The reaction to my last post indicates the sheer amount of interest in grinding concepts. It has also presented me with some completely new and surprising information. Out of all of the concepts and speculation put forward within the article, it was actually a seemingly well-established and widely accepted idea that sparked debate.
Let me summarise:
As we get busy the grinder gets hot and we find that we have to adjust the grinder to a finer setting to achieve a similar extraction to what we initially dialled in, earlier that day. Without going finer, the flow rate of the shot speeds up, with the recipe of dose to shot weight kept the same. Every barista around the world will have worked with this day to day, frequently being irritated by the circles they find themselves running round in.
There are two main theories that are circulated to explain this phenomenon. That is i) the heat of the grinder expanded the metal composition of the grinder and that the burrs move apart or ii) possible change in efficiency of the grinders motor as it beds in for the day.
The world of twitter comes into its own here, allowing for quick, albeit short and sometimes confusing conversational exchanges. Some people questioned, quite aggressively, this notion regarding heat and grind changing. In fact, it was implied to not be true at all: This is very intriguing, because if this is the case, then we want to know why!
It appeared that this knowledge is coming from research carried out by grinding manufacturers; proprietary studies where the results are often not public. But, if the studies are true; that is, the grinder is not changing, nor is the particle sizes coming out of the grinder, the ramifications are massive.
If the grind doesn’t change then some other change is having a huge impact.
The popular counter argument is that hotter grind results in hotter puck temperature. That the grind stays the same and heat energy changes flow dramatically. For instance, yesterday in store we had a medium busy day and we had to alter the grind dial on the EK43 approximately one whole number by midday to achieve the same flow rate that was achieved in the morning. If the grind is not changing, we have then made our grind a hell of a lot finer. This means that the mode of all the particles in the puck is smaller: we will have dramatically changed the surface area within the dose, the contact time will be similar as we have kept the flow the same by going finer. This means our extraction should be considerably higher in the afternoon compared to the morning.
This would then mean that every shop carefully adjusting their grind to achieve a consistent recipe all day long have got it all wrong, myself included. Being proven wrong is a good thing, if it helps us have a better understanding of what we are doing.
It would actually mean that we need to let the flow rate speed up during the day in acknowledgment of the grind size/surface area being the same. It raises some other questions though. Is the increased heat and its potential impact on extraction counter acting the decreased contact time from a speedier flow rate? Or would we need to slow it up, but just not as much as we do.
In all honesty I haven’t consistently tested the extractions at the lowest and peak busy times in store. We have tested it at various times and not noticed a dramatic correlation, but we haven’t steadily tested it with a good data set. I have a coffee post lunch every day though, a coffee I drank in the morning and at the same recipe, it is often a little different but it tastes good. I may not be picking up on something though, which is always a sensory concern. Would the extraction really stay the same if I kept the weights of the dose and shot the same but let the flow rate go as it pleases? One comment on twitter suggested that this new understanding led the individual to actually go coarser in the busy period to counteract the posited increased extraction from a hotter puck….
Wait a minute… what are we talking about? I am now confused. The grinder is not physically changing aperture with temperature, yet I am grinding a lot finer to achieve the same extraction when the grinder is hot. And you are telling me that I the particles coming out of the grinder are not changing in size or distribution as the temperature increases, yet the flow rate suggests otherwise?
Something is fishy here.
The problem with this discourse however was the lack of source material/data. We needed to see the data that the grinding companies have obtained. Grind distribution graphs can look hugely different depending on what axis have been used (i.e. log scales, volume %, total count, etc) and because a lot is going on at once the interpretations can be troublesome. As mentioned in the last post, changing the axis from volume % (where fines are minimal contributors because they are very small) to surface area % (where fines have a huge surface are to volume ratio and thus contribute massively) completely changes what we see as significant. The fines bump which looks so small and uninteresting in the volume % presentation, becomes the most important part of the graph. They contribute the most to surface area and extraction. Not by a little bit. By A LOT. They are 70% of the grind, minimum.
Myself and Chris were kindly sent the graphs which were at the source of the theory that grind distribution does not change with temperature. They are fascinating, and although I want to present that data here, I will not yet, because we want to write this story concisely and scientifically.
However, I will explain our interpretation of this graph whilst referencing the ideas outlined above.
The first amazing thing that they indicate is that yes: The burrs don’t move apart during the day/heating. BUT they also do not show that the grind stays the same.
Part of the grind stays identical, but part of it changes.
The larger particles and therefore the mode stay the very similar. This is expected, because if the aperture of the burrs remains constant, the large particles coming through should also. So what’s changing? Again, it appears to be all about the fines!
The general shape of the particle distribution is pseudo-Gaussian. It looks like an ant hill. However, the hot grind and the cold grind have slightly different shapes. The tails (the fines) of the cold grind are about twice the volume % of the fines tails of the hot grind.
It means the cool dose produces way more fines and therefore more surface area (slower shot) and the hotter dose appears to produce far less fines and therefore less surface area (faster shot) even though the mode is similar (statistically the same). So in essence the original idea about the grind changing is right, just not for the reason we thought. The burrs don’t move, but the coffee breaks up differently.
Why is this happening?
Chris describes it like spaghetti, a hard cold piece of spaghetti shatters into more uneven pieces through being brittle. When heated, the hotter piece of spaghetti (without added water) is more malleable and less likely to splinter into tiny bits. This actually holds true for just about every organic material. This is well founded and published in Physical Review Letters (the premier physics journal); http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.95.095505
It looks as though hot ground coffee is less likely to produce fines. Amazing!
This means when we change our grind to finer we are making more fines to make up for the fact the hot burr ground coffee produces less fines. This will of course change the whole grind profile as we have to bring down our mode size to achieve this. That would infer that a grinder running hotter all the time would be more consistent.
It doesn’t mean we can conclusively say heat of the puck isn’t also affecting flow rate and extraction, but we can say that the grind profile does change at different heats. Chris would speculate this affect is not a major contributor to the extraction.
There we have it, when I wrote that blog and posted it yesterday I didn’t conceive nearly all of which is written above. Thrilling really. May the search for answers continue.