Apparent Overheating of Lever Espresso Machines Measured with Scace Thermofilter

#1: Post by **drgary** » January 26th, 2015, 2:34 am

While comparing lever espresso machines to an Ambient Espresso Vesuvius pressure profiling machine, Mike McGinnis of Compass Coffee attached a Scace Thermofilter to my Conti Prestina commercial lever espresso machine and found incredibly high readings. Here's a video of what he was measuring. Compare this to the offset PID setting suggesting brew temperature on the Prestina.

January 26th, 2015, 2:38 am

We checked the Prestina and the two other lever machines I brought, a current version Elektra Microcasa a Leva, and a 1987 Olympia Express Cremina. All showed apparent flash boiling at the group, which seemed to agree with the thermofilter readings. Yet all of these machines were pulling shots that weren't burned, and this was confirmed by four of us.

January 26th, 2015, 2:43 am

We scratched our heads, and I posted about this in the thread for our meeting. We also compared the thermofilter readings on Mike's Synesso Cyncra and his Ambient Espresso Vesuvius. These showed that the Scace thermofilter was properly calibrated and gave the expected readings on those machines. We ran each without a portafilter installed and saw no flash boiling at the group.

Eric Svendson came to the rescue with this response and an offer of measurement gear.

erics wrote:I have observed the same high temps when testing a Rancilio lever several years back. You need an adjustable check/relief valve on the thermofilter to duplicate the levers flow.

#4: Post by **samuellaw178** » January 26th, 2015, 2:49 am

Just for the record, here's one video I remember seeing on another lever.

January 26th, 2015, 2:57 am

In addition to the lack of a check valve, I think I see what was happening. Each of the lever machines has the group bolted to the boiler. Each has a dipper tube going to the bottom of the boiler, so they are not purging steam from the top. But opening the flow directly from the boiler without a portafilter attached or coffee in the portafilter releases water to the atmosphere that was liquid but is hotter than boiling at one atmosphere of pressure. This is as expected.

If you lock a portafilter filled with coffee into the group, the initial steam would release through the coffee cake and some of it would be cooled by the room temperature coffee particles. The coffee swells with water creating resistance so the flow of boiler water almost stops. Water that has been released into the group now heats the metal parts and the group acts as a heat sink, cooling the water to proper brew temperature. Coffee brewed this way tastes as it should.

The pump machines aren't introducing water directly from the boiler so there is no initial flash boil.

From what Eric wrote, the Scace Thermofilter set up without a calibrated check valve offered too much resistance and trapped steam, giving the reading you saw in the first video.

Does this interpretation seem correct to others?

BTW, just to make sure we were tasting properly Mike set the Vesuvius brew temperature to 212F/100C and that same coffee tasted burned to all of us. This morning I was pulling shots on my Prestina, set up the same way, and the shots tasted very good and could be fine tuned with the PID.

#6: Post by **another_jim** » January 26th, 2015, 3:49 am

The video of the Prestina appears to show the high temperatures throughout the simulated shot with the Scace attached, not just at the start when the check valve is open. The Scace's padding is supposedly thermally close or identical with a double puck shot. So I'm stumped. Maybe I'm missing Eric's point.

But one question about what is not on the video. Was there water flow through the Scace when the lever was down? That would explain the readings, since it would mess up the entire thermal management of a lever group. Lever groups are designed for the water to be static and cooling off during the 10 seconds or so of preinfusion.

http://users.rcn.com/erics/ · #7: Post by **erics** » January 26th, 2015, 7:18 am

So I'm stumped. Maybe I'm missing Eric's point.

The adjustable check/relief valve prevents any flow through the thermofilter until a particular pressure is reached. I also used a smaller orifice to more closely replicate the lever flow.

#8: Post by **Balthazar_B** » January 26th, 2015, 7:32 am

drgary wrote:
From what Eric wrote, the Scace Thermofilter set up without a calibrated check valve offered too much resistance and trapped steam, giving the reading you saw in the first video.

Does this interpretation seem correct to others?

It seems plausible in light of the expected dynamics of a shot pull in a lever machine. There's also the matter of the pressure profile that could affect the readings at the thermocouple, and cause inaccuracies because of the way a puck of coffee will respond and interplay with water as opposed to the static plastic puck in the Scace. It would be interesting to get Greg's point of view on this.

Looking at the first video, I'm amazed anyone was willing to even taste a shot from the Prestina after watching the thermometer!

#9: Post by **OldNuc** » January 26th, 2015, 7:40 am

I think you are looking at the same or similar thermodynamic process that occurred at TMI back in March of 78, look into the adiabatic process.

#10: Post by **rpavlis** » January 26th, 2015, 8:09 am

This is interesting! Remember though, that espresso made at too high a temperature tastes burnt not because we really "burn" it, but because the burnt tasting and bitter tasting components elute from the grounds because they bind less strongly to them at high temperature. Too cold temperatures do not allow some things we like to elute, and we have nasty espresso too.

We must further remember that we really are chromotographing coffee components when we make espresso. We must even further remember that the coffee has a substantial specific heat, about a third that of water per unit mass. As water passes through the puck it thermally equilibrates, but it does not really mix because flow is strictly in one direction like in any functioning chromatography system. The puck is at room temperature when we start, perhaps 22-25 degrees. When hot pressurised water hits the top of the puck it is hot, but as it moves down it cools off from the heat capacity of the coffee. The components that are bitter and burnt tasting tend to bind stronger to the coffee grounds. They do not simply extract and stay in solution but they repeatedly absorb and desorb from the insoluble matrix. As the compounds go downward they encounter lower temperatures, and the materials that bind strongly tend to move down much slower through the lower temperature regions. If we make a very long pull we begin to extract the nasty things even without overheating, and it tastes more and more bitter. When the bottom of the coffee becomes fully hot the nasty things elute fast too, and we end up with wretched tasting espresso.

If one measure the temperature of the espresso as it elutes the temperature keeps rising during the pull. To me espresso tastes best when the temperature of the last espresso to be eluted is between about 92 and 95 degrees. Others might prefer other temperatures.