Oven Based Coffee Roaster

The popper roaster expired. The motor stopped running (and when I dismantled it, the plastic chamber base was quite badly cracked - I think poppers are not designed for the heat and duration they are subject to when roasting coffee). I'm not sure whether the motor expired just because it was thrashed to death, or whether it didn't like the temperatures, or maybe one of the diodes in the bridge gave up. Also, it only ever did very small batches. I want more...

Drum Roasters - Rotisserie - Frankenstein

The way most commercial roasters work is a drum roaster - you have a rotating drum, coffee inside it being agitated (much like the load in a front-loading washing machine) and hot air driven through.

That doesn't sound so different from a rotisserie, but with a drum substituted in place of the forks. Some people modify gas-fired barbecues and can do pounds at a time, but I don't want that scale, so decided to stick with electric oven.

Electric oven rotisseries, however, are reputed to not have enough oomph - if you don't get the temperature profile hot enough fast enough, the beans don't roast, they bake. While I haven't tested it, this may well be true - the oven I bought was 1400W, which is only about 15% more than an unmodified popper (and slightly less than the rewired element I was using in the popper mark 2). I want it to do lots more beans, so it is likely to want more power. So, time for a frankenstein appliance-meld of my own:

Rotisserie - Convection Oven

I thought I'd take a tabletop oven (with rotisserie) and stick a convection roaster top on it. Each has about 1400W on its own. Together, 2800W really ought to do something interesting.

The mechanical and electrical build is on a highly pictorial separate page.

The controller electronics also features on their own page.

Programming the microcontroller is on yet another page.

Roasting

Uncontrolled - not good

Before coding the controller, but after the mechanicals were functional, there was somewhere to put beans, and it was crying out for a test run. So I hooked up the elements to a PID controller at 250C (wire thermocouple just trapped in the door), and connected the roaster fan and rotisserie simply hard-wired to run whenever it was plugged in. 150g of Colombian came out...

... horribly horribly over-done, in under 10 minutes.

More detail on the roast notes page.

However, the good news is that when it was running at a steady 250C it was only using about 30% of the available heat input to do this too-dark, too-quick roast, so I think that, when tamed, it will be up to the job.

First Roast

This is the first controlled roast. The profile is one I used to use in the mark 2 (popper based) roaster. This is a 200g batch, with target (red) and achieved (blue) temperature logging (one sample per second, which is the interval the controller runs at).

What really matters is not the graph, but the coffee. This is the result of that roast. It was still first-cracking when the roast stopped - for a batch this size it really wants longer (or hotter). I've since discovered that very close to this program - but hold the final temperature to 12m15s - works well for a 150g batch.

What I need to do now is play with the profiles. That's going to be slower than the same exercise was with the little batches from the popper - I'll need to drink a lot more for each profile I try. What a tough life!

Information about more roasts and tweaking of profiles is on the roast notes page.

Even more geekily, there's information about a series of trial runs with a dummy load (ceramic beans) to investigate control parameters, heating and cooling rates, and things like that, on the ceramic beans page.


back to the top roaster page
back to Ian's contents page
back to www.astounding.org.uk top contents page

To comment on anything (please do) email ian@astounding.org.uk