One variable changed · simulated on the CoffeeScope V3 engine · updated 18 July 2026
Water temperature is the quietest of the big four dials — no change to grind, dose or ratio, just how hot the kettle is. We held everything else fixed and changed only the water temperature: a cool 85 °C against a near-boiling 96 °C, measured in the simulator.
Extraction is a chemical reaction, and reactions run faster when they are hotter. Cooler water pulls solubles out more slowly, so on a fixed recipe it tends to under-extract — a thinner, more sour cup — while hotter water extracts more of everything, gaining sweetness and body but risking bitterness if pushed too far. The practical question: on the same recipe, how much does 11 °C actually move the cup, and does cooler water leave real coffee on the table?
Both lanes are the baked Hoffmann Ultimate V60 recipe (15 g, 1:16.7, 600 µm, center pour). The only difference is brew-water temperature: 85 °C vs 96 °C.
{
"brewer": "v60", "doseG": 15, "ratio": 16.7, "grindUm": 600,
"roastLevel": 0.28, "roastAgeDays": 7, "grindSpread": 0.24,
"bloomWaterX": 3, "bloomTimeS": 45, "pourRate": 6,
"pourHeight": 16, "pourPattern": "center", "bedLevel": 0.5,
"pourHeadroom": { "stopMm": 10, "resumeMm": 13 },
"resolution": "medium", "numericBackend": "js",
"waterTempC": 85 // Lane A — cool
"waterTempC": 96 // Lane B — near-boiling
}
Every number on this page is read from these two runs, taken to brew completion on the deterministic V3 engine.
| Measure | 85 °C | 96 °C | Takeaway |
|---|---|---|---|
| Extraction yield (EY) | 17.0 % | 19.4 % | Cool water under-extracts below the ~18 % Gold Cup floor; hot water lands squarely inside it. |
| Strength (TDS) | 1.14 % | 1.30 % | The hotter brew dissolves more into the same water — a fuller cup. |
| Drawdown time | 162 s | 162 s | Identical here: temperature changes reaction rate far more than it changes this bed's flow. |
| Cup volume | 223.5 ml | 223.5 ml | Same beverage volume — temperature shifts what dissolves, not how much water passes. |
| Cup temperature at finish | 67.8 °C | 76.0 °C | The cooler kettle also lands a cooler cup — worth remembering for drinking temperature. |
The taste axes are directional calibrations, not sensory-panel scores. A difference is flagged tasteable when it clears the model's own threshold of 0.045 (0–1 scale).
| Axis | 85 °C | 96 °C | Δ | Tasteable? |
|---|---|---|---|---|
| Sweetness | 0.59 | 0.76 | 0.17 | yes — heat unlocks sweetness the cool cup misses |
| Aroma | 0.88 | 0.96 | 0.08 | yes — hotter reads more aromatic |
| Clarity | 0.84 | 0.77 | 0.08 | yes — the cool cup is cleaner but underdeveloped |
| Acidity | 0.62 | 0.55 | 0.07 | yes — cool water reads brighter / more sour |
| Bitterness | 0.24 | 0.30 | 0.06 | yes — modestly higher hot |
| Body | 0.55 | 0.60 | 0.05 | yes — just over the line |
| Astringency | 0.00 | 0.04 | 0.04 | no — below threshold |
Temperature is a broad lever: six of the seven axes shift tasteably, all in the direction extraction theory predicts. At 85 °C the cup is bright, clean and a touch sour and thin — it is genuinely under-extracted; 96 °C fills in the sweetness and body.
Cumulative extraction yield, sampled during the brew:
| Time | 85 °C — EY | 96 °C — EY |
|---|---|---|
| 60 s | 8.7 % | 10.6 % |
| 120 s | 14.7 % | 17.1 % |
| Final | 17.0 % | 19.4 % |
The hot lane is ahead from the very first pour and never gives the lead back — the cooler brew is playing catch-up the whole way and still finishes short.
Each extraction channel in the model carries an Arrhenius temperature dependence: raise the water temperature and the rate constant climbs, so more of every soluble class comes out in the same contact time. Because the bed geometry and drawdown are unchanged, that extra rate translates directly into higher yield (17.0 % → 19.4 %) and TDS. The taste readback follows: sweetness and body build as extraction moves up toward the balanced window, while the cooler cup's higher acidity and clarity are the signature of under-extraction — bright and clean, but missing the fuller flavours that only arrive with heat. See the methodology for the Arrhenius extraction kinetics and heat-transport model.
Load this exact A/B in the live simulator and watch the temperature field drive the extraction field:
Related experiments: Fine vs coarse grind Light vs dark roast 1:14 vs 1:17 ratio