Extraction science, brewing standards and hardware analysis for people who weigh the dose to a tenth of a gram — labelled bench-tested or research synthesis, never blurred.
I weigh the dose, log the shot, and publish the numbers — or say plainly when the gear isn't mine to measure. Everything here carries one of two labels so you always know which you're reading. How we test →
Every guide treats channeling as a fault you either have or you don't, caused by grind, distribution or tamp. The published physics says something stranger: uneven flow between pathways is always present, it feeds on itself, and for most of a shot it hides behind an average that looks fine. Here is the causal chain, from the first 6.7 seconds of wetting to what the spent puck looks like under an x-ray.
Every page selling you a coffee scale quotes the same number: 0.1 g. That number is readability — how finely the display rounds — and it is not a statement about error. Acaia publishes both figures for its own $250 scale, and they are not the same. Which means the case for weighing coffee has to be made somewhere else, and it can be: the error a scale removes is not a tenth of a gram, it is closer to three grams.
Crema is carbon dioxide that was dissolved under pressure and had nowhere to go when the pressure dropped. That makes it a good gauge of how much gas your coffee still had — roast degree, days since roasting, minutes since grinding — and a poor gauge of everything else. In one experiment the thickest foam collapsed fastest, and the strongest cups had the shortest-lived crema.
Extraction is one number: how much of the dry coffee ended up in the liquid. Around 30% of a roasted bean can dissolve at all, and the famous window uses two-thirds of it. The parts everyone repeats — that flavours come out in a fixed order, that a 30-second bloom is settled science, that a shot should take 25 to 30 seconds — are the parts the research does not support.
The physics and chemistry of the shot: pressure and flow profiling, TDS and extraction yield, puck preparation and channelling.
Open pillar → 3 piecesPour-over, immersion and cold brew — grind, temperature, agitation and the paper itself.
Open pillar → 1 pieceTampers, WDT tools, baskets, scales, cleaning chemistry and the small kit that quietly decides shot quality.
Open pillar →
Most gear sites tell you what they think. This one tells you where the number came from. Every piece carries one of two labels before you read a word of it: research synthesis, meaning the hardware is not mine and every figure is attributed to whoever did measure it — or bench-tested, meaning I bought it, measured it, and published the rig alongside the result. Nothing sits in between, and nothing borrows someone else's numbers without saying whose they are.