Sample Roesting

 
 

In this post, I’d like to dive into our sample roasting system at Prodigal. 


Why do we sample roast?

If you want to cup samples before purchasing green coffee, you have two choices: ask suppliers for roasted samples, or roast samples yourself. Many suppliers are willing to send roasted samples, but those samples may not be at your preferred roast color, and they may not be at a consistent roast color. Roast quality and color can have a big impact on perceived cup score and buying decisions, so if you are serious about choosing the right green for your company, it is a good idea to do your own sample roasting, using a system that yields your preferred roast level consistently. 

At Prodigal, our preferred roast level is close to the lightest possible color that does not produce underdeveloped flavors such as hay, grass, corn, and cereal. I’ve never known an importer or green supplier to roast samples as light as we do, so for us it’s essential to do our own sample roasting. Suppliers usually don’t roast that light for many reasons, including local preferences, hedging against underdevelopment, tradition, wanting immediate feedback about samples (very light samples may take time to “open up”), and possibly because adding a touch of development masks some green defects.

Tools Needed for sample roasting

  • Precision scale (I prefer 0.01g)

  • color meter

  • grinder (for color measurement)

  • Spotlight (for trier samples)

  • Green-coffee moisture meter 

  • UV flashlight (See Christopher Feran’s blog post about how and why to use a UV light for green analysis)

What makes a good sample roaster? Here is my wishlist: 

  • Fast roasts

  • High heat-transfer efficiency (the ratio of useful heat transferred to the total heat input)

  • Integrated,  user-friendly software to manage roasts, profiles, and a database of past roasts. 

  • Hot-air roasting with inlet-temperature control

  • Automated BBP for 10/10 consistency

  • Reliable

  • Optional: drops roasts automatically based on a variety of metrics 


The only sample roaster that ticks every one of those boxes is the Roest L200 (née L100). It’s why companies like Tim Wendelboe, Coffee Collectif, Dak, Regalia, Sey, La Cabra, Code Black, Aviary, Prodigal, and many other top-shelf roasters use the Roest. Those companies focus first and foremost on finding stellar green coffee, and if your number-one priority is finding the best green, then you need the best-possible sample roaster. 

Workflow

Before I had ever heard the word, I was obsessed with workflow. I remember as a little kid sitting in a pizza shop, calculating their likely revenues, and thinking about how the shop might be laid out differently to facilitate faster service (I was a weird kid). When it comes to cafe operations and roasting, efficient workflow is critical to profit margins. Sure, a Roest costs more than some sample roasters, but the time savings I get from using the machine are immense. The ROI from that time savings makes the Roest a bargain. Ikawa and Kaffelogic machines can roast almost as fast (though not as consistently), but they cannot roast and cool at the same time. That makes the Roest significantly faster. If one really wanted to push the limits, one could conceivably roast fifteen (!) 200g batches per hour (3kg) in the L200 Ultra. 

(A little ROI math: if one does 40 sample roasts per week and saves three minutes per batch, that’s 120 minutes per week, or 100 hours, per year. At $25/hour, that’s a savings of $2500/year. Whatever the additional cost of the L200 over some other machine, that cost will be recouped in 1—3 years, and that time savings will persist for the life of the machine. I can’t imagine going back to a slower machine or a machine that cannot roast and cool simultaneously.

Our workflow includes precisely weighing green batches to 0.01g, and measuring the weight loss, green moisture, whole bean color, and ground color of every batch as soon as they are cool. We also measure UV fluorescence and log a lot of information about each lot, so we prefer a more leisurely pace of seven batches per hour. To be honest, the Roest is so efficient that once I’m set up with the green batches weighed out, I spend about half of my sample-roasting time answering emails. I only pay attention the machine after first crack begins. That makes the Roest’s time-savings ROI even higher.

How to (nearly) guarantee your samples roast to your preferred color

If your system meets these requirements, you can predict the exact roast time required to produce your preferred (ground) roast color, even when roasting a coffee that is new to you: 

  • use an air roaster with inlet-temperature profiling

  • use an effective BBP that yields perfectly consistent curves

  • use one profile for all samples

  • measure the moisture of the green coffee

We use such a system to predict when to drop sample roasts, such that over 90% of sample roasts are within +/-2 of our ground-color target on our first try. As in: give us a coffee we have never roasted before, and we already know at what roast time it will reach our desired color. We rarely have to re-roast a sample due to missing the target. We could not implement this system with any machine other than the Roest L200. This system not only helps us choose green based on fair comparisons (all samples at the same color), but it saves us time, since we almost never have to re-roast samples. 

Why the “Ultra”

Since day one at Prodigal, we have sample roasted on the Roast L200. I loved the original machine for its ease of use, fast roasts, and precision roast replication via its automated BBP. The original Roest machine was great, but through conversations with Sverre and the team at Roest, we came up with the idea of changing the drum direction to improve heat-transfer efficiency by decreasing air bypass*. Sverre modified our machine to change drum direction after charging. Our machine proved the concept, and led to the development of the L200 Ultra, which is insanely powerful and fast (it can roast full batches as fast as 4:00), eliminates all perceivable effects of conduction, and has incredible energy efficiency. You may notice in the video I am roasting on the L200 Plus, but that machine had been modified to have counterflow (the drum direction reverses after charging.)

What is air bypass, and why do we care about it?

Just like water can bypass the grounds in a V60, air can bypass the bean pile in a roasting machine. Bypass air is the air that manages to go around the bean pile, instead of through it. Air bypass is the reason most air roasters function best near full capacity (larger batches allow less air bypass.) In some machines this is a small concern, but in other machines it is the most important thing to know about using the machine. Higher heat-transfer efficiency saves energy and can yield faster roasts with a given amount of energy input.

In the case of air roasters, bypass air can force the user to increase the heat input (whether via the power setting or the inlet-air temperature), which may make the metal of the roasting chamber hotter. That hotter metal may or may not influence flavor, depending on how hot it gets.

Scott Rao