Man vs Volumetrics

by Matt Perger for St Ali Coffee Roasters

Featuring: Jamie Thomson

Experiment Explained:

Jamie was entirely focused on the shots when he made them. This does not represent a true service situation and is a big leg-up for the Human party. No human can focus on 3 groups during service with any kind of accuracy.
We use VST 22g Precision Filters, Lab refractometers and Extractmojo software. The prism was cleaned with an alcohol wipe between every sample, and the device was calibrated with distilled water regularly. All pipettes and syringes were rinsed twice in the fluid to be measured before taking samples.

Espresso machine was never operating multiple groups during experiment. Every pair of samples was alternated between Jamie or the Volumetric dose being ground first.

The espresso bar was in service across a period of 3 hours. The grinder was adjusted during this time to keep our drinks to spec, as we would every day.

As you can see in the spreadsheet, Jamie chose time as the variable to control. From the accuracy of his brewing ratios, even scales on the drip tray would not have helped him very much. We’re trying to emulate a service environment, hence why we didn’t use scales on the drip tray.

TDS is also an important measure of consistency. This is made redundant when the dose is a fixed variable, so we rely on ext% for our results.

Analysis:

Volumetric no. 7 – The volumetrics chose to stop the shot at 35 seconds. The ext% and brewing ratio were still within parameters, and the shot tasted great.
(NB - Without this sample, the range for time would be 4 seconds, and standard deviation 1.21.)

Volumetric no. 1 – This is the shot in which the volumetrics performed worst (in my opinion). The TDS was lowest, time 2nd highest and yield highest. Taking Jamie no. 1 into consideration, the grind looks to be finer at this time than the rest of the sample set, as Jamie managed to reach an 18.6% extraction with an on-time 54.4% brew.

Jamie no. 6 – Jamie’s worst shot. He went by time, but the grinder foiled him with a spontaneously coarser grind size. In Volumetric no. 6 - the next shot - the volumetrics noticed an increased flow rate and cut the shot earlier, resulting in a correct extraction.

There is an error in all paddle-wheel based volumetric systems which actually works in the baristas favour. When there is lower pressure through the system, a paddle wheel measures all of the passing fluid. When there is higher pressure, the paddle wheel will not measure all of the passing fluid; some will “skip” the paddle wheel. This is because there are small gaps between the enclosure and the wheel (and other fluid dynamics stuff like compression) that allow some passing fluid past without registering.

eg. You’re making espressos with 21g dose and 41g yield in 27 seconds. You accidentally dose 0.5g less for one handle and engage the pump – there will be lower pressure at the paddlewheel due to less resistance in the basket. The volumetric will register more of the fluid, and stop the shot “earlier” than you would have, hence maintaining your brewing ratio. It works the other way round too.

I think it’s safe to say: if you’re not using volumetrics or weighing every single shot’s dose and yield, your espresso bar will be grossly inconsistent. There’s really no excuse, it’s 2012.

If you’re up for the challenge, send me something! - matt@stali.com.au

Loading more stuff…

Hmm…it looks like things are taking a while to load. Try again?

Loading videos…