Last Updated: February 2026
Roasts Included: 22 total (20 standard + 2 sample roasts)
Roaster: Gene Café CBR-301
Baseline: 250 g | 250 °C | Fan 2 → 3 around ~4 min | No preheat
Context
This note supports the YouTube video:
It is not a tutorial.
It is not a set of instructions.
It documents what stopped surprising me after repeating the same process enough times for patterns to emerge.
Roast Distribution
Across 22 total roasts:
- ~19 roasts at 250 g
- 1 roast at 300 g (capacity test)
- 2 sample roasts at ~50 g
Sample roasts:
- 52.5 g → 45.1 g (14.1% WL) Sample Roast 1 – Fermented and …
- 52.5 g → 44.3 g (15.6% WL) Sample Roast 2 – San Isidro 48 …
The statistics and reflections below focus primarily on the 250 g and 300 g roasts.
What Actually Stabilised
By approximately Roast 15–18, several things stopped moving around:
1. Batch Size
250 g behaves predictably.
300 g slightly shifts timing and energy behaviour but remains manageable.
250 g is now the baseline.
2. Temperature
250 °C set temperature works consistently on this machine for:
- Washed coffees
- Naturals
- Honey processes
I did not need to preheat.
The machine behaves consistently from cold start.
3. The “Four Minute Pattern”
Across almost every roast:
- Around ~4:00 minutes
- Overheat protection triggers
- Fan shifts from 2 → 3
- Roast stabilises
After contacting Gene Café, this behaviour was confirmed as normal.
What initially felt dramatic is simply how the machine regulates heat.
The machine was consistent long before I was.
4. First Crack — Helpful, Not In Charge
Early assumption:
First crack would be the primary decision anchor.
Reality:
- On washed coffees → clear, helpful.
- On naturals → sometimes quiet, smeared, ambiguous.
Adaptation:
- If crack is clear → use it.
- If crack is not clear → use colour, aroma, total time.
- Final guardrail → weight loss + cupping.
Crack became information — not instruction.
5. Weight Loss as Guardrail
After each roast, beans are weighed.
Across the 250 g roasts:
- Balanced cups tended to cluster in a consistent weight-loss band.
- Pushed further → heavy, tiring cups.
- Stopped early → thin, watery cups.
The scale proved more reliable than my ears.
Not because numbers are magical —
but because repetition revealed patterns.
Weight loss became the final confirmation layer.
6. Cupping as Validation
Cupping was not used to score coffee.
It was used to answer simple questions:
- Does this feel balanced?
- Does it have body?
- Is the finish clean?
- Would I drink this again?
Cupping confirmed whether roast decisions made sense.
It did not dictate them.
What Didn’t Matter As Much As I Thought
- Perfect crack timing.
- Exact development percentages.
- Roast curves.
- Micro-adjusting every second.
Simplifying improved consistency.
Controlling less improved clarity.
The Role of Green Coffee
This cannot be ignored.
The stability I’m seeing is not just the roaster.
It is also:
- Consistent green coffee.
- Known suppliers.
- Repeatable quality.
Stable input reduces noise.
Reduced noise makes patterns visible.
If the starting material varies wildly, you are learning two chaotic systems at once.
For learning, stable green matters.
What This Means (And What It Doesn’t)
It does not mean:
- I’ve mastered roasting.
- I’ve optimised the machine.
- There is nothing left to learn.
It means:
- The chaos reduced.
- The reference points became familiar.
- The surprises got smaller.
There was no breakthrough moment.
I simply repeated the process enough times that it started to make sense.
Where This Sits In The Project
- These observations stay in Notes.
- Only repeated, stable patterns move into the Guide.
- Nothing is promoted quickly.
- Nothing becomes “truth” without repetition.
The next 20 roasts matter more than the first 22.
Closing Reflection
After 22 roasts:
I understand this machine better.
I understand roasting at a basic, practical level.
And that’s enough.
The first 22 were about understanding.
The next 20 are about exploring.
Not chasing perfection.
Not rushing.
Just building steadily.
One roast at a time.