Cooling and Between-Batch Protocol
Cooling and between-batch protocol determine whether a Roest roast session repeats cleanly from batch to batch. The practical question is not only how fast the roasted beans cool, but how reliably the roaster returns to the same thermal state before the next charge. This page summarizes community guidance for bean cooling, drum-temperature reset, BBP setup, idle cooling fan behavior, and shutdown workflow.
What the Between-Batch Protocol Is For
A between-batch protocol, or BBP, resets the roaster to a repeatable starting condition after one roast and before the next. The main control variable is drum temperature: even though Roest is an air-dominant roaster, consistent drum temperature is a useful check that warmup and between-batch reset are working. Inlet or air temperature still matters, because the next roast can start differently if the drum is at the target charge temperature but the air system is not in the same state.
Tom Roest described the ideal BBP end state as: the drum is at charge temperature and the air is at the profile’s preheat temperature source. That distinction is important: a BBP that only cools the drum may still leave the inlet/air side lagging or overshooting at the start of the next roast.
For charge-temperature selection itself, see Charge Temperature Guidelines. For airflow and pressure side effects from cooling fan settings, see Pressure Management and Airflow and Fan Settings.
Default Starting Protocols
The following starting points are the canonical practical settings for this page. They are not universal recipes; they are baselines to test, measure, and adjust for the roaster, batch size, and profile.
| Use case | Starting protocol | Notes |
|---|---|---|
| 50g sample batches | End BBP around 128°C drum, use 90–100°C air target, 50–55% fan, and a BBP around 1:30 or at least as long as the batch takes to cool source. | Built around cooling-tray timing. Christopher Feran also looks for inlet below 100°C and falling before the next charge source. |
| General programmable BBP | Choose the target charge drum temperature, set the profile preheat and BBP high temperature to the profile’s preheat value, and use a lower cooling step around 100°C if it fits the workflow source. | Tom Roest’s example used 130°C charge, 196°C preheat/high BBP temperature, and 100°C as the low cooling temperature. |
| Manual cooling-profile workflow | After drop, run a cooling profile with 0% power and 100% fan for at least 90 seconds; after bean cooling ends, restart the roast profile and charge when the intended ET/DT condition is reached source. | Denis reported loading at 200°C ET with the drum around 115–120°C in that workflow. |
| 150–185g sessions | Expect a longer reset than for small batches; Denis reported 4:30–5:00 total time for cooling, weighing, entering numbers, and reheating, and elsewhere described needing about 4–5 minutes to reset after high inlet profiles 2 sources. | Larger batches retain more heat and make bean cooling more important. See Batch Size Scaling. |
| P3000-style drum-triggered reset | Set inlet to 40°C and fan to 100% between batches, then start the next roast when drum temperature falls to the intended charge temperature source. | Christopher Feran reported 1:20 BBP, with 1:00 possible when pushing more aggressively source. |
A repeatable workflow should follow the same sequence every time:
- End the roast and drop the beans.
- Cool the beans fully before starting the next roast; do not plan to roast and cool simultaneously.
- Run the BBP or manual cooling profile until the drum reaches the intended reset condition.
- If the reset intentionally cools below charge, reheat back to the target charge condition before loading.
- Charge at the same drum temperature for the same profile.
Multiple contributors discouraged roasting and cooling simultaneously, because it changes pressure, airflow, and workflow timing 2 sources.
Drum Temperature, Sensors, and Repeatability
A drum-temperature reading makes BBP more accurate. Without it, a time-based BBP can still be used, but it is only an approximation because the same ET or inlet condition can correspond to different drum temperatures. Denis explicitly described time-predicted BBP on S100 as less accurate and tied accurate BBP to drum-temperature feedback source.
A useful BBP usually either cools directly to the target charge drum temperature, or cools below target and reheats back to it. The second pattern can be useful when the roast ends much hotter than the desired starting condition, because it flushes stored heat before stabilizing. Changing the starting drum condition changes the finish; inconsistent starts can make later batches darker, faster, or less comparable even when profile settings look identical.
For profiles that rely heavily on BT behavior, see Bean Temperature Profiling. For RoR consequences of inconsistent starts, see Rate of Rise Management.
Bean Cooling: Stock Tray, External Coolers, and Batch Size
Bean cooling and drum reset are related but separate. The beans need to stop roasting; the machine also needs to return to the same thermal state. A BBP should therefore be at least long enough to cover bean cooling, especially when using the built-in tray.
Stock cooling is generally treated as adequate for smaller batches. Denis described the stock tray as cooling 50–120g in about 60–90 seconds, while his 2kW vacuum setup cooled 50–120g in 25–30 seconds source. In another measurement, he reported 185g with the stock trier at 44°C after 1:50, while his external setup reached about the same bean temperature in 30 seconds source.
External cooling becomes more attractive as batch size rises or when very light roasts need a sharper stop. Common community solutions include mesh trays on fans, standalone bean coolers, and vacuum-based coolers using a hopper, mesh, bucket, or colander. curated If using a vacuum-based external cooler, use a heat-tolerant setup that keeps hot beans and chaff out of the vacuum motor and dust bag or canister; hot chaff can be a fire risk, so use metal or heat-safe collection parts, avoid ordinary household-vac contact with hot roast debris, monitor for smoke or embers, and supervise the cooler until beans and chaff are no longer hot. Denis repeatedly reported vacuum-based cooling of 150–200g batches in roughly 30–40 seconds source.
contributors disagree on how adequate the stock tray is for larger batches. Denis found stock cooling acceptable for 50–100g but too slow for 150–200g, while Sorin found the default old tray effective for 185g batches when set up well 2 sources. The practical resolution is to measure actual bean temperature and taste results rather than assume one answer applies to every tray, batch, roast degree, and workflow.
Cooling Fan Idle and Pressure Effects
The bean cooling fan idle setting affects more than bean cooling. It can change airflow and pressure inside the machine, so it belongs in the same control system as exhaust, inlet, and BBP behavior. The stock or factory default discussed in the community was 30% idle (Inferred from conv_1406606919858127000_1413779579704836116; conv_1406606919858127000_1471247456061751347).
idle fan settings are contested. Sam recommended idle 0% because the cooler fan changes airflow inside the machine, while Denis tried changing from stock 30% to 0%, found his roasts worse, and returned to 30% while roasting 2 sources. This is not just preference: Denis reported that on Ultra, turning off the bean cooling fan changed pressure from about -17/-19 Pa to -26/-28 Pa at the same power and exhaust percentage source.
When changing idle cooling fan settings, treat it as an airflow/pressure change, not only as a convenience setting. Repeat a known profile after the change and compare first crack timing, roast length, pressure, and cup result. For pressure measurement and fan interaction, see Pressure Management.
Warmup Before the First Batch
The first roast should begin from a known thermal state, not simply from a fixed clock time. Contributors reported warmups ranging from about 10 minutes to 20 minutes, but the more transferable principle is to warm until the drum reaches the same intended charge condition. Denis reported preheating for a minimum of 15 minutes and 20 minutes before a first roast in one workflow source.
Christopher Feran’s warmup approach is to run the roaster empty with a warmup profile, then proceed into BBP before charging; on L100 this can happen automatically after warmup, while on S100 the BBP may need to be run next source. Tom Roest described either idling at the preheat temperature until drum temperature stabilizes at charge temperature, or running an empty batch to heat the drum past charge temperature before entering BBP source.
Shutdown Cooling
Shutdown should not be treated as the same thing as between-batch reset. Between batches, the goal is to return to the next charge condition; at the end of a session, the goal is to cool the machine safely.
A newer firmware workflow was described as moving the roaster into cooling mode, closing the bean door, dropping power to 0%, setting air to maximum, cooling the drum below 50°C, and then allowing shutdown from the switch source. Production planning should include end-of-session cooling time; Denis suggested accounting for 5–10 minutes of cooling when estimating throughput source.
For cleaning after cooling and longer-term maintenance intervals, see Maintenance and Cleaning.
Common Problems and Fixes
| Symptom | Likely cause | Fix |
|---|---|---|
| Later batches are darker with the same development time | Drum was not reset to the same starting condition. | Extend BBP, cool below target and reheat, or charge only when drum temperature matches the intended charge value. A user reported three roasts where the last was much darker and concluded it was not cooled enough before restarting source. |
| Inlet lags or overshoots at roast start | BBP ended with drum at target but air/inlet not stabilized for the next profile. | Use a BBP that leaves drum at charge temperature and air at preheat temperature, or allow an additional stabilization step before charge. |
| Beans continue popping after drop | Roast energy continues into the tray or cooling is not stopping the roast quickly enough. | Improve bean cooling speed, spread/agitate the beans, or use an external cooler for larger/light batches. Contributors reported beans cracking or popping for 5–30 seconds after drop in some cases 2 sources. |
| Pressure changes after changing cooling fan idle | Cooling fan idle is affecting machine airflow and exhaust pressure. | Treat idle fan changes like profile changes; verify with a manometer if pressure matters to the profile, and retune exhaust/fan settings as needed. |
| 100g profiles run faster after several 150g batches | The machine retains more thermal energy from the larger-batch session than the BBP removed. | Avoid switching batch-size families without a stronger reset, or build separate BBPs for the different batch-size ranges. One user observed 100g roasts running much faster after a series of 150g roasts, even after BBP source. |
Throughput Expectations
BBP time is part of roast time. A nominal 6–7 minute roast with a 3 minute BBP is effectively about a 10 minute cycle under ideal conditions, before accounting for workflow delays. Denis estimated that this produces about 6 roasts per hour if everything is perfect, and suggested that 1kg/hour is more realistic for business planning on S100-sized workflows source.
Tom Roest reported a competition espresso workflow using 200g batches, a 7 minute profile, and 90s BBP, equal to about 1.4kg/hour while maintaining his target quality source. The difference illustrates why throughput should be calculated from the full cycle: warmup, roast time, bean cooling, BBP, weighing/logging, and shutdown cooling all count.