What Decent Espresso has learned from lever machines

Looking back in coffee history, I found that the most exciting time was when direct lever machines ruled. Your bicep is a far more sophisticated controller than any motor (or spring) could ever be. Put a pressure gauge stolen from a bicycle pump on the group, stick a thermometer in the right place, and voilà, you have made a formidable espresso experimentation device.

Like the birth of the guitar, where instrument makers wondered “what kind of music will be made with this instrument”, manual levers were an open-ended coffee-making instrument. The manufacturer did not specify the espresso-making recipe in the hardware of the machine: that was up to you. I believe this is because at that time in history, there were few people who believed in certainties. Much was still to be discovered. So, an open-ended machine was appropriate. Much experimentation rapidly, oh so rapidly, ensued.

The first lever machines used springs, and encoded a “known to work well” lever technique into the hardware. This made good espresso easier to attain, as you didn't need to learn how to use your bicep appropriately. For me, springs were the beginning of the end for levers. With direct levers, the manufacturer intentionally made an open-ended tool. A spring communicates a subtext that “the manufacturer knows best” and once you accept that thinking, further automation (and removal of flexibility) is likely to ensue. This “manufacturer knows best”, and a goal of removing the need for barista skill, is the trend that eventually gave us the great-step-backwards in drink quality: 9 bar espresso machines.

Lever machine users had found essentials and crucial truths about coffee:

• As coffee is extracted, puck resistance will decline due to material loss.

• Therefore, flow will increase naturally if pressure is kept constant, and too high a flow rate gradient doesn’t taste good. 

• So, to respect the nature of espresso is to decrease pressure as the espresso is extracted, to work with the reality of puck erosion, and consequently control the rate of flow rate change.

Here is a chart that shows this in practice. It’s from the Decent Espresso machine software and it’s an “espresso” made with a constant flow rate of 4 ml/s.

Graph showing a declining pressure profile (green) with a constant flow rate (blue)

The green line shows the pressure that results from this flow rate. You can see that puck resistance of this espresso gives a peak of 8.6 bar of pressure, which eventually declines to 3.7 bar at the end of the shot. A grinder that produces a lot of fines can cause a pressure “flick” up at the end, but nonetheless, most espresso pucks behave like this. As pressure and flow are directly linked, one can choose to control pressure (instead of flow) to create a desired flow curve. This is exactly what levers do. It’s no coincidence that spring levers create a pressure curve that is almost identical to the green line above: it’s because this pressure curve will create an almost-constant flow rate of espresso, with medium to dark roasted beans. This however, is also why I am opposed to spring levers. The ideal pressure curve, producing constant flow with a dark roast, is not the same pressure curve that is ideal for a light roast, or perhaps a coffee bean with different solubility, or those using very-low-fines grinders. Manual levers do not have limitation, which is why I always advise people who like light beans, but don’t want to buy a technological machine like the Decent, to buy a manual lever machine.

Cracking the secrets of light roasts
All the coffee experts I approached preferred light roasts, and most, in a candid moment, would admit to preferring pour-overs to espresso. “It’s a better extraction approach” they’d say. Scott Rao had success with light roasts with his “Blooming Espresso” recipe. He applied his understanding of why pour-overs were successful with these beans, and when thinking about pour overs, he noted:

• Pour overs have very even wetting, because the grounds are not compressed, so water can wet all the grounds at almost the same instant.

• Pour overs have a long bloom time, which gives light roasts more water contact time. As darker roasting increases coffee solubility, the opposite is true too, light roasts are less willing to give up their soluble coffee material to water.

• At the “draw down” stage at pour overs, most of the actual extraction has already happened, so you want the draw down to be both even and quick, to have an even extraction.

Applied to espresso making, this led to the following insights:

• What if the exact amount of water that the puck needed to be fully saturated was placed on it right at the start and then the water was quickly pushed into the puck, using a combination of pressure and capillary action? That should create the desired “even wetting”.

• What if a long pause happened after the wetting, letting the coffee “bloom” with long water contact, and thus respecting the poorly soluble nature of light roasts. 

These insights led to the Blooming Espresso recipe, which Rao programmed for the Decent. While typical refractometer readings of espresso are in the 19% to 21% range, Blooming runs typically from 23% to 28%. These very high percentages occur because of this combination of extraction approaches.

Blooming Espresso extraction graph

To replicate a Blooming Espresso on a direct lever all you need is a scale under your espresso cup, and follow these steps:

• Grind quite finely a coffee puck of 14 grams to 16 grams in weight.
  
  
• Raise the lever and drop a column of water on your coffee puck, as fast as possible.
  
  
• Hold a light amount of lever pressure on the puck with your lever, about 5 bar, for about 12 seconds.
  
  
• When drips start to appear on the puck, let pressure decrease naturally into the puck (tip: the ideal grind size for Blooming shots seems to be when you get between 4g and 10g of dripping into the cup, during your Blooming step. The grind is very fine. 8g of dripping seems to produce the best flavor).
  
  
• Wait 30 seconds for the puck to “bloom” and at the end raise the lever again to load some extra water.
  
  
• After the bloom, gently increase the pressure to around 8 to 9 bar, so that flow is around 2 ml/s to 2.5 ml/s.
  
  
• Stop after your in-cup weight is 3.5x to 4x. For ultra-light roasts it is even possible to reach 5x but due to the small size of most direct lever groupheads this can be very difficult.
  
  
• Aim for a total shot time of about 70 seconds.

Espresso’s present and future
The current most difficult challenge for the coffee community is with large-diameter, low-fine grinders, like the Monolith, Lagom P100 and EK43. It turns out that fines greatly help keep puck integrity high during espresso; however, fines also get over-extracted, cause muddiness, and fight the goal of achieving clarity of flavors in a coffee drink. Making espressos with light roasts, and low fines, is still a challenge.

Another current challenge is understanding the most appropriate temperatures, bloom times, and pressures for a given coffee bean. Perfecting a brew takes skill and time.

And finally, while machines like the Decent can instantly change the water temperature, the effect this has on different beans is poorly understood. Current best practice is split: whether to start + 2°C to + 4°C hotter at the start of the shot and decrease temperature, or whether to have constant water temperature during a shot. 

Many espresso machines from history (especially levers) have a decreasing temperature profile built into them. Other machines (especially high rotary pump machines) aim for a constant water temperature during espresso extraction. These different temperature curves during extraction likely cause the characteristic flavor profiles of these different machines, but exactly how this works is poorly known. Some say a hot start is advisable to extract less of the “bad flavors” as the extraction continues. Other think the main merit of this approach is bringing the room-temperature puck up to temperature faster. We’re very much still thinking through what we want to do with temperature during the making of an espresso.

This article was meant to explain my journey as a lever-fan, yet co-designer of a very modern machine. In many ways, I see the Decent as a lever machine, finally giving us back the freedom and flexibility we had with manual levers, but in a digital package. I hope the lever community will continue to see us as friends, working together to have fun, and drinking a few nice cups in our remaining years.