Roasters Rock

Folly of a Roaster Geek

“Dang it! I am out of coffee at home. I need to roast something.” And that is how it started. What happened next was an unfolding of roaster geekness that only those of us that roast will understand.

I needed to roast some coffee. Simple enough. Go get a green sample from the stack in the corner left over from old projects and torch it up. A caffeine fix is needed and the beans are green. Green to brown, grind, steep, press, drink. But alas, I distracted myself.

I was recently in Taiwan and visited an Aboriginal village that is reinventing itself as a third wave coffee community. They bring coffee tourists in and allow them to pulp, wash, dry, roast, and prepare coffee in their processing center / tourist house. To roast the coffee they had a terracotta pot that was modeled after some hundreds of years old thing they used to use to roast coffee over the fire. It is a modern piece now with enhancements in the structure and glazing, but ultimately, it feels like a step back in the past. I bought one of these roasters and brought it back to my home office where it looked pretty on a shelf.

It is about time I test this roaster! So I got an ‘Ov-glove’ (a hot pad you wear like a glove) since the pot will get freakin’ hot and decided to roast some coffee over the flames of my gas stove. I decided that I would use this very large sample of coffee I had left over from a small farm in Concordia, Colombia where I had done a project. The coffee has both a brightness and body so it would be a good test.

But wait… I should compare this terracotta roaster with another roaster to see how they compare! I think I will get the old Hearthware air roaster out and compare a convective heat roaster with the almost 100 percent conductive heat roaster in the pot. That will be interesting to cup the results if I can control the devices enough to match roasting parameters. I decided to use the roasting protocol for samples being graded by The Coffee Quality Institute (CQI): Between 8 to12 minutes at whole bean Agrtron 58 +/- 1 and ground color of Agtron 63 +/- 1.

But wait, if I have a convective roaster and a conductive heat roaster, I should get a drum roaster to see what the difference is when I add a little radiant heat. Since I have not got a Probatino handy in my home lab, I wondered what I could do. I KNOW! I bought an antique roaster from the 1800’s off Ebay about ten years ago and I always wanted to try it. The design is quite simple. Picture a soup can with a slider door cut into the side. Get a metal rod with a wood handle and stick it into the can and then you essentially had a little barrel roaster. There is a baffle inside of ‘the can’ and when you spin it, you get action of the beans being moved and tumbled. Add a wire hook to the outside so you can hang it on the hearth in the fireplace and there you have it, an 1800’s home barrel roaster.

Time to Roast!!! I started with the 1800’s roaster. First problem with this roaster, you cannot see the beans once they are inside. Oh well, roast by sound. I also consulted “The Frugal Housewife,” published in 1820, which explained that the proper way to roast coffee was to stop when the beans were ‘dark and oily.’ Not much help; just go for it. As it turns out, in about 9:30 minutes smoke poured from the roaster so I dumped the ‘dark and oily’ beans. I gave this coffee to my neighbor who does not know any better and tried again. This time I used the Ov-Glove to sneak peeks inside and got a nice 11:30 roast into the 1st crack.

Next up was the Taiwan terracotta pot. There is a hole in the top of the pot about the size of a sight-glass on a conventional roaster. The handle is hollow. I first thought this was for smoke to be carried away. As I approached the end of the roast I realized that this was actually for dumping the beans. You just lift the pot and they dump out of the handle. I found that by adding a 21st century iPhone with a flashlight app that I could get a good look at roast development inside the pot. I got this roast out in 9:45 just into 1st crack.

Last was the roaster I was most familiar with. I set up the Hearthware outside since I modified it a bit. In order to get a roast, even approaching eight minutes, I had taken off the chaff collector on top so the air just flows straight through the beans and carries the chaff all over the place. I also periodically switch to cool mode to slow it down. I also manually agitate the machine the whole time to get an even roast. This roast came out in 8:45 minutes just into the 1st crack.

Time to Cup!!! I stole a little of my samples for the days coffee, but waited 24 hours to officially cup my results. I got my daughter to ‘blind’ the coffee for me. I used the SCAA form and was surprised at how different they were.
A: Bright and thin: 83.5
B: Flat and a bitter finish: 81.25
C: Nice balance of acidity and body: 85.25

Then the reveal surprised me.
A=Hearthware. I pretty much guessed that A was the air roaster due to the brightness and lack of sugar development.
B=Taiwan Pot. Wow! I did not expect that. With all of the agitation I did to keep the coffee moving I thought it would have developed nicely rather than take on such a baked profile.
C=1800’s roaster. Well I will be damned! Sometimes older is better. It has more technology than a pot but a far cry from an electric air roaster. I was totally impressed with the results!

I do wish I had not burned my thumb checking roast color development though! My inner roaster geek is going to try a few more experiments now…

Rocky Rhodes is an 18 year coffee veteran, roaster, and Q-Grader Instructor, and his mission now is to transform the coffee supply chain and make sweeping differences in the lives of those that produce the green coffee. Rocky can be reached at

Suggested Reading