150g if under beans appear in that style source. |
| Roasty, burned bread, savory, or baked notes | Excess heat pulses or too much heat at the wrong phase | Remove abrupt spikes around yellow/development and reduce peak or post-peak heat; big pulses at yellow and development were linked to baked notes source. |
| Uneven exterior on 100g | Too much hot air or power early for the low mass | Use a slower ramp and avoid blasting early; high early power on 100g was associated with unevenness source. |
| Crash near first crack | The roast is arriving with the wrong momentum, not necessarily too little heat at that instant | Avoid panic heat additions during the crash; one recommendation was never to increase heat in the crash because it exacerbates the problem source. |
| FC too late after copying a shared power profile | Machine power, voltage, ambient conditions, or batch behavior differ | Increase power modestly, often 3-5% or 5-10%, then re-evaluate timing rather than assuming the shared number transfers exactly 2 sources. |
| FC too early or roast too fast | Peak is too high or too early for that bean/batch | Reduce peak by a few percent, move the peak later for low batch sizes, or slow the first phase through Charge Temperature Guidelines. |
| Converted inlet profile from a power log burns or maxes at 100% | Conversion or PID behavior is not matching the original heat application | Rebuild manually or simplify the profile; one converted power-to-inlet profile topped at 100% and burned beans source. |
Manual Power Profiles and Step Changes
Manual power control is useful when learning a coffee because it allows the roaster to change one thing at a time and watch the response. A practical manual profile described in the community used 5% power for 45 seconds, then 80%, then a drop from 80 to 70% at 4:10 and from 70 to 60% at 5:00 source. Another manual suggestion was 50% for 1 minute, then 75-80% for a maximum of 3 minutes, then slowly down source.
Small adjustments matter. A 5% difference can be large, and several users reported useful changes from 1-2%, 3-5%, or 5-10% adjustments depending on the context source. Power reductions also change pressure and airflow behavior, so large drops near FC should be coordinated with Airflow and Fan Settings and Pressure Management.
Power Profiles, Inlet Profiles, and Hybrid Control
Power profiles give direct control but do not compensate for voltage, ambient temperature, hot or cold greens, or different machine behavior. Inlet profiles can compensate for some temperature variance by asking the machine to hit an air-temperature target, but they introduce PID response and can create aggressive power behavior after turning point. BT/PWR and BT/IT profiles add bean-temperature feedback when BT is reliable, but BT is less reliable at low batch sizes.
For 50-100g on L100/S100, BT was repeatedly described as unreliable, making inlet or power strategies more practical than pure BT control source. At larger batch sizes, BT/PWR can be useful because it lets the user set power at a given bean temperature and curve between set points source. For more on BT-based control, see Bean Temperature Profiling.
Meaningful Disagreements
CONFLICT (Unresolved): Voltage and power-profile transfer. One view is that power profiles are too dependent on socket voltage and household load; lower voltage or running appliances can change roast behavior even with identical power values source. Another view is that voltage differences do not explain profile differences as long as the supply can deliver the same power, summarized as “watts = watts” source. The practical compromise is to treat power profiles as local recipes and log voltage/ambient conditions when repeatability matters.
CONFLICT (Unresolved): Airflow’s effect on flavor. Some experiments suggested that airflow choice does not strongly affect flavor if temperature is matched accordingly source. Other users found higher or lower airflow changed cup character, development, roastiness, or clarity in their setups. Because airflow changes power draw, pressure, inlet temperature, and heat transfer together, it should not be evaluated only from the power curve; see Airflow and Fan Settings.
Practical Logging
Power-curve work depends on comparisons. Useful logs include batch size, power curve, fan/exhaust settings, RPM, charge temperature, FC timing, drop timing, weight loss, color, and tasting notes. Weight loss and sensory evaluation are especially important because similar-looking curves can taste different, and identical total roast time can still produce one roasty and one clean cup if heat is applied differently source.
For judging outcomes, combine this page with Weight Loss Targets, Color Reading and Measurement, Cupping and Sensory Evaluation, and Roast Defects Troubleshooting.