Gene Café CBR-301 | Active Exhaust Baseline

In / Out: 250 g → 214.3 g
Weight Loss: 14.3%
Set Temp: 250 °C
End Decision: First crack + 80 seconds
Cooling: Initiated deliberately at planned stop


Why this roast exists

After an extended sequence of naturals, this roast was designed as a structural reset.

Not a contrast roast.
Not a development push.
Not a Lazy Mode candidate.

A calibration.

The purpose was simple:

After many naturals, how does washed structure feel again?

Specifically:

  • Does acidity present more clearly?
  • Does clarity feel different from sweetness?
  • Does structure show up without flavour labels?

Plan (Pre-defined Guardrails)

  • 250 g batch
  • 250 °C set temperature
  • No preheat
  • Standard airflow progression (fan 2 → 3, system-driven)
  • Development target: 75–95 seconds post first crack
  • Weight loss target: 14–15%
  • Total time guardrail: ~12:30–12:50 window (cool at intent, not fear)

No mid-roast intervention was planned.


What happened

First crack was clearly audible.

The roast was ended at first crack + 80 seconds, consistent with the planned development band.

Colour progression was described as “text book” in the log, resolving post-cooling to an even medium brown.

No manual airflow changes were required.
The system behaved predictably.

The stop decision was bounded:

  • Crack used for orientation only.
  • Fixed development window applied.
  • Colour and aroma used as confirmation.
  • No extension “just in case.”

Structural Position

At 14.3% weight loss, Roast 19 sits cleanly inside the intended washed baseline band.

It is:

  • More restrained than development-forward naturals.
  • Structurally lighter than your upper natural guardrails.
  • Fully within the 14–15% washed calibration window.

This makes it a valid washed reference for this system.


What this confirms

Roast 19 reinforces several stabilised patterns:

  • Weight loss remains the primary structural anchor.
  • Crack audibility is descriptive, not decisive.
  • Washed coffees behave more cleanly and predictably than many naturals under the same airflow conditions.
  • A fixed post-crack development window can be executed calmly without chasing sound.

Most importantly:

The roast felt controlled.


What this roast is not

It is not:

  • A development contrast.
  • An optimisation attempt.
  • A new guardrail.
  • A Lazy Mode default.
  • A video roast.

It is a calibration reference.


Role in the dataset

Roast 19 marks a reset point in the project timeline.

After learning how naturals behave differently across origins and structures, this roast re-centres perception around a washed coffee under stable system conditions.

At 14.3% WL, it becomes:

Washed Baseline — v1

Not dramatic.
But stable.