The tamping pressure during tamping has a minimal effect on the extraction of espresso. We wanted to provide scientific evidence for this insight gained from daily practice. To this end, with the support of a food technologist, we set to work and extracted 150 espressos, took measurements, and varied the tamping pressure. In this article, we present our findings to you.
What is tamping?
Tamping is the act of compressing ground coffee in the portafilter. To “tamp,” a tamper is used, which consists of a handle and a metal base that is pressed into the portafilter like a piston. The metal base matches the inner diameter of the basket. If the gap between the basket wall and the tamper diameter is too large, a portion of the coffee powder will not be sufficiently compressed.
Compressing the grounds allows for even water distribution. The brewing water should encounter the same resistance across the entire surface of the coffee grounds at all contact points. If the coffee grounds are not compressed enough in certain areas or if there are air pockets beneath the tamped surface, there is a risk of channeling. This can be prevented through careful preparation of the grounds, vertical compression, and even tamping.
Guide: Step-by-step espresso preparation (articles and videos).
Three classic tampers with a 58mm diameter.
Research question: What is the influence of tamping pressure?
We investigated two sub-aspects. Firstly, we wanted to find out whether a change in tamping pressure influences the extraction time. Secondly, we examined whether the tamping pressure has an effect on the extraction yield and, consequently, on the concentration of dissolved coffee particles in the cup.
To obtain the most accurate results possible, we precisely defined the work steps and methodology.
Materials
Rancilio Specialty, professional espresso machine with a precise flow meter. However, every espresso shot was cross-checked with the Acaia Luna.
Etzinger etzMax, espresso grinder with an integrated scale. All grinds were double-checked, and if there was a discrepancy, the process was repeated (see methodology).
Puqpress, automatic tamping machine for professional gastronomic use. Tamping pressure is adjustable.
VST refractometer, used to measure TDS as a percentage, i.e., the concentration of dissolved particles in the drink relative to the water.
VST syringe filter, to filter solids from the coffee solution so that only the dissolved particles are measured with the refractometer.
VST syringe
Etzmax espresso grinder and Puqpress
Methods
The espresso grinder is set to an output of 18.5g. We use grind results between 18.4g and 18.6g.
The Rancilio Specialty espresso machine is programmed to an output water volume of 45ml and a brewing temperature of 93 °C.
The corresponding amount of coffee is ground directly into the portafilter. Afterward, the coffee is leveled with the fingers and tamped with the Puqpress. The portafilter is locked into the machine and the extraction is started.
The amount in the espresso cup is measured using the scale. After extraction, the portafilter is removed and knocked out.
A flush shot is run on the espresso machine.
After extraction, the espressos are set aside to cool. Once cooled, they are stirred and the TDS value is measured with the refractometer.
To do this, the espresso is dispensed into the refractometer using the syringe via the syringe filter.
This process is repeated 50 times each for tamping pressures of 10kg, 15kg, and 20kg.
The data analysis for throughput time and TDS values is performed using a one-way ANOVA with a set error probability of 5%.
Results: Throughput time and tamping pressure
The variance analysis for throughput time showed no significant difference. In other words, whether we tamped with 10kg, 15kg, or 20kg, our measurements did not show any noticeable variation based on the tamping pressure.
Impact of tamping pressure on extraction time.
| 10 kg pressure | 15 kg pressure | 20 kg pressure |
| 26 | 20 | 27 |
| 21 | 23 | 25 |
| 23 | 26 | 29 |
| 24 | 27 | 28 |
| 23 | 25 | 24 |
| 24 | 26 | 27 |
| 25 | 26 | 23 |
| 24 | 29 | 25 |
| 26 | 20 | 26 |
| 31 | 24 | 27 |
| 25 | 23 | 25 |
| 24 | 23 | 21 |
| 25 | 22 | 26 |
| 27 | 23 | 21 |
| 24 | 30 | 24 |
| 24 | 24 | 21 |
| 25 | 22 | 24 |
| 27 | 21 | 25 |
| 23 | 21 | 28 |
| 31 | 24 | 22 |
| 25 | 24 | 22 |
| 24 | 25 | 24 |
| 22 | 30 | 23 |
| 27 | 20 | 25 |
| 22 | 21 | 22 |
| 25 | 23 | 25 |
| 25 | 21 | 21 |
| 24 | 21 | 22 |
| 24 | 27 | 21 |
| 23 | 22 | 24 |
| 27 | 23 | 22 |
| 21 | 22 | 30 |
| 24 | 24 | 21 |
| 25 | 21 | 25 |
| 24 | 25 | 29 |
| 25 | 24 | 26 |
| 23 | 24 | 23 |
| 23 | 23 | 24 |
| 23 | 23 | 24 |
| 22 | 23 | 23 |
| 21 | 22 | 24 |
| 26 | 24 | 26 |
| 23 | 25 | 24 |
| 25 | 22 | 20 |
| 24 | 23 | 27 |
| 20 | 23 | 22 |
| 24 | 26 | 25 |
| 24 | 22 | 26 |
| 22 | 21 | 22 |
| 23 | 24 | 20 |
| Groups | Count | Sum | Average | Variance |
| Column 1 | 50 | 1212 | 24.24 | 4.51265306 |
| Column 2 | 50 | 1177 | 23.54 | 5.60040816 |
| Column 3 | 50 | 1210 | 24.2 | 6.08163265 |
| Source of Variation | SS | df | MS | F | P-value | F crit |
| Between Groups | 15.4533333 | 2 | 7.72666667 | 1.43133301 | 0.242301 | 3.05762065 |
| Within Groups | 793.54 | 147 | 5.39823129 | |||
| Total | 808.993333 | 149 |
| Fk(0.05) | 3.05762065 |
The test statistic F is smaller than the critical F-value, meaning there is no significant result. The actual p-value exceeds the set error probability of 5%. Therefore, no significant difference can be determined between the groups.
Results: TDS values and tamping pressure
The variance analysis for TDS values also shows no significant difference. The concentration of dissolved coffee particles was not significantly affected by the varying tamping pressure.
Impact of tamping pressure on the concentration of dissolved particles.
| 10 kg pressure | 15 kg pressure | 20 kg pressure |
| 8.53 | 7.77 | 8.15 |
| 7.99 | 7.94 | 8.15 |
| 8.03 | 8.02 | 8.16 |
| 8.13 | 8.03 | 8.27 |
| 8.01 | 8.08 | 8.2 |
| 8.01 | 8.06 | 8.08 |
| 8.1 | 8.07 | 8.08 |
| 8.05 | 8.26 | 8.01 |
| 8.2 | 8.12 | 7.96 |
| 8.24 | 7.96 | 8.03 |
| 8.17 | 7.98 | 7.99 |
| 8.15 | 7.99 | 8.03 |
| 7.91 | 7.85 | 8.11 |
| 8.09 | 8.02 | 8.02 |
| 8.09 | 8.88 | 8.02 |
| 8.09 | 8.12 | 8.01 |
| 8.11 | 8.11 | 7.93 |
| 8.15 | 7.99 | 8.04 |
| 7.96 | 7.99 | 8.04 |
| 8.3 | 8.05 | 7.99 |
| 8.25 | 8.15 | 7.99 |
| 7.99 | 8.2 | 8 |
| 8.07 | 8.28 | 7.99 |
| 7.99 | 7.81 | 8 |
| 7.81 | 7.84 | 7.97 |
| 7.8 | 7.92 | 8.01 |
| 7.93 | 7.9 | 8.01 |
| 7.94 | 7.87 | 7.66 |
| 7.86 | 7.97 | 7.8 |
| 7.93 | 7.88 | 7.89 |
| 8.06 | 8.02 | 7.99 |
| 8.08 | 8.02 | 8.02 |
| 7.99 | 7.95 | 7.92 |
| 8.04 | 7.7 | 7.99 |
| 8.05 | 7.89 | * |
| 8.13 | 8.01 | * |
| 8.2 | 8.11 | * |
| 8.19 | 7.95 | 7.91 |
| 7.98 | 8.02 | 7.96 |
| 7.94 | 8.07 | 7.7 |
| 7.82 | 7.94 | 8.01 |
| 8.08 | 8.04 | 8.08 |
| 8.08 | 8.05 | 7.92 |
| 8.07 | 7.96 | 7.83 |
| 8.06 | 7.97 | 8.03 |
| 7.88 | 7.78 | 7.97 |
| 7.88 | 8.01 | 7.97 |
| 7.93 | 7.81 | 7.91 |
| 7.97 | 7.73 | 7.65 |
| 7.96 | 7.85 | 7.69 |
TDS values. *missing measurements in these fields
| Groups | Count | Sum | Average | Variance |
| Column 1 | 50 | 402.27 | 8.0454 | 0.01824167 |
| Column 2 | 50 | 399.99 | 7.9998 | 0.03163465 |
| Column 3 | 47 | 375.14 | 7.98170213 | 0.01635356 |
| Source of Variation | SS | df | MS | F | P-value | F crit |
| Between Groups | 0.10546284 | 2 | 0.05273142 | 2.37573217 | 0.0965805 | 3.058928 |
| Within Groups | 3.19620383 | 144 | 0.02219586 | |||
| Total | 3.30166667 | 146 |
| Fk 0.05 | 3.058928 |
| Fk 0.1 | 2.33979948 |
The test statistic F is smaller than the critical F-value, meaning there is no significant result. The actual p-value exceeds the set error probability of 5%. Therefore, no significant difference can be determined between the groups.
Raw data and measurements of the tamping test
The raw data and further information can be found in the Excel sheet. If you uncover any other interesting insights, we look forward to the results.
Download tamping project results
When analyzing the data, a linear correlation between the extraction time and the determined TDS values was found for the following datasets, determined using the Bravais-Pearson correlation coefficient:
- Tamping pressure 10 kg: extraction time and TDS: weak positive correlation -> r=0.5068
- Tamping pressure 15 kg: extraction time and TDS: strong positive correlation -> r= 0.6854
- Tamping pressure 20 kg: extraction time and TDS: weak positive correlation -> r= 0.5423
where: |r|= 1 -> perfectly linear correlation
|r|= 0 -> no linear correlation
Tamping pressure, taste, and additional information
Within the scope of this test, the taste of the espresso was not examined. Therefore, no statements can be made regarding it. However, based on the findings of the test, we interpret that varying tamping pressure in the range we studied has no or only a marginal influence on taste.
Also, due to our test, we cannot make any statements about how pressure lower than 10 kilograms affects extraction. Based on research, our partners, and our own experience, we recommend tamping with at least 6 kilograms of pressure.
Discussion
When investigating the question of whether tamping pressure influences the extraction time of espressos, no influence could be proven. This does not mean that no influence exists, only that no influence was proven within the test model used. Interference factors include, for example, inaccuracies in preparation and the insufficient precision of the testing machines used.
The equipment used is professional-grade equipment designed for gastronomic use and not precise laboratory equipment.
However, in practice, there are far greater interference factors in preparing an espresso than those in our test procedure.
Thus, it cannot be assumed that tamping pressure has a large impact on extraction in practice, as there is a multitude of inconsistent parameters affecting the extraction process. In this context, tamping pressure is just one parameter among many.
---
The tamping test was conducted by Michel Aeschbacher in collaboration with a food technologist who wishes to remain anonymous. You have our gratitude!
We are open to further research partnerships and are happy to work on insights related to the topic of coffee using our materials and expertise.
