phil_w_robertson wrote:I'm a professional roaster, but I frequently turn to the home roasting community for thoughts and theories that are founded in science and not roaster-specific empirical data.

Welcome to H-B's Home Roasting, Phil! I'm sure I speak for many when I say that we look forward to some roaster-specific empirical data from that sweet Probat UG15 Roaster you drive

Some thoughts from the peanut gallery. With air roasting, heat replenishment is pretty quick and easy. With drum roasting it's seems more about back up reserves of heat and keeping heat travelling through the entire drum surface at maximum efficiency. The mass and amount of surface area of the drum compared to the bean mass is one form of the reserve and I've read that drum roasters also have heat exchangers around the inside of the body of the roaster that are a second reserve. Without this second reserve, could this add to an ET drop being measured in the air close to the outside of the drum? Is enough of a bank of reserve needed to get to the start of 1st crack where then air flow can regulate the finish?
I see efficiency as having the bean mass accessing heat from as much of the area inside the drum as possible at all times. As heat is taken from a spot on the drum surface it then is available to be replenished. Drum speed and balanced agitation helping to increase this heat transfer while maintaining a reasonable MDT?
I see the bean agitation as maybe being more effective than the fan air flow in providing some convection and getting the heat spread out in the drum by disrupting it and moving it from the inner drum surface. This somewhat being what Max is seeing with minimal fan flow?

Max, In terms of accurate measurement of ET, that is fairly straightforward. If the x* and y* are the real measurements, then the taken measurements x and y have to have y same scale and offset y = a + b*y* and x = a + b*x* If they don't, the basic heat transfer measurement doesn't work. This is easy to check with a sacrificial load of beans. Let the ET settle at 350 to 375 and hold it there. Now drop in the beans. They shpould asymptote at very nearly the same temperature on the BT probe, and the minute by minute rise should remain proportional to the intial temperature difference. If your ET and BT probes are obeying this relation well; the roaster is properly instrumented.

Ed, the simulation curves have the simplifying assumption of instantaneous heat transfer from ET to BT (there can be lots of storage and delay going from the heat input to ET, that is not part of the simulation). the basic premise of the simulation is that degrees rise per minute per 100 degree difference in ET and BT is a constant for a particular roaster at a particular bean charge and airflow setting.

• Delta BT = k*(ET - BT)

This simplifying assumption has been very useful for me computing profiles on my air roaster. I can input my desired profile and get my stats program to spit out the ET ramp soak settings I need to 30 second accuracy.

I do not know how accurate it will be on the drum, but I have no reason to suspect it will be a lot worse. Storage effects in the transfer from ET to BT will show up as a difference in the gain rate when the ET is changing trajectory. This only happens at the end of the drying period and the heading into the first crack, so it should not be a major error source. I hate to introduce time dependency, since it turns the simple intuitive one line functions I'm using now into longer unintuitive programs with state variables and lots of places for bugs to hide.

Jim
Your approach and reasoning seems sound to me. I see no reason why the drum should be less consistent all things considered. I also think it's good your are figuring these things out with only the insulation mod. being done. This gives a good baseline for the roaster as stock.

The insulation has either no effect or only a very mild one the roasting dynamics. It saves on power, and it stabilizes the ET readins, making it easier to adjust the power and air. I think anything I find will work on a stock roaster fitted with an ET probe.

When I'm done, I'll get in touch with Molly and see if they'll offer a an option of drilling a tapped hole in the front plate at the edge, so one can install the same dial thermometer or TC well for ET as they use for BT.

The good news is that the simulation works, and I was able to get the profile shape I wanted using exactly the environmental ramps shown in the previous post. I now know how to operate this roaster.



The mixed/bad news is that the M3 only has a heat transfer rate of 15 degrees per minute per hundred degree difference. So this roaster will never set speed records, and this roast was about 3 to 4 minutes slower than on my P1.

The other good news is that this was a good roast: sweet, no ashiness, fairly lively fruit, although not quite at the level of the P1, and the nice complex roast tastes I can never quite get on the P1. Neither going high on the ET before the first crack, nor dropping it seems to be a deal breaker.

I'm ready to start roasting some good coffees and doing taste tests.

Hi Jim.

Nice progress.

I have the feeling that installing the exhaust mod made a faster roast last weekend. Haven't done any measurements. Maybe it has to do with the back pressure (or the pipe being hot).

I would like to know your opinion in using BT rate of change as a kind of thermal inertia indicator. This measurement, reported in another thread, should help anticipate changes to power. But I don't know how it compares to ET. I have use it in real time and it looks very helpful. Using the rate of change as second indicator would also allow to use only one thermocouple instead of two.

Cheers

The ET responds much faster to control inputs, and it determines the BT; so from a control quality viewpoint, you can always roast better using it rather than BT alone. Moreover, it tells you if you are in danger of charring the roast.

It's not necessary once you've settled on standard charge weights and profiles; but I can't see any better way of learning the roaster or experimenting with profiles and weights.

Thanks Jim. I'll lower a thermocouple in the mouth of the feeder at drum level next weekend to see what readings I get.

Cheers

I'm not sure what you mean, it sounds like an exhaust sensor. Mine is beneath the skin, but outside the drum.