Coffee | One Roast at a Time

Roast 32 & Roast 33 — Brazil Santos
These two roasts were deliberately paired.

The intention was simple:
Test repeatability using Total Roast Time (TRT) as the primary control variable, while acknowledging machine thermal state.

What followed was more instructive than expected.

Context

Same coffee.
Same batch size: 250g.
Same set temperature: 250°C.
Same airflow behaviour (fan stepping to 3 around ~200°C).
No mid-roast adjustments.

The only meaningful variable:

  • Roast 32 → Cold start (~21°C ambient)
  • Roast 33 → Warm start (~63°C machine retained heat)

Roast 32 — Cold Start Baseline

Ambient: 21°C
Time to 200°C: 4:31
Time to 230°C: 7:21
TRT: 13:20
Weight Out: 209.5g
Weight Loss: ~16.2%

Observations

  • 200°C arrived slightly later than historical cold baselines (~4:10), suggesting slightly slower early ramp.
  • 13:20 TRT landed cleanly in the intended development band.
  • No visible oil.
  • Structurally consistent with previous 16% range Santos roasts.
  • Cooling cycle normal.

Interpretation

This roast behaved predictably.

Cold start + 13:20 TRT continues to sit comfortably in the 15.8–16.3% WL lane.

This confirms:

  • TRT is stable when thermal state is controlled.
  • 250g remains the most predictable batch size.
  • 200°C arrival is still the best early phase indicator.

Roast 32 reinforces the working Santos guardrail:

Cold start (~20°C) → 13:15–13:20 TRT ≈ ~16% WL.

This is now repeatable across multiple sessions.

Roast 33 — Warm Start Push

Machine Start Temp: ~63°C
Time to 200°C: 2:57
Time to 230°C: 6:06
TRT: 13:05
Weight Out: 204.8g
Weight Loss: ~18.1%

Observations

  • 200°C arrived extremely early (sub-3 minutes).
  • Despite only reducing TRT by ~15 seconds compared to Roast 32, weight loss increased dramatically.
  • New highest WL recorded for Santos.
  • Minor oil spotting visible on isolated beans next morning.
  • Roast did not appear dramatically darker visually.

Interpretation

This roast confirms something important:

Warm start energy compounds faster than intuition suggests.

Although TRT was reduced slightly, the early ramp (200°C at 2:57) front-loaded the roast with significantly more energy. That early acceleration altered internal bean development enough to push weight loss to ~18%.

The relationship is nonlinear.

Shortening TRT by seconds does not linearly offset increased early energy.

This roast defines the upper Santos boundary for this setup.

Key Lessons Reinforced

1️⃣ Thermal State Is Not Cosmetic

Starting at ~60°C is not a minor variation.
It materially shifts phase behaviour.

The 90-second difference in 200°C arrival between Roasts 32 and 33 was decisive.

2️⃣ TRT Is Contextual, Not Absolute

13:05 TRT means something different at:

  • 4:30 to 200°C
    vs
  • 3:00 to 200°C

TRT without phase context is incomplete.

TRT with phase context becomes powerful.

3️⃣ Early Energy Multiplies Late Effect

Roast 33 demonstrates:

Front-loaded energy cannot be fully “undone” with minor TRT reductions.

The system accumulates energy.

This reinforces the working adjustment model:

  • Cold start → ~13:15–13:20
  • Warm start → ~12:40–12:50

13:05 was not enough reduction for a 2:57 ramp.

4️⃣ Oil Spotting = Upper Envelope Marker

Minor oil spotting appeared on isolated beans in Roast 33.

This is not a defect.
It is a signal.

It marks the transition from:
“Upper espresso development”
toward
“Approaching second crack structure.”

That visual confirmation is valuable.

The upper guardrail is now physically defined.

5️⃣ Santos Is Structurally Forgiving

Despite reaching ~18% WL:

  • The roast did not appear dramatically dark.
  • The beans retained structural integrity.
  • Surface oil was limited.

Brazil Santos tolerates development better than some washed Central Americans.

That matters for house espresso calibration.

Evolution Observed

Earlier in this journey:

  • Crack sound was relied upon.
  • Peak temperature felt significant.
  • TRT was considered in isolation.

Now:

  • 200°C timing anchors phase understanding.
  • Thermal state is deliberately considered.
  • TRT is adjusted probabilistically.
  • Weight loss confirms structure.

This is a shift from reactive roasting to controlled testing.

Strategic Outcome

Roast 32 confirms the stable Santos lane (~16%).
Roast 33 defines the upper boundary (~18%).

Between those two:
Your house espresso profile likely lives.

The tasting will decide preference.
The data has already defined the envelope.