Key Metrics Summary

MetricRoast 39Roast 40
CoffeeBrazil Santos (Wanted)Brazil Santos (Wanted)
Batch size250 g250 g
VentilationOven hood extractionOven hood extraction
Start conditionCold startWarm start
Starting temp / ambient20°C75°C
Time to 200°C04:2803:13
Time to 230°C07:3005:46
Total roast time13:1013:10
Weight out209.9 g205.2 g
Weight loss16.0%17.9%
Cooling time07:4208:13

Both roasts were manually cooled at exactly 13:10 to isolate the impact of starting temperature.


Context

These two roasts were performed back-to-back using the same coffee, batch size, and temperature setting on the Gene Café CBR-301.

The only intentional variable was starting temperature:

  • Roast 39: cold start (~20°C)
  • Roast 40: warm start (~75°C)

This experiment was designed to test something that had begun to emerge from earlier Santos roasts:

If total roast time is identical, does a warm start produce a darker roast?

The results strongly suggest yes.


Roast 39 — Cold Start Reference

Roast details

ParameterValue
Batch250 g
Start temp~20°C
200°C reached04:28
230°C reached07:30
Cooling initiated13:10
Weight out209.9 g
Weight loss16.0%

Observations

The roast followed the expected thermal pattern seen previously with Santos beans on this machine.

Drying progressed gradually, and colour development appeared steady. Aromatically the transition followed the familiar sequence:

  • early hay / warm grain
  • then toasted cereal
  • gradually moving toward sweet, nutty roast aromas

The beans exited the roaster a medium brown, visually consistent with earlier Santos roasts in the ~16% weight-loss range.

This roast therefore serves as a useful baseline reference for the bean.


Roast 40 — Warm Start Comparison

Roast details

ParameterValue
Batch250 g
Start temp~75°C
200°C reached03:13
230°C reached05:46
Cooling initiated13:10
Weight out205.2 g
Weight loss17.9%

Telemetry note

During this roast the Gene Café Android app failed to connect at the start of the roast.

The roast had already begun before telemetry logging started. The application required a force close / restart on Android before reconnecting to the roaster.

Because of this:

  • early telemetry is incomplete
  • manual timing was recorded for key milestones

The manual timings for reaching 200°C and 230°C align closely with the telemetry curve once logging resumed, suggesting the remaining data is reliable.

This kind of operational detail is worth recording as part of the project.


Phase Comparison

Time to 200°C

RoastTime
Roast 3904:28
Roast 4003:13

The warm start shortened the drying phase by over one minute.

That is a significant difference in a roast of this duration.


Time to 230°C

RoastTime
Roast 3907:30
Roast 4005:46

Again the warm start advanced the roast by almost two minutes at this stage.

This confirms that the beans enter the Maillard and browning phases earlier when the drum starts warm.


Time above 230°C

Because both roasts ended at the same TRT:

RoastTime above 230°C
Roast 39~5 minutes 40 seconds
Roast 40~7 minutes 24 seconds

The warm-start roast therefore spent nearly two extra minutes at high temperature.

That explains the difference in weight loss.


Weight Loss Comparison

RoastWeight loss
Roast 3916.0%
Roast 4017.9%

A difference of almost 2% weight loss is significant.

In practical roasting terms this likely means:

  • Roast 39 → medium espresso roast
  • Roast 40 → noticeably darker roast

This aligns with earlier Santos experiments where warm starts produced weight losses in the 17–18% range.


Colour and Aroma Expectations

Based on the roast metrics:

Roast 39

Expected profile:

  • balanced sweetness
  • moderate body
  • lower bitterness
  • good espresso baseline

Roast 40

Expected profile:

  • fuller body
  • deeper roast character
  • reduced acidity
  • possible slight bitterness depending on extraction

Cup testing will determine whether the additional development is beneficial or excessive.


Ventilation Observations

Both roasts were performed using the kitchen oven hood extraction system recently installed above the roaster.

Across multiple recent roasts the system has shown:

  • stable airflow
  • reduced smoke accumulation
  • fewer overheating protection events

The hood appears to provide a consistent roasting environment, which is likely contributing to the increasingly predictable roast curves.


Lessons Learned

Starting temperature has a major impact

Even when total roast time is identical, a warm start changes the roast significantly.

This is primarily because:

  • drying finishes earlier
  • browning begins sooner
  • beans spend longer at high temperature

TRT alone is not a universal control

Total roast time works as a guardrail only when starting conditions are similar.

If the roaster begins warm, TRT must usually be shortened to reach the same roast level.


Weight loss continues to be the most reliable endpoint

First crack remains difficult to hear on this system, especially with ventilation active.

Weight loss provides a consistent reference for roast development.


The Gene Café shows consistent thermal behaviour

Across multiple Santos roasts a pattern is emerging:

Cold starts:

  • 200°C ~4:15–4:30
  • 230°C ~7:10–7:30
  • TRT ~13:10
  • WL ~16%

Warm starts:

  • 200°C ~3:00
  • 230°C ~5:45–6:10
  • TRT ~13:10
  • WL ~17–18%

This growing dataset is making the machine’s behaviour increasingly predictable.