This article is based on research by Stéphane Ribes, from a presentation in June 2019, which you can find here.

Pressure during espresso extraction is not determined only by the resistance of the coffee puck. Rather, the compression of the air trapped above the puck is a major factor governing the initial pressure curve. Three variables determine this compression: the initial headspace (1), the water flow supplied by the pump (2), and the water flow entering the coffee puck (3).

Considering a perfect gas model, the following formula can be established:

Blind-basket experiments with a DE1PRO showed acceptable agreement with this model.

The following questions can be inferred from Ribes' presentation:

Tests were carried out with a DE1PRO v1.1 Decent Espresso machine. Blind-basket experiments isolated the effects of headspace and pump flow. Headspace was varied by adding water to the blind basket, producing estimated air volumes of 36, 26, 16 and 10 cm³.

Flow-rate experiments used initial rates of 1.0, 1.5, 2.0 and 4.0 mL/s. In these tests, water flow was switched to 0 mL/s when pressure reached 7 bar.

Coffee-puck experiments used 12 g in a 15 g VST basket, an IMS SI 200 IM shower screen, and EK43 grind settings from 0.9 to 1.2. Further practical comparisons investigated different basket sizes and IMS CI 200 versus SI 200 shower screens.

First, the blind basket was used to determine how changing the trapped air volume alters pressure build-up independently of coffee. Second, the pump flow was varied while headspace remained comparable. Third, coffee was introduced to separate the effect of water entering the puck from the effect of puck resistance itself. Finally, basket size and shower-screen configurations were changed to examine whether these principles also explain differences observed in realistic espresso preparations.

The graph below shows a clear positive relationship between headspace and the time required to reach 7 bar at 1 mL/s.

Approximately 10 cm³ required 15 s, 16 cm³ about 21 s, 26 cm³ about 28 s, and 36 cm³ about 32 s. Thus, a smaller trapped air volume is compressed more rapidly and causes pressure to rise sooner and more steeply.

In other words, as headspace decreases, the time required to reach 7 bar at 1 mL/s decreases significantly.

The inverse relationship with flow rate is even stronger. The time to 7 bar fell from roughly 32 s at 1 mL/s to 23 s at 1.5 mL/s, 17 s at 2 mL/s, and only 9 s at 4 mL/s. Higher flow also produced progressively greater pressure overshoot after the 7-bar shut-off trigger, showing that the hydraulic system continues to respond after pump flow is stopped.

In other words, higher input water flow rates significantly reduce the time needed to reach 7 bar. Also, more pressure overshoot occurs as the input water flow rate increases.

Changing the EK43 setting from 0.9 to 1.2 produced a large decrease in maximum extraction pressure, from approximately 12 bar to 3–4 bar. In contrast, the additional time needed to reach 3 bar compared with the blind basket stayed close to 4.5–5.0 s. An excessively fine grind produced a pressure curve resembling the corresponding blind-basket curve, shifted later in time. Most of this initial delay was due to water entering the coffee puck rather than to grind-dependent hydraulic resistance.

This supports the principal observation that, during relatively slow preinfusion, the pressure-rise delay caused by puck saturation is only weakly affected by grind size.

With the same 18 g dose and EK1.8 grind, changing from a VST 20 g basket to the taller VST 22 g basket increased headspace. The measured additional pressure-rise delay increased from about 5.0 to 7.5 s, while maximum extraction pressure decreased. Despite the different pressure curves, both examples reported an extraction yield of 19.9%.

With a 15 g dose in the same basket, changing from the IMS CI 200 to SI 200 screen had little effect on headspace and therefore little effect on the pressure delay: approximately 6.0 versus 5.5 s. Extraction yields were 20.4% and 21.3%, respectively. The SI 200 produced a gentler pressure decline; the proposed explanation is its smaller distance to the puck surface, although further testing is needed to confirm this hypothesis.

In a second comparison, changing simultaneously from CI 200 + VST 15 g to SI 200 + VST 20 g, with an 18 g dose and the same EK1.8 setting, increased the time to 2 bar from 14 to 16 s. The lower maximum pressure was attributed to the combined effect of increased headspace and the larger holes of the VST 20 g basket.

The experiments indicate that the early espresso pressure curve can largely be interpreted as the compression of a trapped air volume by the net accumulation of water above the puck. Headspace and pump flow determine how rapidly this compression occurs, while water entering the coffee delays it. Basket geometry therefore affects pressure indirectly through its effect on headspace.

1. How does initial headspace influence pressure build-up?

2. How does pump water-flow rate influence the pressure curve?

3. How do puck penetration and grind size influence the initial delay?

4. How do basket size and shower-screen configuration affect the pressure profile?

#research #Ribes #coffee #experiments