I have been using an IR thermometer with my air popper with good results, however, it means pretty much dedicating one hand to the task of taking readings with the IR thermometer constantly. The readings also shift considerably depending on where I take the reading; so I tend to throw out high or low readings and just average the readings in my head on the fly. This is a rather tedious process and I was looking for an easier way of doing it.

I bought a K type thermocouple with an extended probe, then fabricated a delrin cylinder with a center hole that would just fit the thermocouple's shaft:


The center hole is then slit along one side so it can be cinched down tight. A small threaded hole perpendicular to the axis is added for the 'cinching' bolt.


Next, I just freehand grounded a slight angle onto one corner and by trial and error fit it to the air popper's slanting upper escutcheon:


I then glued it onto the air popper. This was the only difficult step, as the cylinder is made of delrin, which is like nylon in feel and character, but stronger. It is notoriously difficult to glue, though, as the plastic is self lubricating and most types of glue don't like to stick to it. Mine fell off about a million times with the glue still wet and even after the epoxy set semi-rigid, the thing was still sliding off in slow motion, s-l-o-w-l-y r-e-m-i-n-d-i-n-g m-e t-h-a-t I-'m n-o-t t-o-o b-r-i-g-h-t. I propped it roughly in place after studying my Rube Goldberg manual and leaning all kinds of crazy tools against it in the hope THIS TIME it would stay roughly in place. My aim was to have the probe slightly above the air entry slits and at a slight angle (the bean mass rotates counter clockwise in the popper) so the beans would slide against the probe and not knock it off. In hindsight it may have been easier to just drill a hole and mount it through the sidewall... but I might need a longer probe for that. Its drying now, so I won't know if it'll hold or not until I actually try to roast some beans and see if it falls off when the beans are in there.

Ooops, this is the fixture on the popper:

Arrrrgh! OK, I tried to roast a 90g batch of beans... by the very end, the bottom of the fixture had turned translucent, clear, and you could see through it! very interesting...if I didn't need any other beans roasted in the future. Also, the bottom of the fixture 'peeled' back and the glue didn't hold. I'll have to remount it. Also, the plastic of the probe encasing the metal looks melted from the heat, so the plastic end of the probe needs to be shielded from the heat.

Drill through the casing it is, then.

Here's a different approach to the problem: drill a hole directly in the side of the popper and pass the thermocouple probe through it. There are two points of support for the probe now, the roast chamber wall (metal) and the popper housing (plastic). Careful drilling and alignment allows placement of the thermocouple tip right in the center of the bean mass, away from the chamber walls, or whereever you like.

My thermocouple shows 180deg and coffee is ready in 5 minutes. Quite surprising. I must stirr manually to get even roast. Because there is too much heat I pause after 2 minutes. Going to use temperature regulator with written program for coffee. Can you point me roasting curve I can tweak for my needs? This one is good?

Hi jzagaza! I don't have any roasting curves, sorry. At first it was because the air popper is so simple, just turn on or off. Later, I realized I wanted it to warm/dry the beans slowly and bring them up in temp slower; so I got the router speed control, around $18 on amazon. Aside from that, plus the thermometer, that's all I use and they work fine. But I have to manually ramp up the control (which increase both air flow AND temp same time, they are tied together). During a normal 90g or 100g batch, I ramp it up (starting at medium) about 5 times, in very small steps and my whole roast process is around 13 minutes, sometimes a bit longer. I usually terminate when the beans reach close to 425F or 430F.

With the speed control, lower temps/air flow requires that you stir the beans manually, there isn't enough air to move them well, especially at first when they are heavy. But about halfway through there is enough airflow to spin the bean mass so you don't have to stir anymore. If your beans don't stir by themselves, try smaller batches. I use no more than 100g per batch due to the low settings I use. I have heard other people using chimney extensions, running the popper at an angle, and separating the airflow from heat control (separate controls) doing 30-40% larger batches.

good luck,
walt

Walt - indeed ON/OFF method is much improvement over taste. I bought router controler for 8$ and will test soon

Controler works very well - temperature changes 1 degree Celsius per second, histeresis +-1,5 degree.

This is what I get: 5 minutes 150degC then slowly up, 10 minutes total. 50/50 Kolumbia Kachalu Organic + Indian robusta.

Update 15-12-2015: it taste so great! First time I'm really happy.