24 different profiles of the same coffee on four different roasters.

The following is the information pack that accompanied 24 individually packaged cupping samples being sent all over the globe.

The research is an informal look at how machinery and set up will affect roasting flavour.

A lot of hard work went into sourcing, planning, traveling, organising, roasting and packing this coffee. We hope you learn something valuable.

The Intention of the experiment originally, was to work out the kind of roaster we wanted to install in our business that will open in Oslo in 2017. I believe that investing in the best equipment for your needs the first time is imperative.

However, after reading Scott Rao and Rob Hoos’ work and seeing Martin Wiesinger speak at Re:Co Dublin, we realised these experiments could have multiple applications. Martin spoke about the way that roasters are using Cropster to build profiles, so Jørgen and I thought that our experiment could involve looking at more variables than just the machine. I am especially interested to see if it is possible to use specific curves to influence particular flavours, as it’s been much discussed but rarely tested on a large scale.

The Method was to take 4 roasting machines that have good reputations within the specialty coffee industry. A Loring S15 Falcon at The Factory in Copenhagen. The Diedrich CR25 at Nord in Oslo. The 1958 Probat UG15 at Kaffa in Oslo and the Giesen W15 in Ålesund at Jacu on the West coast of Norway. I wanted to meet the people that use these machines and emulate the way they get the best out of their roaster. I also wanted to test the boundaries of what each model was capable of. So rather than developing one profile with the same specifications, I hoped to push them to the limit to discover the best and worst of each machine.

The Coffee is a delicious 88 point Kenyan Thunguri from Dormans that has been sourced by Collaborative Coffee Source. They were generous enough to offer a discount so that we could purchase 240kg of this coffee with the intention of having 6 roasts on each machine. So thank you CCS!

The reason we chose this coffee is because #1. Kenyan coffees are delicious and I have always loved roasting and drinking them. #2. It has an intense sweetness and acidity that can be manipulated to display different characteristics during roasting with great clarity. #3. It had a great consistency and density. #4. It was vacuum sealed and packed in boxes which made it great for traveling. #5. I believe Kenyan coffees cup roast defect with an excellent transparency.

Coffee Specs:

Thunguri Factory Kibirigwi FCS

Kirinyaga, Kenya

Area: Kirinyaga County

Nearest town: Karatina Town

Altitude: 1600m

Number of farmer members: 1115 in total

The Limitations of performing an experiment like this are significant and I had many respected roasters warn me of the problems we may encounter. Perhaps this is why an experiment of this kind has not been attempted before.

But being aware of the differences also brings understanding. That is why each of the different roasters brings its own strengths and weaknesses, therefore there can be no answer as to which machine is “best”, only which machine is best for you.

So here is a list of some of the factors we had to account for:

  • Different probe thickness, length and placement (will relay heat differently, therefore create different curves)
  • Lack of dual probes on certain models (Some profiles will only have bean temperature reading)
  • Restrictions on flame addition (two of the machines had only 30%-60%-100% flame)
  • Piping to the roaster that manipulated airflow
  • Lack of ability to manually manipulate airflow (some machines had full control whereas some had none)
  • Speed of the cooling function in the tray (some were very fast, others slow)
  • Different drum thickness or lack of drum
  • Different drum speed
  • Some models made it very difficult to hear first crack so development time could be slightly off on those roasters
  • Different generations of machine and different configurations/options on each machine

With all of this in mind, we decided that rather than create a profile to try and emulate the same settings on each machine, we would consult with the person using that specific roaster and get their advice as to its optimum usage. We would then follow their lead on some but then push outside their boundaries on other profiles, to get a broad sense of its capabilities.

One of the main takeaways from this experience is that no one roaster used the same language or communicated their method in the same way. For me, having a shared vocabulary is the first step towards growth and learning. This is something the barista community excels at but it was only possible once people started collecting data and sharing.

Strengths and weaknesses. This is by no means an exhaustive list, nor does this apply to all models or makes.


We tested a Loring Falcon s15 at the Factory in Copenhagen and my favourite thing about this machine was its temperature stability between roasts. You are not tethered to the machine once starting your roast day as it can idle for as long as you need. They also have software that can control the machine from your phone, so if you need to be away from the roaster for an emergency whilst in the middle of a batch, it’s not the end of the world.

The cooling tray was also extremely effective. It was the most effective of the machines by far.

There are also a lot of great accessories with this machine. I was able to weigh the roasts directly into a device that then loaded the coffee into the roaster. So no lifting. It also automatically records temperature and gas data, so you don’t have to spend time logging comments into Cropster, you can just focus 100% on what is happening during the roast.

Some things I found challenging were the software and display that comes with this machine. To me, it felt slow to react so I found myself “double clicking” and turning off the flame entirely. The software interface also seemed at odds with the crisp design of the rest of the machine. The trier also seemed short making it difficult to use, I had multiple occasions when beans got stuck and I struggled getting it back into the machine resulting in spilled coffee in the cooling tray. I am sure you get used to it over time. The nice thing about it is that it doesn’t affect the temperature during roast. My other gripe is that there is no way to manipulate the flame without it being directly tied to airflow. All the other models we tested had these functions separated. Listening to first crack and accurately ascertaining when it occurs was also extremely challenging for me.


Jørgen and I joined Nikolai at Nord and test out his 2015 Diedrich CR25 for this test.

The design and accessories that come with the Diedrich are beautiful and make roasting very functional. Their distributors also do a great job with support and installation.

We found the software that comes with this machine much easier to use than the Loring. It is also very intuitive and reactive. The cooling tray is extremely effective and there is precise control over flame and airflow.

The machine we were testing with included a destoner that allowed some bean sized stones to make it through. When using the load function and the cooling tray simultaneously, this affected the readings we were collecting in Cropster. We saw a dramatic drop in bean and exhaust temperature midway on the first few profiles and as a result we stopped using the load function and were emptying the cooling tray by hand. We’ve asked Diedrich about this experience and will update this article with any new information.


Refurbished Probats are the darlings of the specialty coffee industry. Since I learned to roast on one, they will also hold a nostalgic, warm place in my heart. However, like classic cars they each have their own personalities.

My favourite thing about them is how manual they are. It feels a lot like driving an old sports car. You feel every gear change and bump in the road. Once the drum is hot, there is a lot of ambient temperature so that you can ride conduction through your roasts really effectively. The trier is simple and easy to use but can affect temperature on the bean probe. Listening to first crack is simple as the drum is thin enough for the crack to be very audible. When cooling 10kg of coffee, it took around 5–6 minutes.

However, once you start roasting on a classic like this 1958 UG15, you are absolutely at the mercy of the machine. If you cool for slightly longer than intended, it will affect your next roast significantly, which makes repeatability difficult. Often the first few roasts of the day will be very different to the rest of the profiles because it takes a lot of energy to warm the machine. They are also very susceptible to the weather and conditions in your environment. Changing profiles between the seasons can be challenging and frustrating.

The model we were using had a 3 step flame, so we were locked in at either 30%, 60% or 100% which made incremental flame increases impossible. There was also only one probe in the face plate, which meant less recordable data.

Today’s Probat is a very different animal to this classic from the 50 years ago. But due to their popularity we are all drinking a lot of coffee from these classics and given the impact that have in a roastery/cafe we included one because, why not? It’s a route many of us consider.


We tested on a 2010 Giesen W15 which includes some really useful software if your intention is to set profiles to roast automatically for you. Setting automated roast profiles is a feature more manufacturers are including with their machines. I liked the repeatability when employing this software. However, when moving over to manual roasting, it becomes a little challenging with the functions on this model as they are clunky and hard to move between when making changes.

This machine also had limitations in flame, same as the Probat. So even though you were choosing 40% on the software, the machine itself only had 3 settings of 30%, 60% and 100%.

During the first roast on the Giesen, we had setup issues and Cropster froze, so I lost all information on the profile, which was frustrating. But I took it as a good opportunity to roast “the old fashioned way”, by sight, smell and sound. We included the sample in the pack, although there is no profile information attached to it. It still had the same moisture loss (10.4%) and a similar colour using Colour Track (50.46). Cropster gave us incredible support and backup during this entire week, I was really impressed with their ability to adapt to so many one off scenarios when they were several countries away.

I also struggled with the cooling tray with the Giesen on this model. Half of the tray sits beneath the burners and the drum so that the coffee never really cooled for us. It took up to 12 minutes to get some of the roasts out of the cooling tray. The airflow of the tray is also tied to that of the drum (there is only one fan on this model), so that if you want to restrict airflow during a roast, it’s not possible if you are also cooling a previous roast. The green also loaded very slowly into the drum when charging and there was no way to prop open the door if we wanted to between roasts.

My Beliefs are that as a community, the roasting industry has so much more to gain by sharing ideas and experience. Cropster is a perfect tool to facilitate that because it collects tons of data which can be shared and compared. Even though we can’t control all the variables, we can learn a lot from sharing our experiences.

Every batch is another opportunity for learning if only we collaborate and create a shared vocabulary.

If you wish to have a chance to taste these samples you can participate in any of the following events: Woman Barista Connect in London during September. The International Coffee Organisation in September or Roasters guild of Europe in October.

Thank you to Andy and Paul at Cropster for access to the software and the amazing technical support through our challenging week. Thank you to Michael and Jens at The Factory. Gunar, Anne Birte and Oliver at Jacu. Nikolai at Nord and Bjørnar at Kaffa. We couldn’t have done it without you wonderful people!

Talor Browne.

Co-Founder and Head Roaster .

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.