by Marijn Dokter

Data when roasting coffee

Roasting coffee is all about two key data: temperature and time. Whereas the latter is simply measured from the start of the roast, the former is obtained through temperature probes. These probes measure temperature in certain places. Most of the time there is a probe in the bean mass, and a probe outside of the bean mass, either in the exhaust or above the bean mass. The data on every roaster will differ a bit because of probe placement. For instance at Shokunin (a Giesen W6) we have a first crack temperature of 190°C, whereas my home roaster (a Quest M3S) has a first crack temperature of 203°C. This data is essential if you want to replicate roasts, which is important in a production environment like a roastery. If you’re roasting at home, data is less essential, but can still help you a lot to get better roasts. 

Quick sidenote here: temperatures depend on probe readings and differ for every roaster. The temperatures mentioned in this article are just to give an impression.

Stages of coffee roasting

Before coffee is roasted, it is green and very hard and it smells somewhat dusty. As the coffee is roasted it transitions from green to yellow and eventually to brown. These colour changes are milestones during roasting and can help guide you influence the flavour of the eventual coffee. The first phase during roasting is the drying phase (also called yellowing). This phase lasts until the coffee turns yellow, generally around 150°C.

When the drying phase ends, the maillard phase (also known as browning) begins. In this phase, Maillard reactions start to occur, giving the coffee its brown colour and constantly changing the coffee’s flavour. This phase ends when first crack starts, which is typically around 200°C. First crack is usually easily recognizable: the coffee “cracks” and it makes a sound similar to popcorn popping. The phase that happens after the beginning of first crack is the development time (or post-crack phase). 

Effects of roast time on taste

Before roasting coffee cannot be consumed as we know coffee, so the roasting process has big impact on the flavour of coffee. We usually say a coffee needs to be developed before it can be brewed. If you pull the coffee out shortly after first crack has started, there will be a lot of characteristics left that the coffee originally has. The coffee will have a lot of acidity and a lighter body and mouthfeel. If the coffee is developed more, many of the lighter aroma and acidity compounds will slowly disappear. Sugars are caramelized and therefore the coffee gets less acidic and sweet and more bitter. At Shokunin we roast quite light, because we like to showcase the flavours that different coffees from different farmers have. We try to keep as much flavour intact as we can while still balancing all those flavours out and not imparting any recognizable bitter flavours with the roasting.

Most of the differences in flavour and roast level happen after first crack, during the development phase. But what about the drying phase and maillard phase? They also influence the taste of the coffee. If you push the coffee through these coffees quickly, the coffee will be more lively and “funky”. If you take it a bit slower, the coffee will be a bit more mellow. This is something we play around with when making roast profiles, based on every coffee’s unique flavour profile.

Roasting at Shokunin

At Shokunin we roast fairly light, but we try to properly develop the coffee for espresso and filter, an omniroast. Most of our coffees get to first crack in a little over 8 minutes. We go into first crack at 190°C and most of our coffees are dropped around 195°C to 196°C. Important to note is that even a difference of 0.2°C already gives a noticeable taste difference. This is why we have strict guidelines for our end results to determine if we can sell a batch of roasted coffee or not. We use a fixed airflow during the roast, which depends a bit on the coffee, for some it’s a bit higher than for other coffees. Our drumspeed and batch size, which are always related to each other, are also fixed so that our coffees’ flavours are as consistent as possible.

Roasting at home

Roasting at home is totally different from roasting in a production environment. The best part about home roasting is that there is only one person to please and that is yourself! When home roasting it is not that important to exactly replicate your roasts. We have experience with two different homeroasters, the Gene Cafe CBR-101 and the Quest M3S. For both machines we’ll give you some tips to get the best out of them!

The Gene Cafe is a little limited in its control. It doesn’t have a lot of power, but you can still get a good cup out of it! To get the most out of the Gene it should be preheated to the maximum temperature. Once it reaches the maximum temperature (250°C) you should stop the roaster, load it as fast as possible with 200 grams of coffee and put the chamber back in and turn the heat to max again. The reason for this is that this roaster heats the coffee somewhat slowly, so you need to retain as much energy as possible for adequate development. The only temperature reading the Gene gives you is the air temperature, so you need to pay attention to the beans. The most important thing with the Gene is to use your sensory skills. Depending on the color of the beans and the smell of the beans you have to make the decision of when to end the roast and drop your coffee. To extend the development time you can drop the temperature a bit after first crack occurs.

The second home roaster we have some experience with is the Quest M3S. The Quest is a small drum roaster working on electricity. Compared to a roaster on gas it’s a bit less nimble, but there is no hassle using gas or propane stoves or bottles. A huge advantage of the Quest M3S is the possibility to data log your roasts using probes. With these probes you can track your roasts in data logging software like Cropster or Artisan. This makes it easier to replicate roasts, or to adjust your roast profiles. With the Quest we mostly use maximum airflow throughout the whole roast since the fan is not very strong. The Quest allows for batch sizes from anywhere between 150 grams to 250 grams for proper development and probe readings. Charge temperature is anywhere from 180-200°C, depending on the charge weight. We try to hit the coffee with a lot of heat from the start, until the coffee turns yellow. Then we gradually lower the heat until first crack. With the Quest you need to plan ahead a bit; its electrical elements take a while for an adjustment to kick in. This is a disadvantage of an electrical roaster compared to a gas roaster. However, if you plan your roast well and get good results, replication is easier and you’ll get some delicious brews.

Views on roasting before and after Shokunin

Before I joined Shokunin I was a bit sceptical about omniroasts, mostly because the ones I had I did not like very much. They were either too sour for espresso or a bit roasty for filter coffee. Doing almost exclusively omniroasts at Shokunin is quite a change for me! Before I joined Shokunin I worked at the Stielman coffee bar, which used Shokunin coffee a lot. At this point I was a bit surprised with the coffee, because it worked for filter and for espresso. At Shokunin I learned that to do omniroasts well the body of the roasted coffee is really important. It needs to hold up in an espresso, but it should not be overwhelming for a filter coffee. Using this balance, both the filter coffee and espresso turn out nice. The effect of the omniroasts on our coffee is that our espressos are a bit more punchy whereas our filter coffees are a bit more soothing and smooth. 

Ever since I started roasting coffee under the Shokunin label, one of my key characteristics was the low drumspeed setting that I used. As far as I knew, no one used such a low setting, which is one of the things that made my roasting style unique. Lately, however, I’ve revisited this variable, using all the extra knowledge that I’ve gathered over the past year.

Let’s start with the beginning. Drumspeed is something that is rarely written about, even less so than other roasting variables. Perfect Daily Grind recently wrote about it, but still there is very little substantial content to the article. My view on it is that a lower drumspeed allows for a more subtle, silky mouthfeel, whereas higher drumspeeds create more body and depth to the coffee. This only includes the defect-less range; going too high or low causes scorching on the beans and therefore bitter flavours. This silky mouthfeel is something I always appreciated in coffee, and it what I wanted to incorporate in my flavour style.

Over the past year, I’ve also been experimenting with the Soaking technique. This allows a roaster to transfer heat more effectively to the coffee, and play with the amount of conductional heat in relation to convectional heat. Whereas implementation of this technique did not change any other roasting parameter, it did add a little extra smoothness and aftertaste to my coffee.

Now that I’ve fully incorporated soaking into all roasting profiles, I wanted to revisit the drumspeed issue. Higher drumspeeds used to make the start of the roast too quick, therefore making the coffee a bit too harsh. However, I now wanted to know if there would be a way to prevent this with the use of soaking. The goal was to retain the silky mouthfeel, but to create the extra depth to the coffee, which would make the sweetness even more lingering.

So I did experimental roasts with three different coffees. These, I cupped, I pulled espresso shots at multiple locations, I brewed filter coffee, and with every method, the results were conclusive: the higher drumspeed (48hz / 70rpm instead of 40hz / 60rpm) with a longer soak (1:30 instead of 0:40) was exactly what I sought after. Even though I can be reluctant to change such an integral part of my quality assurance, I feel that this is another step forward in the pursuit for the best quality. Another benefit that made me very excited, though it did not influence the choice, was that it saved 20% gas per roast. In my opinion quality should never suffer from such cost savings, but in this case it was a very welcome extra benefit.

For every roaster, I would recommend to try and find a drumspeed setting with which the coffee doesn’t scorch or tip and that makes controlling the ROR easiest. Even within the range of defect-free roasting, one can make a big difference in the way the coffee reacts to the air temperature. Playing with this in the beginning to get the graphs you want prevents a lot of trouble later down the line!

Having a proper warm-up protocol for your roasting is essential to making your first few batches of the day work. There’s surprisingly many roasters that have such a time-efficient warm-up protocol but who keep having issues with their first roast, be it reaching the targets or even properly developing a high-density coffee.

Every roaster needs a different warm-up procedure, based on size, mass of the material, roasting system, etc. Generally this takes at least 30 minutes. What’s important to realize is that your bean temperature probe reads mostly the air temperature, but only little of the drum’s temperature. What I tend to go for in a warm-up protocol, is a stable air temperature of roughly 40 – 60 degrees Celsius above the charge temperature. Even when the air temperature has reached that point and is stable, you will see the bean temperature slowly rising towards it. Over time, the bean temperature’s increase will stagnate. Usually this takes about 30 minutes, but in the winter it may take longer. Having reached this point, your roasting system loses about as much energy as you’re putting in, meaning you have a stable energy. This would be a good point to start roasting, though one needs to first reach the desired charge temperature, possibly with a between-batch-protocol.

I suggest any roaster to properly put some effort into their warm-up protocols. It saves coffee by creating a stable and reliable first roast of the day. An unstable first roast means more unstable roasts afterwards. Your roasting schedule also gets more flexible if you have the freedom to start your day with any coffee rather than your least important one.

For any roaster, a between-batch-protocol (BBP) is necessary to stabilize the system’s energy before starting a new roast. Since every coffee and every roast will have a different residual energy, one needs to carefully think about a way to reach the same starting energy, not temperature, for every new roast.

We used to go about this using time. For example, after finishing a batch, one would leave the gas off for a minute or so to remove excess heat from the roaster, then gradually bring the bean temperature up to the charging point. This is an easy way to remove some energy from the system and create some stability. However, differences in residual energy are not sufficiently covered this way.

Rather, one should use their gas setting to reach a certain temperature roughly 20 degrees Celsius below the charging point. For example, if I charge batches at 180 degrees, after finishing a batch I would want to reach 160 degrees in about 3 minutes. Then, I would increase the gas setting to bring the temperature up to 180 degrees and charge again.

This way, you’re basing your energy input during the BBP on a coffee’s residual energy. More residual energy means less energy input in those three minutes to reach 160 degrees, so you’re balancing things out. Of course this is just the basic mindset for a BBP; actual data points depend on one’s roasting system and style of roasting.