A Pattern Emerging in Roasts 28–38
Over the past ten roasts, something interesting began to appear in the data.
At first it felt like the roasts were simply getting easier. The machine seemed more predictable. The roast outcomes felt more repeatable.
Looking more closely at the numbers, it turns out there is a clear reason for this.
The Gene Café itself hasn’t changed.
But the starting conditions of the roaster have become consistent.
Once that happened, the machine started behaving much more like a stable system.
The Dataset
The following roasts provide a useful comparison window.
| Roast | Bean | Start Temp | TRT | Weight Out | WL |
|---|---|---|---|---|---|
| 28 | Santos | 21°C | 13:10 | 210.5 g | 15.8% |
| 29 | Santos | 61°C | ~12:45 | 210.9 g | 15.6% |
| 30 | Santos | 63°C | 12:45 | 208.1 g | 16.8% |
| 31 | Santos | 63°C | 12:30 | 209.8 g | 16.1% |
| 32 | Santos | 21°C | 13:20 | 209.5 g | 16.2% |
| 33 | Santos | 63°C | 13:05 | 204.8 g | 18.1% |
| 34 | Kenyan | 23°C | 13:00 | 209.5 g | 16.2% |
| 35 | Kenyan | 21°C | 12:40 | 210.5 g | 15.8% |
| 36 | Santos | 20°C | 13:11 | 210.4 g | 15.8% |
| 37 | Santos | 71°C | 12:41 | 208.1 g | 16.8% |
| 38 | Jess | 21°C | 12:28 | 213.5 g | 14.6% |
Several patterns begin to emerge immediately.
The First Pattern: Cold Starts Are Predictable
Looking at the cold start roasts:
| Roast | Start Temp | TRT | WL |
|---|---|---|---|
| 28 | 21°C | 13:10 | 15.8% |
| 32 | 21°C | 13:20 | 16.2% |
| 36 | 20°C | 13:11 | 15.8% |
These results are extremely close.
In fact, they are close enough to suggest something important:
When the Gene Café starts cold, it behaves in a very consistent way.
Typical cold-start behaviour in this setup:
- 200°C reached around 4:10–4:40
- 230°C reached around 7:10–7:40
- total roast time around 13:10
- weight loss around 15.8–16.2%
This level of repeatability is surprisingly strong for a small home roaster.
The Second Pattern: Warm Starts Change Everything
Now look at the warm starts.
| Roast | Start Temp | TRT | WL |
|---|---|---|---|
| 29 | 61°C | ~12:45 | 15.6% |
| 30 | 63°C | 12:45 | 16.8% |
| 31 | 63°C | 12:30 | 16.1% |
| 33 | 63°C | 13:05 | 18.1% |
| 37 | 71°C | 12:41 | 16.8% |
When the drum begins warm, the roast progresses much faster.
The machine already contains thermal energy, which compresses the early roasting phases.
Typical warm-start behaviour:
- drying phase accelerates
- Maillard reactions begin sooner
- development begins earlier
- weight loss increases quickly
In other words, the roast curve shifts forward.
If roast time is not adjusted, the beans will roast further than intended.
The Hidden Variable: Starting Drum Temperature
The key difference between these two groups is simple:
starting drum temperature.
This variable is rarely discussed in home roasting, but in practice it turns out to be extremely important.
Two roasts with identical settings can behave very differently depending on whether the roaster begins at:
- 20°C
or - 60–70°C
The Gene Café appears particularly sensitive to this.
Why This Matters
Once the starting conditions became stable in this project, something interesting happened.
Roasts started becoming easier to predict.
Instead of feeling like each roast was a new experiment, the machine began behaving like a system with known boundaries.
That opens the door to a more structured approach.
Two Operating Modes
Based on the data so far, the roaster appears to operate in two practical modes.
Cold Start Mode
Starting drum temperature:
~20–22°C
Typical behaviour:
- 200°C ~4:10–4:40
- 230°C ~7:10–7:40
- TRT ~13:10
- WL ~15.8–16.2%
This has become the baseline roasting condition.
Warm Start Mode
Starting drum temperature:
~60–70°C
Typical behaviour:
- faster drying phase
- earlier Maillard
- earlier development
- increased weight loss for the same TRT
In this situation, roast time must be shortened to maintain the same roast level.
What This Enables
Once this behaviour is understood, it becomes possible to think about roasting differently.
Instead of chasing:
- crack timing
- colour guesses
- intuition
you can control the roast using three variables:
- Batch size
- Starting drum temperature
- Total roast time
These three together give surprisingly good control of roast outcome on the Gene Café.
Where Roast 38 Fits
Roast 38 was the first full-batch roast of the Colombia Regional Jess (Huila washed blend).
Key metrics:
| Metric | Value |
|---|---|
| Batch | 250 g |
| Start temp | 21°C |
| TRT | 12:28 |
| Weight loss | 14.6% |
This roast was intentionally stopped earlier to produce a lighter profile.
It now serves as the reference roast for this coffee.
A Broader Lesson
One of the most useful discoveries in this project so far is simple.
The Gene Café CBR-301 is far more repeatable than many people assume.
When the starting conditions are consistent:
- the heat behaviour stabilizes
- the roast phases align
- the outcomes become predictable
Understanding the system matters more than chasing profiles.
Closing Reflection
This project is not about finding the “perfect roast”.
It is about understanding what is happening inside the machine.
Roast by roast, the patterns slowly reveal themselves.
And sometimes the most useful discoveries are not about the coffee at all.
They are about the behaviour of the system doing the roasting.