Bean Temperature Profiling

Bean temperature profiling on ROEST is useful, but it is not the same as measuring the internal temperature of the bean. The BT trace is a sensor signal shaped by bean mass, airflow, RPM, probe placement, batch size, software control, and machine generation. This page explains when BT is useful for profile control, when it should be treated cautiously, and how to use it without over-reading the graph.

What ROEST BT Represents

ROEST bean temperature is best treated as a repeatable machine signal, not a literal internal bean temperature. The probe can read a mixture of bean surface temperature, hot air, and contact with the bean pile; it is also delayed by probe response and affected by the physical behavior of the bean mass. Denis explicitly described ROEST BT as surface-related rather than internal bean temperature, and Christopher Feran described RoR as a derivative of BT rather than an independent physical measurement 2 sources.

Turning point is especially easy to overinterpret. Christopher Feran described it as a measurement artifact: the point where the preheated probe and cooler green coffee reach equilibrium before heating together again; thinner or more responsive probes and higher air speeds make it occur sooner 2 sources. For broader interpretation of the derived RoR trace, see Rate of Rise Management.

The main practical consequence is that absolute BT numbers should not be transferred blindly between machines, batch sizes, or models. ROEST units can differ by several degrees in BT readings, and sensor placement changes on Ultra/L200-class machines make curves and settings not directly comparable to S100/L100-style data 2 sources.

Batch Size Determines Whether BT Is Actionable

Batch size is the dominant condition for deciding how much trust to place in BT. Low batch weights leave the probe less consistently buried in beans, so the signal can become a mixture of air and bean contact. Higher batch weights give more stable BT behavior, but can create other constraints around ET/exhaust probe usefulness, RPM limits, and bean movement.

For S100/L100-style machines, 50–100g batches are repeatedly described as unreliable for BT/RoR-driven decision-making. At 50g, BT and RoR should generally be ignored and the roast should be controlled by inlet, ET, time, visual cues, and taste; first crack may appear around 225–230°C BT on small batches, which is a sensor-context number rather than a transferable “true” crack temperature 2 sources. At 100g, BT can produce useful clues but is often unstable enough that a good roast may have an ugly graph; this is especially true with higher RPM or higher airflow, where the probe reads more air 2 sources.

Before applying any 120g+ batch-size guidance, stay within the rated capacity of the specific ROEST model and configuration. S100 users should not exceed their machine’s specified maximum batch size; larger 150–185g guidance applies only to models or configurations rated for those loads. From roughly 120–125g upward, where the machine is rated for that batch size, BT becomes more usable. Several contributors report that ROEST or the ROEST team recommended 125g as a strong practical weight, especially when splitting 250g samples into two roasts, and Denis recommends a minimum of 120g for bean-temp/inlet profiles because BT must be accurate enough to modulate inlet 2 sources. Around 150–185g, on models/configurations rated for those loads, BT readings are commonly treated as much closer to useful roast feedback; Denis described 150–185g as more accurate for BT and better for consumption roasting, while Patrick noted that graphs make the most sense at 120–150g 2 sources.

Ultra/counterflow machines are a special case. Multiple contributors report improved BT behavior at lower batch sizes in counterflow or Ultra contexts, but the probe positions and airflow behavior differ enough that Ultra data should not be applied directly to S100/L100 profiles 2 sources. For model-specific notes, see ROEST Ultra Guide.

Practical Workflow and Starting Rules

The safest BT workflow is to choose a batch size where BT is usable and within the rated capacity of the machine, keep that batch size fixed, and use BT as a relative control signal rather than a universal temperature truth. This section is the canonical procedure for applying BT on ROEST; detailed batch-size tradeoffs live in Batch Size Scaling.

Batch-size rules for BT use

Recommended BT-based procedure

1. Select one batch size and keep it fixed. For BT/inlet or BT/PWR work, start at 125g minimum where that is within the machine’s rated capacity; for drinking/consumption roasting, 150–185g is the more stable working range on models/configurations rated for those loads 2 sources.

2. Build the first profile around inlet or power, not BT-only PID. ROEST’s strength is often described as inlet profiling, while BT-only PID control depends on both good PID tuning and a responsive BT probe source. Use Inlet Temperature Management or Power Curve Strategies for the primary heat plan.

3. Use BT as an adaptive trigger only when the reading is stable enough. In an inlet/BT profile, inlet changes when BT reaches defined points. Raising a BT trigger applies its associated inlet value later; lowering it applies that value earlier. The roast may speed up or slow down depending on whether the moved inlet step increases or reduces heat 2 sources.

4. Compare early BT timestamps when transferring a profile. A practical transfer check is to compare BT at 2, 3, and 4 minutes against the reference log. A large difference indicates the IT/BT relationship needs adjustment rather than blind copying source.

5. Avoid major RPM or airflow changes near first crack if interpreting BT/RoR. RPM and air changes can create fake flicks or wrong graph readings by changing how much air the BT probe sees source. See Drum Speed / RPM Settings and Airflow and Fan Settings.

6. Use drop temperature as a repeatability anchor, then verify by color and cup. Denis’s workflow is to drop at a chosen BT, test color, cup, brew, and then raise or lower drop temperature based on taste and color rather than rewriting the entire profile source. Drop decisions are covered in more detail in Development Time and Drop Decisions.

7. Treat first crack as a cue, not the main control setpoint. ROEST automatic first crack marking is based on crack count thresholds, and crack audibility varies by coffee and process. For BT-based workflows, first crack should be cross-checked but not treated as more consistent than BT drop temperature 2 sources. See First Crack Management.

Profile Types That Use BT

BT-only PID profiles

BT-only profiles ask the machine to follow a bean-temperature curve directly. They are the most intuitive in theory, but the most sensitive to sensor delay, PID tuning, and unrealistic target curves. Multiple contributors report overshoot, lag, or power saturation when BT curves are not realistic or when PID parameters are poorly matched.

Denis shared BT PID constants of P 15, I 2, D 20 and described those constants as determining how the heater follows a profile curve source. However, he later described BT PID as wasted time for his own workflow and recommended inlet profiles or BT/inlet instead because BT is influenced by batch weight, probe delay, bean shape, volume, and agitation source.

BT/PWR profiles

BT/PWR profiles apply power changes at BT points rather than time points. The concept is attractive because it can adapt power changes to how quickly a coffee is moving through the roast, but it still depends on a usable BT signal. This can be helpful above the lowest batch sizes and less reliable when the probe is reading too much air.

IT/BT or inlet/BT profiles

IT/BT profiles use BT as the trigger axis and inlet temperature as the controlled output. In practice, they are often treated as adaptive inlet profiles: when the roast reaches a BT point, the machine moves toward the associated inlet value. Denis reported strong results using inlet/BT profiles and described changing only drop BT, and sometimes charge temperature, once a profile concept is established source.

Sorin summarized IT/BT as a consistency tool after finding the best approach with IT, while Christopher noted that IT/time and IT/BT profiles can be modified to become effectively identical if the heat application is matched 2 sources. This makes IT/BT most useful after a roaster has already established a good inlet shape.

Reading BT Around First Crack and Drop

BT often becomes hardest to interpret near first crack because moisture release, crack audibility, airflow, RPM, and probe contact all interact. Some coffees barely crack audibly; some processed coffees do not crack clearly; and automatic crack detection can miss audible cracks or trigger from early isolated cracks. Because of this, BT drop temperature and cup/color feedback can be more repeatable than development time counted from an inconsistent FC mark.

For larger and more stable batches, first crack often appears in a lower BT range than on small batches. Denis described 150g+ first crack as commonly around 190–195°C in his context, while small 50g batches can show first crack much higher, around 225–230°C BT 2 sources. These numbers are machine- and profile-dependent; they should be treated as local landmarks, not universal chemistry.

A useful rule is to record the BT drop temperature that produced a desired color and cup, then adjust future drops near that value. Denis suggested that once a liked result is found, drop temperature often works across many beans with small exceptions of about 1–1.5°C depending on bean shape and size source.

Machine, Probe, and Setup Variables

BT can shift when hardware changes. Probe placement differs between Ultra and older models, counterflow changes measurement behavior, and some machines have different sensor positions or probe arrangements. ROEST temperature placements may be consistent within certain machine generations, but changes to fans, impellers, heating elements, and probe locations can affect readings 2 sources.

RPM affects BT by changing bean contact and air exposure. Higher RPM can expose coffee to heated air more efficiently while reducing bean-mass density around the probe, so the BT sensor can pick up more hot air and less direct bean contact 2 sources. Lower RPM can improve probe contact in some contexts, but can also change roast dynamics and evenness; see Drum Speed / RPM Settings.

Airflow and pressure also affect how the probe behaves, but not always in intuitive ways. Temperature-based profiles can increase power draw when airflow rises because the heater must heat more air, while in manual or power profiles increasing airflow can drop temperatures source. Pressure-specific setup belongs in Pressure Management.

Profile Sharing and Calibration

BT-based profiles are less portable than they appear. A shared curve may fail because another machine reads BT differently, has different inlet behavior, uses different voltage, runs at different heater fan settings, has a different batch size, or is roasting a coffee with different heat absorption. Inlet profiles are often easier to share than BT/IT profiles, but even inlet profiles may need offsets.

A practical sharing approach is to treat shared BT profiles as shapes, not prescriptions. Match the batch size, compare early BT timestamps, verify the actual yellowing/FC/drop behavior, and then adjust inlet offsets or BT trigger points. For broader profile transfer guidance, see Profile Sharing and Starting Points and Calibration and Environment.

Common Misreads