By Paul Golding
Head Roaster, Gabriel Coffee
Can’t get your grind setting right? You think you’ve dialled it in, but a few minutes later it’s wrong again? Clumping coffee driving you mad? These are common problems in many cafes, and in the last couple of months we’ve been looking hard for some answers that make sense. Our suspicion was that rapidly changing moisture levels in the coffee could be the cause, so we arranged some experiments to test the theory.
Water Activity and dose
Moisture level is just as important in coffee as it is in most dry food products. This goes for green, un-roasted coffee and for the roasted product. Many roasters and green coffee suppliers use agro-industry moisture meters to test for total water levels, which is good, but taking it a step further involves testing for the dynamic movement of free water available for evaporation, or use by bacteria and mould. This sort of dynamic testing is known as water activity, and requires some special equipment, but is worth the extra trouble for the insight it provides to the movement of water into and out of the coffee beans.
Water Activity, or Aw as we’ll abbreviate it from now on, is a critical quality control factor which food scientists watch like hawks. Too much moisture in a product supposed to be dry can cause premature spoiling and staling. With roasted coffee Aw is a prominent factor involved with early staling and, as we’ll see here, inconsistencies in grinding.
The water activity meter we use at Gabriel is technically called an Equilibrium Hygrometer. Put simply, it places a sample of coffee in an enclosed environment, then measures the vapour pressure of water evaporating from the sample. From this it calculates the “free water” available for evaporation and presents it as a decimal value on a specific water activity scale. The scale runs from zero, which would be completely dry air, to 1.0, which is the water activity of pure distilled water at a given temperature.
So, how does this factor influence grinding performance? It makes sense that if coffee is very dry, it will be more brittle, and will grind more quickly. As coffee gains moisture, one would expect it to become denser, and therefore grind more slowly. Our testing in the lab reveals this is indeed the case, and a moderate amount of change in Aw will result in a significant change in the performance inside the grinder. This effect is mainly relevant to Electronic type grinders like the Robur E or Mythos One, which use an electronic timer to control grinding time and therefore dose.
The testing revealed that our coffee, when packed fresh out of the roaster has an Aw of around 0.18 to 0.26, depending on the weather. This range is recommended for the packaging of roasted coffee to avoid premature staling. While the coffee is in the bag, it degasses, exhaling carbon dioxide via the one-way valve. Coffee loses a fair amount of moisture during the degassing process, with a freshly opened bag at 6 days after roasting reading an Aw of around 0.12.
Once the bag was opened, we tested the Aw at 5 minute intervals, at the same time measuring the dose dispensed by a Mythos One grinder without any changes to the settings. We took three test doses from the Mythos each time, and reported the Average of the three, to smooth out any general variations from dose to dose.
Aw increased over the testing period as the coffee absorbed moisture from the air. As the Aw readings increased, so the volume of coffee dispensed by the Mythos decreased, by about 1 gram over 45 minutes, as you can see in the graphs below, where the horizontal axis is time:
A 1 gram reduction in dose is a pretty significant drop, and if you’re not paying attention, one that will quite seriously affect the flavour of your coffee.
So, we have established that after opening a bag of coffee, water activity will increase steadily, and with it, the time taken to grind a set dose of coffee.
The next stage involved checking the effect of increasing water activity on extraction times.
Water Activity and extraction
For this section we shifted to the Mahlkonig EK 43 grinder, as we could pre weigh the dose and pour it straight into the machine. The objective was to extract a 22gm dose every 5 minutes, while keeping grind setting unchanged, to determine if the extraction time would shift in correlation with Aw.
The graph below illustrates the results. We opened a bag of coffee dated 6 days after roasting, and immediately took samples for Aw reading and extraction. Every 5 minutes after, we repeated the process, obtaining a series of extraction times and Aw values over a 45 minute period. The machine used was a La Marzocco PB using a volumetric button set to 300 pulses.
The graph shows the extraction times for a dose with these parameters:
22 grams in, 44 grams out, over x seconds, where x varies according to changing Aw.
We can see the initial shot time was 18 seconds. Without making any changes to the grind setting on the EK, over a 45 minute period the extraction time increased by 22 seconds to a final time of 40 seconds for the shot. At the same time the water activity meter recorded an increase in Aw from .087 to .288.
The progression is not perfectly linear, which we attribute to inconsistencies in the coffee beans absorption of moisture, small error from the water activity meter, and slight variations in the LM’s delivery of extraction water and pressure. The trend line is quite clear, however, and backs the theory that coffee with a higher in-built moisture level will provide greater resistance to extraction water and promote a longer shot time. One aspect we found interesting was the way this test fell into two halves- the first 20 minutes, where Aw increased but shot time did not, and the latter 20 minutes where Aw continued to rise, and shot time increased dramatically.
Our interpretation here was that the coffee immediately started to draw in moisture from the air, but this was limited initially to the surface layer of each bean, leaving the bulk of the coffee inside the bean still comparatively dry. While the increasing moisture registered on the Aw meter, there had not been enough penetration of moisture into the coffee to significantly affect the shot. After 20 minutes, enough time had elapsed for the moisture to penetrate closer to the core of each bean, meaning more and more of the ground coffee would have an increased moisture level. (The Aw readings are made with whole beans in the sample cup).
In both series of tests, subsequent checks after the 45 minute mark showed only very small change. 45 minutes was roughly the amount of time required for the coffee to reach equilibrium moisture content with the air, after which there were only tiny changes to the grind performance or shot time.
The testing showed us clearly that from the time of opening a bag of coffee, Aw will increase steadily for around 45-50 minutes, until equilibrium with the air is reached. The higher the ambient humidity, the higher the coffee’s Aw will go.
It is also clear that there is a relationship between the water activity in the beans, and the performance in grinding and extraction.
- Grinding will take longer, meaning an electronic grinder dispensing dose based on operating time, will progressively throw a smaller dose as Aw increases.
- Density of the grounds increases via moisture content, providing more resistance to espresso extraction, and longer shot times as Aw increases.
Now let’s take the findings into a café situation. You might be working all dialled in, halfway through your current bag of coffee, dosing 22gm in, 44 out in 34 seconds, for example. When the bean hopper gets low, you open a new bag and pour it in. Within 10 minutes, you are grinding the new coffee with a very low Aw, and suddenly, your dose seems to have increased, and at the same time your shots are gushing a bit. You respond by decreasing your grind time and perhaps make it a bit finer. Under normal circumstances this would be the right response, but in this case you’ll be a victim of dynamic Aw change. As the new coffee draws in moisture, it will act as we’ve seen above, and your new grind settings will soon be producing under-dosed shots which still over-extract and leave a muddy mess in the group basket.
The solution is to follow a practice of opening the next bag of coffee earlier, and getting it into the grind hopper while it’s still at least half full. The procedure in many cafes we’ve seen, of opening the bag and placing it, inverted, into the half full hopper to drain out as service progresses, is revealed as a pretty clever practice in terms of combating this issue. The coffee has more time to equilibrate with the air before it hits the grinder blades, meaning less adjustment required.
If you are encountering problems with inconsistent grind performance in your workplace, it’s worth considering how dynamic change in moisture levels may be affecting the coffee. Coffee will always perform best when it’s stored in a cool, dry place, allowed some time to equilibrate after opening, and when the grinder is placed in a location with the most stable environment.
We hope this has been an interesting read, and some useful information. We’ll be posting further articles on Water Activity in the future. We also have articles planned on cupping and roasting, plus a series of brewing videos. Keep an eye on the Coffee Research section of the website for updates, and our @gabrielcoffee instagram feed.
All the best,