Eleven Ways to Prepare an Espresso Puck: Rigorous Measurement on a Home Lever Machine

In this paper, I compare eleven espresso puck-preparation techniques (a wedge distributor, three needle-and-tamp protocols, and seven robot-arm shaking programs forming a kinematic basis set) in 110 manually pressure-profiled extractions on a home lever machine. Screening on the pressure-time integral, reproduced to 1.9 % by the manual lever, enforces input equality and certifies a human arm as a controlled pressure source. Mean extraction yield and strength differ, but every significant pairwise difference crosses a roast-batch boundary carrying a comparable systematic, so conclusions are robustness statements: across every batch assumption, zone convention, and input-screening band examined, a horizontal sweep with the cup axis aligned with the direction of travel (V7) is the only preparation that holds first or second place on a reader-weighted preference map, and it is the recommendation that survives. The double-tamp question is answered in the negative: the one-sided 95 % upper confidence bound on the extraction-yield difference, double minus single tamp, is -0.12 percentage points (+0.04 after multiplicity correction), with no detectable spread effect. The kinematic contrasts are bounded null at ten runs per technique, and a second experiment targeting the popular densification hypothesis is a quantified negative: the year-stored pellet tracks dissolved solids, any residual preparation signal below 0.11 percentage points at 95 % confidence. Three practical findings follow: (1) distribution, not tamping, is the active lever of puck preparation, and V7 with a single tamp is the batch-robust choice; (2) grind consistency, not machine cost, sets the reproducibility ceiling, favoring investment in the grinder over the machine; (3) preparation effects match the roast-batch systematic in size, so roast consistency must be controlled alongside technique. The dataset and deterministic pipeline are deposited; the study is offered as a model of honest measurement practice on an accessible system.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-13
DOI
https://doi.org/10.5281/zenodo.24208058
Primary Topic
Coffee research and impacts
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Eleven Ways to Prepare an Espresso Puck: Rigorous Measurement on a Home Lever Machine

Maximilian Alberth
1 links
Zenodo (CERN European Organization for Nuclear Research)
Coffee research and impacts
preprint

Eleven Ways to Prepare an Espresso Puck: Rigorous Measurement on a Home Lever Machine

Maximilian Alberth
preprint en

Abstract

In this paper, I compare eleven espresso puck-preparation techniques (a wedge distributor, three needle-and-tamp protocols, and seven robot-arm shaking programs forming a kinematic basis set) in 110 manually pressure-profiled extractions on a home lever machine. Screening on the pressure-time integral, reproduced to 1.9 % by the manual lever, enforces input equality and certifies a human arm as a controlled pressure source. Mean extraction yield and strength differ, but every significant pairwise difference crosses a roast-batch boundary carrying a comparable systematic, so conclusions are robustness statements: across every batch assumption, zone convention, and input-screening band examined, a horizontal sweep with the cup axis aligned with the direction of travel (V7) is the only preparation that holds first or second place on a reader-weighted preference map, and it is the recommendation that survives. The double-tamp question is answered in the negative: the one-sided 95 % upper confidence bound on the extraction-yield difference, double minus single tamp, is -0.12 percentage points (+0.04 after multiplicity correction), with no detectable spread effect. The kinematic contrasts are bounded null at ten runs per technique, and a second experiment targeting the popular densification hypothesis is a quantified negative: the year-stored pellet tracks dissolved solids, any residual preparation signal below 0.11 percentage points at 95 % confidence. Three practical findings follow: (1) distribution, not tamping, is the active lever of puck preparation, and V7 with a single tamp is the batch-robust choice; (2) grind consistency, not machine cost, sets the reproducibility ceiling, favoring investment in the grinder over the machine; (3) preparation effects match the roast-batch systematic in size, so roast consistency must be controlled alongside technique. The dataset and deterministic pipeline are deposited; the study is offered as a model of honest measurement practice on an accessible system.

Zenodo (CERN European Organization for Nuclear Research)
University of Southern Mississippi (US)
Coffee research and impacts
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon