I've been hunting through old threads on this site trying to figure out my upgrade path for a home roaster. I've been heat-gun/dogbowl to home-made roasters 1 & 2 to iroast2 to behmor. I've been looking at small shop roasters, but any that have significant instrumentation or are capable of profiling are quite expensive. I've been looking at the Hottop, which is probably the top of the "home" machines, but if i buy it i'm going to use a variac on the fan and heating element.

Anyway, reading all of these old threads, and considering my experience going from the iroast2 to the Behmor, I wanted to throw the following hypothesis out there. I do not have the means to test it right now, nor am i asking anyone else to, but depending on how this thread goes it may be something i try to test in the future. OK, here's the hypothesis:

It is best to keep the environmental temperature slightly above bean temperature and control the rate of heat transfer by varying the airflow.

I don't have a precise answer for "slightly" but my guess would be about 10 degrees. I also imagine the above statement if true, would only hold true for a range of ramps - don't get all technical about the extremes.

What does everyone think? Personally, I don't believe the above statement is true, but I think there's a possibility it may be true, and it may be the way forward for roasting. What i'm looking for is someone to either knock off my philosopher's hat or pat me on the back. Any thoughts?

I agree that minimizing the difference is a good way to increase roast quality, but 10F may be a little too ambitious.

The speed at which the beans rise in temperature depends on how much higher the environmental temperature is, and on how fast the airflow is. I run about 35F environmental temperature above bean temperature at the end of the roast to get a 12.5 degree rise per minute. Before the first crack, where I'm running at 30F rise per minute, the supply temp runs around 55F higher.

So if you shoot for a 10 degree difference, you'll either need to design a hurricane roaster, or go for very slow roasts.

On the other hand: a pure air roaster is not really ideal for fast convective heat transfer, since the beans are moving along in the air current; it is, in effect, an iso-flow (?) heat exchanger. A convective system like the SC/TO could be better, since the airflow is up and down, and the beans move sideways, so that it works like a cross-flow heat exchanger. Of course, the controls need to be much better, and the airflow design probably also needs improvement, but for a minimum |ET - BT| system, that would probably be the most promising initial design route.

Hmm, thanks for the thoughts. I was just throwing 10 degrees out there as an example, but i think you're right Jim, it's probably got to be more than that. Are there any good resources out there that talk about the different mechanisms of heat transfer and how they affect the beans differently? I guess one of my underlying assumptions is that it is always best to heat the bean as uniformly as possible - that the core of the bean should not significantly lag its surface temp. I know this is true for the extreme - I had a couple of roasts on my iroast that were almost second crack on the bean surface and barely past first for the center of the bean (at least that's how it seemed). They weren't too great. But i suppose that smaller beans would roast better if this were true in all cases.

I wonder if it's worth taking the time to chop grean beans into smaller particles before roasting them - not so small that they stale as they roast, but small enough to roast them next to full size beans and see if bean size is any factor. I guess a drum roaster would be out for this, because the holes in the drum would be too big. Has anyone done this?

On the surface though, it appears that Jim, you favor a fixed temperature differential while Ed, you don't think that's optimum and would rather play with different types of heat transfer. Is that accurate?

The speed at which the bean temperature rises is proportional to the difference between the environmental and bean temperatures. So speeding or slowing the roast profile means changing the difference between the two.

Typically, the bean center is a good eal lighter than the surface, and getting them even requires a roast that is too slow. A typical commercial drum profile yields about a 5 agtron color difference between whole bean and ground. If you get a lot mroe than that, the roast was either blindingly fast, or the the drying period was inadequate.

The chapter in Illy is hard going but good reading. Schenker's roasting dissertation is a good introduction to the scientific literature