The Equation to a Perfect Extraction

The Equation to a Perfect Extraction

27 November 2014

Back when I used to train baristas (about 15 years ago) it was difficult to avoid overwhelming newcomers with the complexities of espresso extraction. We definitely spent more time training them on the intricacies of extraction than we did on how to texture milk. Most new baristas seemed to be more interested, however, in improving their milk texturing skills than in coming to grips with the seemingly endless variables of espresso extraction.

They seemed to naively think that espresso extraction was relatively simple, while latte art was more of a challenge. Latte art does seem to be a great motivator in engaging people with an espresso-based menu. However, in recent years, latte art has become more commonplace, helping to bring the complexity of extraction back into focus.

For cafes to compete in today's market, it's not enough just to pour a rosetta any more. Baristas have to know all about their coffee and how to get the best out of it. Roasters have tried to help baristas by listing the details of coffee varietals, farm locations and roast styles on packaging. A recent development, however, are roasters listing the recommended espresso brew 'recipes' or ratios. The ratios are written in a way that assumes baristas are familiar with, and using, gram scales for grinding and weighing liquid shots and TDS (Total Dissolved Solids) meters.

While on the surface, this may seem like a logical approach to assisting baristas in achieving ideal extraction, I would argue that it opens up a Pandora's Box of issues.

The TDS refractometer has been marketed globally to baristas with a great cloud of seemingly creditable scientific fanfare. I believe, however, that the end result is not that reliable. In my experience, I've found that the refractometer can be highly inaccurate and that the methodology is so inconsistent that the information for the most part is pretty useless.

So, it's unfortunate for me to witness coffee roasters using the refractometre to almost bully baristas into directing them on how to use their coffee beans. Well-intentioned baristas are set a task endlessly chasing an unattainable goal, like Sisyphus, the Greek king punished for not being able to perform the impossible task of rolling an immense boulder up a hill.

Let me explain. Refractometers work on the basis of measuring how much soluble substances bend the light as it passes through a liquid. Given that different substances bend light at different rates, it is essential to know what substances are already in the liquid apart from coffee.

Damien Armstrong, from water filtration company Pure Coffee Australia, has explained to me that TDS in Sydney water can vary as much as 70 - 160 parts per million (ppm) and Melbourne water can be 30 - 110 ppm.

But the refractometer only measures the total TDS, not the type. In taste tests we have conducted with individual minerals that are commonly found in water, each individual mineral can affect the taste of coffee like chalk and cheese (or perhaps sugar and salt would be a better analogy).

On top of this, a refractometer has to be temperature-stablised by allowing 30 - 60 minutes of ambient temperature prior to use. Otherwise, it will give you an incorrect reading. The temperature of water will also affect the rate at which light is bent as it passes through water.

To further complicate matters, certain coffee compounds in the TDS can only be read on a meter at specific temperatures. So, as the coffee starts to cool, they will disappear on your reading. A coffee has to be read at exactly the same temperature every time to get a consistent reading. To do a really accurate reading a barista should deduct the TDS that is contained in the water to start with, to get a true coffeeonly extraction reading. To date, I have yet to see any coffee professional do this properly.

To top it off, there are different micron filters that are used to remove colloidal material in the brew that affect the refractometer reading. Each different size filter will give you a different result. There is no standard recommended filter size for the coffee industry refractometer. So, many coffee professionals are doing the equivalent of passionately comparing Velencia oranges to Pink Lady apples and expecting them to be the same.

All in all, this can start to really do your head in. And yet, it is this analysis that is driving all the new brew ratios.

It is pretty amazing how espresso extraction has come full circle in terms of a more scientific approach versus an artisan approach. Fifteen years ago we were setting grinders with gram scales and timers. Then, we found we got a more consistent result by using a more artisan approach, focusing on volumes rather than weights and visual flow rates rather than extraction yield percentages. Now we are back at it with our scales, timing and measuring. But I'm not sure if espresso coffee is tasting much better.

I do know that when I came across the TDS evaluation kit with its recommended brew ratios in Milan four or five years ago, the closer we got to their recommended 'ideal' extraction ratio window, the worse the coffee tasted. Now it may have just been the local water that caused the problem, but it made me very wary of the supposed benefits of this "renaissance" of the scientific approach to brewing coffee.

One good barista in a highly sought after city cafe told me that he altered the brew recipe-ratio from the one that the coffee roaster had insisted - like a dictator - that they use. When the coffee roaster came by and unknowingly tasted his altered brew, the roaster proclaimed how good it was, completely oblivious to the fact that he had just tasted a completely different brew recipe.

When it comes to espresso brew ratios it is interesting to note, now that shots are being weighed, that we have moved from traditional visual ristretto shots of 15 - 25 millilitres, espresso shots of 30 millilitres and lungo shots of 40 millilitres all using 7 - 10 grams of ground coffee (and all including crema), to shots that can be made from a double 22 grams of ground coffee and 50 grams of weighed liquid espresso. This would equate to two traditional lungo shots including crema, given that the volume of crema will be greater due to the volatility of the younger "new-wave" beans.

The trouble with lungo shots is that these shots tend to taste more astringent than either espresso or ristretto shots. This is because coffee beans are predominantly made of cellulose (wood fibre) and the dynamics of filter baskets don't efficiently deal with this material.

In defence of the old-school doubleristretto, or "rizzy", it did become confused with what can only be called a "super ristretto", where a barista uses a naked portafilter and brews with approximately 22 grams of coffee to produce as little as 10 - 15 millilitres of liquid. This is an extreme ratio and tastes way too intense for most people. A modern approximate ratio of 30 millilitres of liquid from 20 grams of coffee is the rough equivalent of an old-school double rizzy and is still a nice tasting shot.

Among all this confusion, there are a few simple tools to ensuring a proper extraction. One is water treatment. I recommend Reverse Osmosis water that has the precise natural minerals added back in every time we brew. Just make sure they are the right minerals (tip: magnesium makes coffee taste bad). Also a clean machine makes a big difference. And this is one thing that science has thankfully improved greatly in recent years.

This article was also published at BeanScene Magazine