Puck preparation
How ground coffee is dosed, distributed and tamped into an espresso basket. The aim is an even bed, so that water passes through all of it instead of channelling through a few paths.
Contents
In brief
- The puck is the compact bed of ground coffee in the basket; its permeability decides how fast water flows through it and at what pressure.1,3
- In one permeability study, packing more coffee into the same chamber (360 to 480 kg/m³) cut bed permeability about three- to fourfold.3
- In one 2017 espresso study, tamping pressure had no detectable effect on the extraction of caffeine and trigonelline; grind size did.4
- Channelling is flow focused in one or a few paths instead of through the whole bed; it lowers extraction yield and reproducibility.6,7
- Needle stirring breaks up clumps in research and in practice, but no peer-reviewed study opened for this entry compared distribution tools by their effect on the cup.6,14
What the puck is#
In espresso, hot water under pressure is pushed through a compact bed of ground coffee held in a perforated basket.1 Baristas call this bed the puck. Puck preparation is everything done to the coffee between the grinder and the machine: dosing, distributing and tamping.
The puck matters because it is the main resistance in the system. According to a coffee science textbook, the bed’s permeability sets the pressure on top of it and the flow through it. Permeability means how easily water passes through.1 Too low a permeability gives high pressure, low flow and long shots that may over-extract.1,3 Grind size is the main way to change permeability; the grinding and particle size entries cover it.1,10
Preparation decides whether that permeability is the same everywhere in the bed. According to the textbook, the pressure drop depends on the degree of tamping and on the homogeneity and shape of the bed. An even bed avoids channelling.1 How the machine delivers water and pressure is covered in the espresso machine entry.
The steps#
Practice varies, but published studies prepare pucks in a similar order. In a 2026 study, the authors chose their routine to match current practice in espresso brewing.6
- Dose. Ground coffee goes into the basket. Café grinders dose by pre-weighed beans, by time or by a load cell that stops at a set mass.2
- Distribute. The grounds are spread evenly and clumps are broken up. In the 2026 study, coffee was shaken for 10 s in a closed container, stirred with a needle tool, and clumps were removed.6
- Tamp. A flat tamper presses the grounds into a level, compact bed.6 According to a grinding chapter, the tamper sets the density and porosity of the bed, as a densifying machine does in industry.2
- Lock in and brew. The first stage is infiltration: water soaks into the initially dry bed.12 Pre-infusion and shot parameters are covered in the espresso recipes and ratios entry.
Dose and basket fill#
The dose sets how much coffee the basket holds, and with it how tightly the bed is packed. A 2026 permeability study notes that portafilters have a standard diameter of 58 mm.5 Its authors conclude that dose is the variable that mainly controls how thick the puck is.5
Packing density changes the flow strongly. In a 2015 laboratory study, 7.2, 8.0 or 9.5 g of coffee was packed into the same 20 mL chamber.3 The densest beds let water through about three to four times less easily than the loosest (see the chart in the Deep dive).3
A basket can also be overfilled. According to a coffee science textbook, too much coffee leaves no room for the bed to expand as it wets.1 The result is overcompaction, disturbed percolation and solids in the cup.1 No study opened for this entry measured the effect of underfilling a basket.
Distribution and clumps#
Clumps. Grinding charges coffee particles electrically, and charged fines and larger particles cling together.11 In a 2024 study, a little water added to the beans before grinding reduced the charge and the clumping at once.11 In the same study, an espresso ground this way ran nearly 50% longer and was stronger.11 Static is covered in the grinding entry.
Stirring with needles. Practitioners stir the grounds with thin needles, often called WDT, to break clumps and even out the bed.6,14,15 In a 2026 flow study, the authors described removing clumps with a needle tool as necessary for good reproducibility.6 One practitioner test reported more consistent shots with full-depth needle stirring than without distribution.14 No peer-reviewed study opened for this entry compared distribution methods.
Tapping and levelling tools. In the 2015 permeability study, beds were tapped 100 times before being compressed with a tamper.3 No study opened for this entry tested levelling tools that spin on the surface of the bed.
Tamping#
Force. How hard to tamp is debated (see Contested). In a 2017 espresso study, tamping pressure had no detectable effect on the extraction of caffeine and trigonelline.4 Particle size, by contrast, affected it significantly.4 A 2026 permeability study, citing the 2017 study, treats tamping force within the range a barista can apply as a second-order effect compared with dose.5 A coffee science textbook lists the degree of tamping among the factors that set the pressure drop (see Contested).1
Level. In the 2026 flow study, the grounds were tamped level.6 Practitioners stress a level tamp rather than a hard one.15 One tamper patent explains why in its background: unevenly compressed coffee lets water flow through a channel and extracts unevenly.13 No study opened for this entry measured the effect of a tilted tamp.
Repeatability. Tampers with force sensors are built to show when a target force is reached.13 In the 2026 flow study, an automatic tamper set to 20 kg was used for repeatable tamping.6
Screens and paper filters#
No peer-reviewed study or textbook opened for this entry described or tested metal puck screens or paper filter discs in espresso, so this entry does not cover them.
Channelling#
Channelling is flow that is focused in one or a few channels instead of passing evenly through the whole bed.6 Parts of the bed are then extracted more than others.8,13 The authors of a 2026 flow study note that channelled shots keep a low extraction yield even at high flow.6
Why small flaws grow. A 2023 model shows that extraction itself makes channelling worse. Dissolving coffee opens the pores, so a path that carries more water becomes more permeable still.8 Small differences in the bed at the start can therefore grow during the shot.8
Evidence from grind. In a 2020 study, extraction yield peaked at a medium grind and fell again at very fine settings. An even-flow model had predicted otherwise.7 The authors attribute this to uneven flow at fine settings, which also explains poor reproducibility.7
Reading a spent puck. A spent puck is not a record of the shot: in the 2026 flow study, cracks formed partly through backflow when the pump stopped.6 Two practitioner tests also found that what a barista can see is a poor guide. Holes in the spent puck did not reliably show bad distribution or tamping.16 Even-looking flow from the basket did not reliably predict yield or taste.14 A refractometer reading of extraction yield is a more direct check (see the extraction yield entry).
Flow through a porous bed#
A packed coffee bed is a porous medium. At low flow, Darcy’s law describes it: the flow rate Q equals k A ΔP / (μ L), where k is permeability, A the cross-section, ΔP the pressure drop, μ the viscosity of water and L the bed height.3 The 2015 study fitted its flow and pressure data to Darcy’s law and found steady-state permeabilities of about 2.6 × 10⁻¹⁴ to 4.4 × 10⁻¹³ m² for arabica beds at 80 °C.3 Its permeabilities fitted a corrected Kozeny–Carman equation, which links permeability to porosity and particle size, once one parameter was fitted to each grind.3
Density matters more than the fit suggests. In that study, four grinds (Sauter mean diameters about 80–131 µm) were each packed at 360, 400 and 480 kg/m³.3 Going from 360 to 480 kg/m³ lowered permeability by a factor of about 3–4 for every grind (our arithmetic from the published table).3 The authors describe permeability as extremely sensitive to bed bulk density.3 At the highest density, the coarsest and finest grinds also differed less than at the lowest (our arithmetic).3
Data table
| Series | Initial bed bulk density (kg/m³) | Permeability (× 10⁻¹⁴ m²) | Source |
|---|---|---|---|
| D, coarsest | 360 | 33.6 | 3 |
| D, coarsest | 400 | 43.8 | 3 |
| D, coarsest | 480 | 8.95 | 3 |
| C | 360 | 23.9 | 3 |
| C | 400 | 19.3 | 3 |
| C | 480 | 6.44 | 3 |
| B | 360 | 13.7 | 3 |
| B | 400 | 11.8 | 3 |
| B | 480 | 4.87 | 3 |
| A, finest | 360 | 7.65 | 3 |
| A, finest | 400 | 4.94 | 3 |
| A, finest | 480 | 2.59 | 3 |
| Grind (Sauter mean diameter) | 360 kg/m³ (7.2 g) | 400 kg/m³ (8.0 g) | 480 kg/m³ (9.5 g) | Source |
|---|---|---|---|---|
| A, finest (79.73 µm) | 7.65 × 10⁻¹⁴ m² | 4.94 × 10⁻¹⁴ m² | 2.59 × 10⁻¹⁴ m² | Corrochano et al.3 |
| B (101.60 µm) | 13.7 × 10⁻¹⁴ m² | 11.8 × 10⁻¹⁴ m² | 4.87 × 10⁻¹⁴ m² | Corrochano et al.3 |
| C (112.86 µm) | 23.9 × 10⁻¹⁴ m² | 19.3 × 10⁻¹⁴ m² | 6.44 × 10⁻¹⁴ m² | Corrochano et al.3 |
| D, coarsest (131.36 µm) | 33.6 × 10⁻¹⁴ m² | 43.8 × 10⁻¹⁴ m² | 8.95 × 10⁻¹⁴ m² | Corrochano et al.3 |
The coarsest grind was the one exception to a steady fall: it was more permeable at 400 than at 360 kg/m³.3 All four grinds came from the same medium-dark arabica, and the doses were packed into a chamber of 20 mL.3
Darcy’s law breaks down at espresso pressure. In a 2026 study, flow through a puck rose linearly with pressure only below about 5 bar.6 At usual brewing pressures the flow saturated: it no longer scaled linearly with pressure.6 The authors explain this by the interplay of the bed’s elasticity and porosity, which they call poroelastic regulation.6 A Kozeny–Carman estimate for 200 µm particles overestimated the flow by a factor of about 1,000; the authors attribute the gap to swelling.6
Porosity. Reported porosities depend strongly on how the bed was packed and when it was measured (see Reference values). Loose, untamped grounds in a 2026 imaging study had a total porosity of about 0.48–0.69 and permeabilities of about 1.6 × 10⁻¹¹ to 1.9 × 10⁻¹⁰ m².5 These are roughly 35 to 7,000 times higher than the wet, compacted beds of the 2015 study (our arithmetic).3,5 For espresso pucks, the 2026 authors assumed a connected porosity of 0.3–0.5, depending on tamping and the spread of particle sizes.5 In the same study, porosity did not vary systematically with grind setting.5
Close to turbulence. The 2026 imaging study estimated a Forchheimer number of about 0.016–0.064 for espresso conditions.5 That is laminar flow, but close to the regime where inertia starts to matter.5 Finer grinds or tighter packing could push a shot into it.5
How the bed changes during a shot#
The puck at the end of a shot is not the puck that was tamped.
- Infiltration. A 2025 study followed the water front into a dry espresso bed with time-resolved X-ray tomography.12 A one-dimensional model of flow into an unsaturated porous medium, with wet and dry regions, agreed well with the measurements.12
- Swelling and consolidation. As the particles wet, they swell and rearrange under flow and pressure; porosity falls, flow drops and the pressure drop rises.1 In the 2015 study, beds consolidated as soon as water flowed: their height fell by 0–31%, least for the coarsest, densest bed and most for the finest, loosest one.3
- Fines. More fines lower the permeability of the bed and lengthen the shot.10 Fines migration, the idea that water carries fines down to the filter, would lower porosity at the bottom.9 In a 2023 tomography study of capsule beds, porosity after extraction always decreased from the outlet at the bottom to the inlet at the top: the opposite profile.9
- Cracks after the shot. In the 2026 flow study, tomography of a spent puck showed horizontal delaminations and cracks, and the puck had peeled off the filter mesh.6 The authors link this partly to backflow through the machine’s three-way valve when the pump stops.6
- Stopping and restarting. In the same study, restarting the brew several times raised the flow without extracting more.6 The authors read this as the bed reorganising each time pressure is applied, and advise against brews with pauses or repeated pressurisation.6 They did not test pre-infusion as such.
Channelling as a feedback#
Two lines of work explain channelling. The 2020 study combined an extraction model that assumes even flow with measurements on real shots.7 Measured yield peaked at an intermediate grind and fell at finer settings, while the model predicted yield to keep rising as the grind got finer; the authors take the gap as evidence of uneven flow at fine settings.7
The 2023 model gives a mechanism. It treats the bed as two parallel paths of slightly different porosity.8 Because extraction raises porosity and permeability, the faster path opens further and takes a growing share of the water.8 In the model, the drop in yield at fine grind comes from one part of the bed being completely extracted.8 The authors suggest making the flow as even as possible as the first thing to check.8
What this means for preparation follows from the model but has not been tested step by step. An even starting bed gives the feedback less to amplify.8 No study opened for this entry measured how much each preparation step reduces channelling.
Reference values#
| Parameter | Value reported | Context | Source |
|---|---|---|---|
| Bed permeability (wet, packed) | ≈2.6 × 10⁻¹⁴ to 4.4 × 10⁻¹³ m² | Four grinds, 360–480 kg/m³, 80 °C, steady state | Corrochano et al.3 |
| Packing density effect | 360 → 480 kg/m³ lowered permeability ≈3–4-fold | 7.2, 8.0, 9.5 g in a 20 mL chamber, 37 mm diameter | Corrochano et al.3 |
| Bed consolidation under flow | 0–31% of bed height | Coarser grinds consolidated less | Corrochano et al.3 |
| Steady-state porosity | 0.12–0.33 | Same study; calculated from dry-grain density, corrected for consolidation | Corrochano et al.3 |
| Porosity, untamped dry packs | ≈0.48–0.69 (total) | X-ray microtomography, two coffees, 11 grind settings, not tamped | Wadsworth et al.5 |
| Permeability, untamped dry packs | ≈1.6 × 10⁻¹¹ to 1.9 × 10⁻¹⁰ m² | Simulated on imaged packs | Wadsworth et al.5 |
| Connected porosity, espresso puck | 0.3–0.5 (assumed) | Depends on tamping and particle spread | Wadsworth et al.5 |
| Forchheimer number | ≈0.016–0.064 | Espresso conditions; laminar, near inertial | Wadsworth et al.5 |
| Darcy-like flow | Below ≈5 bar | Flow saturates at usual brew pressure | Waszkiewicz et al.6 |
| Research puck preparation | 18.5 g, 58 mm basket, 14 mm bed; 10 s shaking, needle stirring, 20 kg automatic tamp | One flow study | Waszkiewicz et al.6 |
| Tamping pressure | No detectable effect on caffeine and trigonelline extraction | Fine espresso grinds; pressures in full text | Kuhn et al.4 |
| Basket diameter | 58 mm (industry standard) | Portafilters | Wadsworth et al.5 |
| Water before grinding | Shot nearly 50% longer, stronger | Less static and clumping | Méndez Harper et al.11 |
Open questions#
- Each preparation step. No peer-reviewed study opened for this entry compared distribution methods, levelling tools, puck screens or paper filters by extraction yield or sensory test. Practitioner tests exist but are not independent or peer-reviewed.14
- Tamping force. One study found no detectable effect on two compounds4; whether force changes flow time, aroma or sensory scores at fixed dose was not measured in any study opened for this entry.
- Level tamping. The effect of a tilted bed on channelling has been argued from mechanism but not measured in any study opened for this entry.13,15
- Fines migration. The one imaging study found a porosity profile opposite to migration, but in capsules.9 Tamped baskets have not been imaged in the same way in any study opened here.
- Seeing channelling. Channelling is inferred from flow and yield data or from spent pucks; one study notes it had no visualisation of the bed reorganising during brewing.6
Key studies#
Corrochano et al., 2015 – Permeability of coffee beds. Packed four grinds at three densities into a small brewing chamber and measured steady-state permeability with water at 80 °C. Permeability was extremely sensitive to packing density, and the beds consolidated as soon as water flowed.
Kuhn et al., 2017 – Grind and tamping pressure. Followed the extraction of caffeine and trigonelline through espresso shots at high time resolution, varying particle size and tamping pressure. Finer grinds extracted more per gram of beverage; tamping pressure had no detectable effect.
Lee et al., 2023 – Uneven extraction. Models two flow paths through an espresso bed with different porosity. Extraction opens the faster path further, so small differences grow during the shot and can explain why yield falls at very fine grinds.
Waszkiewicz et al., 2026 – Poroelastic flow in espresso. Measured long-time flow through standard-prepared pucks at pressures from about 1 to 12 bar and imaged spent pucks. Flow stopped scaling with pressure at usual brew pressures, spent pucks showed cracks, and restarting a brew raised the flow without extracting more, which the authors read as the bed reorganising.
Wadsworth et al., 2026 – A permeability model for pucks. Imaged loose packs of two coffees at 11 grind settings and simulated flow through them. Permeability followed connected porosity and specific surface; the authors point to dose, not tamping force, as the main control of puck thickness.
- Tamping force: a coffee science textbook states that the pressure drop through the bed depends on the degree of tamping1, and one study found permeability extremely sensitive to bed density.3 A 2017 espresso study found no detectable effect of tamping pressure on caffeine and trigonelline extraction4, and a 2026 permeability study, citing that 2017 study, treats tamping force within the range a barista can apply as a second-order effect compared with dose.5 In the density study, density was set by putting more coffee into a fixed volume, not by tamping harder.3
- Fines migration: if water carried fines down to the filter, the bottom of the bed would become less porous.9 The one imaging study opened for this entry found the opposite profile after extraction, with the bed more porous at the bottom than at the top; the authors attribute it to uneven erosion of the bed. Its beds were capsules, not tamped baskets.9
Coffearium (2026). Puck preparation. Coffearium – The open coffee compendium. Retrieved 22 October 2026, from https://coffearium.org/entries/puck-preparation/
@misc{coffearium2026-puck-preparation,
author = {{Coffearium}},
title = {{Puck preparation}},
year = {2026},
howpublished = {Coffearium -- The open coffee compendium},
url = {https://coffearium.org/entries/puck-preparation/},
urldate = {2026-10-22},
note = {Retrieved 22 October 2026}
}Entries are updated when the evidence changes; the retrieval date tells your readers which version you used. Changes are listed in the Change log.
Change log
- – Removed the claims on puck screens, paper filters and underfilled baskets, and their two practitioner sources. Reason: Each rested on a single Tier C source; no Tier A or B source was found.
- – Total porosity of the untamped imaged packs corrected from 0.50–0.78 to 0.48–0.69; ratio to the 2015 beds corrected from 30 to 35 at the low end. Reason: The upper value had been read from a neighbouring column of Table 1 in the 2026 study.
- – Porosity of 0.12–0.33 in the 2015 study described as calculated from dry-grain density and corrected for consolidation, not measured on wet beds. Reason: Wording of the source.
Sources
- BMestdagh, F.; Glabasnia, A.; Giuliano, P. (2017). The Brew – Extracting for Excellence. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 15, pp. 355–380. doi:10.1016/B978-0-12-803520-7.00015-3 · other copies
- Bvon Blittersdorff, M.; Klatt, C. (2017). The Grind – Particles and Particularities. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 13, pp. 311–328. doi:10.1016/B978-0-12-803520-7.00013-X · other copies
- ACorrochano, B. R.; Melrose, J. R.; Bentley, A. C.; Fryer, P. J.; Bakalis, S. (2015). A new methodology to estimate the steady-state permeability of roast and ground coffee in packed beds. Journal of Food Engineering 150, 106–116. doi:10.1016/j.jfoodeng.2014.11.006 · other copies
- AKuhn, M.; Lang, S.; Bezold, F.; Minceva, M.; Briesen, H. (2017). Time-resolved extraction of caffeine and trigonelline from finely-ground espresso coffee with varying particle sizes and tamping pressures. Journal of Food Engineering 206, 37–47. doi:10.1016/j.jfoodeng.2017.03.002 · other copies
- AWadsworth, F. B.; Vasseur, J.; Zhang, J.; Dobson, K. J.; Gagné, J.; Buckland, H. M.; Coumans, J. P.; Heap, M. J.; Theurel, A.; Kendrick, J. E.; Lavallée, Y.; Hendon, C. H. (2026). A model for the permeability of coffee pucks validated using X-ray computed micro-tomography. Royal Society Open Science 13(4), 252031. doi:10.1098/rsos.252031 · other copies
- AWaszkiewicz, R.; Myck, F.; Białas, Ł.; Puciata-Mroczyńska, M.; Dzikowski, M.; Szymczak, P.; Lisicki, M. (2026). Under pressure: Poroelastic regulation of flow in espresso brewing. Physics of Fluids 38(6), 063113. doi:10.1063/5.0319611 · other copies
- ACameron, M. I.; Morisco, D.; Hofstetter, D.; Uman, E.; Wilkinson, J.; Kennedy, Z. C.; Fontenot, S. A.; Lee, W. T.; Hendon, C. H.; Foster, J. M. (2020). Systematically Improving Espresso: Insights from Mathematical Modeling and Experiment. Matter 2(3), 631–648. doi:10.1016/j.matt.2019.12.019 · other copies
- ALee, W. T.; Smith, A.; Arshad, A. (2023). Uneven extraction in coffee brewing. Physics of Fluids (AIP Publishing). doi:10.1063/5.0138998 · other copies
- AMo, C.; Johnston, R.; Navarini, L.; Suggi Liverani, F.; Ellero, M. (2023). Exploring the link between coffee matrix microstructure and flow properties using combined X-ray microtomography and smoothed particle hydrodynamics simulations. Scientific Reports 13, 16374. doi:10.1038/s41598-023-42380-y · other copies
- ASmrke, S.; Eiermann, A.; Yeretzian, C. (2024). The role of fines in espresso extraction dynamics. Scientific Reports 14, 5612. doi:10.1038/s41598-024-55831-x · other copies
- AMéndez Harper, J.; McDonald, C. S.; Rheingold, E. J.; Wehn, L. C.; Bumbaugh, R. E.; Cope, E. J.; Lindberg, L. E.; Pham, J.; Kim, Y.-H.; Dufek, J.; Hendon, C. H. (2024). Moisture-controlled triboelectrification during coffee grinding. Matter 7(1). doi:10.1016/j.matt.2023.11.005 · other copies
- AFoster, J. M.; Lee, W. T.; Moroney, K. M.; Prjamkov, D.; Salamon, M.; Smith, A.; Petrassem-de-Sousa, J.; Vynnycky, M. (2025). Dynamics of liquid infiltration into an espresso bed using time-resolved micro-computed tomography: Insights from experiment and modeling. Physics of Fluids 37(1), 013383. doi:10.1063/5.0245167 · other copies
- A/BWeber, D. (2021). Tamper for espresso machine (US Patent 11,006,777 B1). United States Patent and Trademark Office; filed 1 July 2020, granted 18 May 2021.
- CRibes, S. (2020). Which puck preparation technique is best?. Decent Espresso (documentation).
- CAdam (Clive Coffee) (2026). What Causes Channeling in Espresso & How to Fix It. Clive Coffee (blog).
- CBarista Hustle (2021). Looking at Spent Pucks. Barista Hustle.
Tier A · authoritative · B · trade and expert · C · general. Supporting quotes for every source are in the data version of this entry. If a publisher’s site refuses access, try the doi link or “other copies”.
Discussion
A space for readers to share experience from the farm, the roastery or the bar, ask each other and offer other views. The Coffearium team does not reply here. Please keep to the community rules; to change the entry itself, use Suggest a change.