So, a massive roadblock for our experiment to test the distribution efficacy of distribution tools has been the analytical scales we need. They’re here now, and we’re pretty excited to get going ️:microscope:

Here’s a rundown of the experimental design:

We’re thinking we should try and mimic real-world situations, so we will be testing a 20 sample set from a Mazzer grinder. There are other grinders we can use for sure, but hey — time and resources (in saying that we’ll gladly accept all EK43’s and Mythos grinders sent to us — c/o Matt Perger, St Ali, South Melbourne, make it happen peeps). We’ll dial the grinder in for espresso, then follow the protocol Sasa Sestic has helpfully laid out here, as we test the OCD distribution tool first.

Just for clarity, we’ll clean the OCD with boiling water then dry thoroughly. We’ll adjust the depth by ensuring the propellors are clear of grinds after using the OCD, and check the puck itself has not been compressed too much, ensuring it’s nice and fluffy. Once set, we’ll get into testing.

We’ll first sample 20 doses using the OCD protocol and a Mazzer grinder before using the tool. This will be our control. We’ll then sample 20 doses using the Mazzer grinder, the OCD protocol, and then use the OCD to distribute the coffee.

The sampling process will involve our nifty 3D stainless steel laser-printed tool. This tool is a circle that fits in a 20g VST basket, snugly fitting up against the bottom filter of the inner basket. The circular device is divided into eight parts, with four segments in each part. The idea here is to insert the tool firmly into the puck, remove the tool with the coffee in each segment, then weigh each segment. This will give us a good idea of how the coffee was distributed horizontally across the basket.

Matt has done a really good job here of ensuring each individual inner segment has a volume equal to each individual outer segment. So the inner circle (four inner segments added together) is also equal to the outer circle ( … that’d be the four outer segments added together …). A good distribution tool should, in theory, have a relatively equal amount of coffee in each outer and inner segment, and therefore the inner circle should also be equal to the outer circle.

A VST basket is slightly sloped up outwards from the bottom filter to the top lip, and our tool is aligned from the inner bottom circumference of the basket straight up. So there’ll also be some wastage from the gap between the outer side of our tool, and the inner sides of the basket. We’ll also measure this, and this measurement will be important.

So here’s where the math gets a little tricky. Here’s what we’ll end up recording:

4 x inner segment weights, x 20

4 x outer segment weights, x 20

1 x inner circle weight (4 inner segment weights added together) x 20

1 x outer circle weight (4 outer segment weights added together) x 20

1 x wastage weight, x 20

These measurements will be taken for the 20 samples taken pre-OCD, and 20 samples post-OCD.

We’ll then have an average or mean for the 20 measurements taken for inner segments, the outer segments, the inner circle, the outer circle, and wastage.

We’ll subtract the mean wastage from 20 (our initial 20g dose used for each sample) and divide this number by 8 (corresponding to the 4 inner and 4 outer segments). This number will be our hypothetical “perfect distribution in each quarter” number. We’ll then subtract each outer and inner segment weight from this number (20 x for each condition) and find the mean difference.

Finally, we’ll compare this mean difference between pre-OCD and post-OCD, and see if there’s a statistically significant difference.

We have other avenues we want to pursue. Puqpress have kindly sent us a unit to use for testing, so we’ll also explore adding tamping to the design, then measure and weigh each segment. At some point we may head into blind taste testing or refractometry testing — but for now we want to stick with direct empirical measurement.

So to all our kickass funders for this project — thank you for your support! And we’d love to hear your thoughts in the comments below!