It’s an exciting time to be data driven in coffee. A little more than a decade ago a coffee refractometer would set you back 700 USD and here we are where devices are entering the market at a quarter of that price. Price had always been a barrier to entry in coffee refractometry but now that more and more enthusiasts and professionals will be getting their hands on one, it’s also important to ensure the data they get from it has as little error as possible. The workflow I describe below mainly builds off Jonathan Gagné’s recommendations and findings, and is adapted for my Atago (although I’m hoping owners of other devices will also benefit). I will first describe the equipment and workflow and then some thoughts on why I find refractometry beneficial.

Equipment

• Two tablespoons. I personally prefer cupping spoons and am a sucker for the Little Dippers (not to be confused with Mini Dippers) from Umeshiso. Tablespoons have a large thermal mass that help bring the sample to room temperature. I learned the tablespoon technique from Jonathan on his discord.
• Lint free microfiber cloths. Use at least two – one for wiping coffee and the other for wiping alcohol and water. It’s important to separate the two to prevent contamination.
• A demitasse spoon or teaspoon in case of espresso (just for convenience)
• Isopropyl alcohol to clean the refractometer lens (I will advise against alcohol wipes since they will scratch the lens over time)
• Distilled water to clean and zero-calibrate refractometer lens
• Two dropper bottles, one each for alcohol and distilled water
• A coffee refractometer that can measure TDS in the range you brew coffee and has sample temperature information
• Patience

Workflow

The most critical thing to ensure in refractometry is having 0.1°C between zero calibration temperature and sample temperature. Jonathan Gagné explains it best halfway down his post on precision refractometry. Not having such a small delta will be the main reason behind most of your errors in refractometry. I know this sounds too painful a thing to ensure, but I’m hoping the following steps will make it easier.

• Measure the weight of the empty carafe or vessel in which you collect brewed coffee. You can later subtract this weight from combined weight of vessel and beverage to get beverage weight.
• Store your refractometer, alcohol dropper bottle, distilled water dropper bottle in the same room. Having all of these at the same temperature will be key, especially when it comes to the rim of your refractometer.
• Assuming a dirty lens, put a few drops of alcohol on the lens and wipe with lint free cloth. Make sure you’re cleaning the walls of the rim thoroughly.
• Put a few drops of distilled water on the lens and similar to the previous step clean thoroughly with lint free cloth. DO NOT do zero calibration with the distilled water you just put in because remnant alcohol can skew the reading.
• Brew your coffee. Stir the brewed coffee vigorously to prevent layering of solubles. I prefer a back-and-forth motion to not create a whirlpool. I recommend against swirling for the same reason. At this point, measure your combined vessel and beverage weight (and later subtract vessel weight you measure in the beginning to get beverage weight).
• For filter coffee scoop a small sample (no more than 2 ml) onto a tablespoon. For espresso this might be easier to do with a demistasse or teaspoon.
• Pour only 1 ml from the spoon you scooped with onto another tablespoon. This is your main sample (refer to photo at top of page for example sample size).
• On the refractometer lens you cleaned earlier, pour a few drops of distilled water and hit the button for measurement (on an Atago this is the “Start” button)
• Press the “Zero” button (or equivalent) to calibrate to zero against the distilled water you just poured. Note the temperature it zeroed against.
• Clean thoroughly with a lint free cloth. Spending upto three minutes between when you scooped out the sample and when when you finish zero calibration is okay. I prefer keeping the difference in duration at about a minute. In my experience that usually results in less deviation from zero calibration temperature. The small sample size poured onto the tablespoon and the large thermal mass of tablespoon will ensure that the sample gets cooled to room temperature almost immediately and prevents further evaporation.
• Gently swirl the tablespoon brew sample a few times and pour just a few drops (just enough to cover the refractometer lens; refer to photo at top of page).
• Hit “Start”. Your refractometer will now measure tds based on the delta of your sample from zero calibration temperature and refractive index. Although refractometers have inbuilt temperature compensation for a given delta between zero calibration temp and sample temp, this can be error prone and inconsistent. If your sample temperature delta is within 0.1°C, continue pressing “Start” untill you get three consecutive readings of the same value. This is your brew’s tds.
• If for some reason your delta is more than 0.1°C despite all of this (usually this happens if your sample size was too large and the rim heated up a fraction too much), unfortunately the only recourse is to wipe the lens with a separate lint free cloth and go through the trifecta of alcohol cleanup + distilled water cleanup + distilled water zeroing. In most cases I have noticed that the rim in the second round registers a temperature that’s closer to the sample temperature previously measured (Atago be a bit finicky that way. As Lance Hedrick will tell you, “Unlike a VST, one really needs to babysit the Atago”). I’m personally curious if the newer generation of refractometers are less finicky.
• Once you’re done measuring, clean with alcohol and distilled water as previously described. If you store your refractometer in a clean environment, you probably won’t need to clean it with alcohol before your next usage. For cleaning between back-to-back uses, I clean with distilled water if measuring for filter coffee, and with both alcohol and distilled water if measuring for espresso. For each sample I keep an eye out for the delta between zero calibration and sample temperature.

Isn’t this painful?

I’m not gonna lie that this is at best inconvenient. But the reason we should strive to get our error margin down is simply because as small an error of 0.02 tds error translates to on an average of 0.25 EY error margin if not more in filter range. That in my experience is a lot when choosing what direction to follow when dialing in. Once can argue that this isn’t as much of an issue in espresso range, however remember that in absence of syringe filters (or centrifuge), you’ve already introduced some error in your reading, and having the refractometer do temperature compensation only adds to that error.

Where’s the Magic?

After learning the cupping spoon technique from Jonathan I have effectively stopped using an Atago Magic for a bunch of reasons. For starters it requires additional cleanup, I have seen it contaminate samples without thorough cleanup and using a tablespoon is equally fast for the same sample size and is better at cooling samples IME. Additionally I’m almost certain that getting any nicks or scratches on the surface will skew the reading to an extent that it’s rendered effectively inaccurate in filter range.

Is it worth it?

To me the refractometer has been an invaluable tool to help guide dialing in. However I’m a big proponent of accompanying tasting notes when providing EY numbers. EY only tells you that something is extracted, not how it tastes to you (I fully realize taste descriptors can be subjective but something is better than nothing). A good example of this is saying something like “I found my tricolate brew to taste sweet with fair amounts of acidity at 22.5% EY, muted at 23% EY and having more flavor separation while losing some sweetness at 22% EY”. Or alternatively “I was able to push to 22% EY on SSP-MP64 burrs with high flavor separation whereas on a Lagom Mini with Moonshine burrs I could only push to 20% with some flavor separation before it got muddy.”

The more you start noting extraction yields for your drinks, the more you’ll learn about every varietal, processing technique, grinders, burrs, changes in technique and brewing device capability among other things. Again, all of these are based on what tastes good to you and that is the only framework you need to care about for context. Similar to espresso recipes, an EY number is only useful in reference to one’s own preferences.

More refractometry resources

In addition to Jonathan’s work I’ve linked to previously, here’s some other resources you may find useful:

• Mitch Hale’s Universal Extraction Calculator (Mitch also talks at length about refractometry across different posts on his blog, including sample filtration)
• Lance Hedrick’s exhaustive video comparing refractometers (he uses the technique I describe above; also watch his brewing tutorials for examples of how to use taste descriptors with EY numbers)

If you’ve read till here, consider supporting NOT ME but Umeko Motoyoshi by buying their fabulous Gay Cupping Spoons (among other things)

Subscribe to keep your pocket updated with coffee science.