Community espresso profile

Dark & Steady 9-6

by jonnvice

A pressure-controlled profile for dark roasts and arabica-robusta blends. Short, defined pre-infusion, a ramp that reliably reaches ~9 bar, then a linear decline toward 6 bar until target weight. WHAT IT'S FOR Dark and medium-dark roasts, robusta blends, supermarket espresso beans. In my cups: chocolatey, balanced, good body, no sharp acidity. Not for light roasts. WHY IT EXISTS I wanted a dark-roast profile that behaves the same way every shot, with as few moving parts as possible: a defined start, a short saturation (dark, rested beans don't need a bloom), a ramp that reliably reaches ~9 bar, and a gentle declining finish. It's pressure-led on purpose – the profile sets the pressure, the grind only moves the shot time. After trying a lot of other profiles, it's the one I keep coming back to for dark beans. WHY THE DECLINE As solubles are extracted, the puck loses resistance. At constant pressure, flow tends to creep up late in the shot, and a channel that has formed gets pushed harder. Lever machines deal with this by letting pressure fall – this profile does the same, with a 5 ml/s cap as a hard stop if a channel opens anyway. That's the design intent; see below for what I can and can't show. HOW IT WORKS - Stabilize (≤3 s): pump off until ≤0.7 bar – defined start, no residual pressure - Fill (≤8 s): 4 ml/s, 4 bar limit, exits at 3 bar - Saturate (3 s): 3 bar hold, no bloom (dark, rested beans are largely degassed) - Ramp to 9 bar (≤10 s): linear ramp to a 10-bar target, exits at 8.7 bar. Safety exits: pumped volume or target weight - Decline 9-6 bar: starts at the measured pressure, declines toward 6 bar over 30 s, 5 ml/s cap as channel protection - 92 °C flat (8 °C offset) THE ONE TRICK: OVERSHOOT TARGET A pressure-controlled ramp approaches its setpoint asymptotically. With target 9 / exit 8.7, the exit was never reached in 2 of 4 shots. With target 10 / exit 8.7 it triggered in 5 of 5. The decline starts from the actual measured value, so the overshoot never leaks into the extraction. TESTED ON Gaggia Classic Pro E24 + GaggiMate Pro (firmware v1.9.0), ULKA vibration pump, stock OPV, temp offset 8 °C, Kff 0.3. Normcore High Extraction basket (double layer, 18 g) with puck screen. DF54 with RDT, WDT. Developed on Morris Espresso 1.1, later confirmed with several dark to medium-dark blends and straight arabica (grind adjusted only). ~30 analysed shot logs across eight versions. LIMITS – AND AN INVITATION One machine, one main development bean, one (unblinded) taster. A Bluetooth scale is required. Rotary pumps, other boilers and other firmware versions are untested. The 18 g variant is untested by me (with a puck screen I can't fit 18 g into my 18 g basket) – same curve, target 36 g, slightly more generous safety exits. Please check that the Ramp exits on pressure (≥ 8.7 bar), not on "pumped". If you try it on a different setup, please share a brew here – especially the ones that don't work. That's how v2 gets better. WHAT I'M NOT CLAIMING I'm not claiming the declining pressure itself makes it taste better, or that it measurably reduces channeling – I haven't run an isolated A/B against flat 9 bar, and puck flow is fairly flat between ~5 and 11 bar, so the effect may be modest. What I can show is a robust, repeatable pressure build-up. The taste is my judgment; yours may differ. CREDITS Stabilize phase and overshoot-target idea inspired by Automatic Pro (modsmthng) and Adaptive v2. Built by me from GaggiMate logs and a lot of iteration.