Chatting away with Jonathan at Repack espresso(conveniently located near my house) this morning was the beginning of this blog. I thought the conversation provided a neat narrative for considering the impact evenness and extraction can have on how we percieve brew ratios.
We chatted about the conceptual differences between an EK coffee burr espresso recipe and other grinders, in this case, his Anfim grinders.
What is your real dose and for that matter real extraction?
Let’s say that we both start with a 16g dose and dial a coffee in to achieve a desired extraction and strength.. Jonathan dials in a single origin coffee in his store (a roaster and coffee we also use in our store). I however dial this coffee in at our store and use an EK with coffee burrs. My recipe that achieves a flavour of both strength and extraction that I am happy with is predictably a longer shot – 16g in to 39 out. This tends to be common practice for an EK. Most of our shots are around this mark. Jonathan's is predictably shorter - 16g in to 34g out.
There’s a touch of dryness on Jonathan's shot but its really good. After chatting about our recipes he tries a longer shot but as is often reported it was thin and over extracted tasting.
It’s easy to think we are using the same dose in and exploring the same yields out. But we are not.
The difference is all about evenness. Conceptually its easy to recognise that a more even grind means we allow the water to access more of the coffee at once. This then means that the we can hit a higher extraction without the over extraction taints we have correlated with certain refractometer numbers. For example EK shots often taste under extracted at 20 percent, whereas as similar under extracted taste is achieved at a 17 percent reading on another grinder. This isn't exclusive to grinders though. The physics of even extraction that can be improved with lower flow rates/pressure (as explored in the previous two posts) present a similar phenomenon. Extractions that peak at a higher number when the brew is tasting its best compared to before.
In both cases we are extracting more evenly, and nowhere is evenness more impactful than when brewing espresso.
As these change, so does the use of reference points, such as brew ratio and extraction.
So, back to the title, In both cases – lower extractions that taste balanced and higher extractions that taste balanced- are the doses of coffee we are starting with actually the same? Well, as a dry weight on the scales, Yes. But, as far as the water is concerned, No.
If the grind is less even or we are extracting from less of it from some reason, then we are not actually using 16 grams, maybe we are only using 14g. I don’t know for sure, but for arguments sake, let’s assume it’s 14. From a grinding point of view , if we get a load of boulders, as we may from a Robur, then that “coffee material” is not fully accessible to the water. Part of the 16 grams is tied up in the bolder, essentially taking down the dose you are actually brewing with. The same goes for the physics of brewing, lets say the water passes through the coffee less evenly, either because of distribution, flow rate or evenness of grind. The water will be using less or more of the dose, effectively completely altering the brew ratio which we so often discuss.
This then impacts hugely on the refractometer reading, especially when converted to represent extraction.
When we type in the TDS% reading to the mojo app, we also type in our starting dose and shot weight. The calculation that takes place considers that we were extracting from the whole dose, in this case 16 grams. What if using a less even grind actually meant we only really used a dose of 14 grams? If we then type this as our starting dose into the mojo we would get a much higher extraction reading.
So in essence, are EK shots really that much of a weaker ratio than other espresso shots? For example the two recipes myself and Jonathan decided upon when using different equipment seem very different on paper, but in reality they may be much closer than they first appear.
