Jim
Thanks for your help! It's not easy when I have a roaster unlike any other. No one to compare notes with. I believe my roaster is closest to the Loring in abilities but there are few using one and public data is scarce. As George and Mike get more experience with theirs I hope future chats with Mike will be helpful. My build goals being maximum and adjustable heat transfer combined with energy efficiency. With the last 1.5 lb roast I posted I still had some head room with my heat transfer settings and my available energy. I got a hot experimental race car but I'm a green driver with much still to learn. Knowing what I'm trying to accomplish at every point in the roast along with best ET and times helps to keep me on the road.
cheers,
Ed

another_jim wrote:Sounds like your roaster is running very sweet, since you are getting fast roasts with low ETs.

Jim, would you mind elaborating on that? I have a very wide ET-BT gap on my RK drum setup, it is up to 300F. The reason for this is I prefer to do 4lb roasts for "lifestyle" issues (no time for lots of small batches), and even with the 55rpm motor the RK drum doesn't have the most efficient heat flow outside-to-inside the drum. Inside the drum itself those beans are bouncing like wild so the heat flow seems pretty good.

My perhaps naive view is what matters is the air temp in the drum, and that is what I am basing my roast profile on. The air temp outside the drum seems about as relevant as the air temp 1" above the flames. Note I am measuring the external temperature on the top of the drum about 1" from the spinning drum itself. The internal probe is in the middle of the drum (with respect to right/left) and at a midpoint between the rod and the outside.

Scott

The environmental temperature is the hottest temperature the beans experience, their immediate heat source. The speed of the roast is determined by the difference between that temperature and the current bean temperature multiplied by a heat flow factor. The higher the airflow, the better the heat flow, and the lower the required ET-BT delta to maintain roast speed. If your environmental temperatures get very higher, they affect the outside of the bean, worst case charring them, best case deepening their degree of roast relative to the bean centers. In essence, the lower the required ET for a given profile, the more evenly the bean is roasted.

A fast roast, all other things being equal, will be better than a slow roast, since less of the aromatics are cooked off. Howver, things are rarely all-else-equal. In particular, a roaster has a speed limit for a given charge weight determined by the required environmental temperatures. In a drum, lower charge weights can give you faster roasts for the same ET, and it's your call how many extra roasts you want to do go with lower weights.

Jim, that logic makes perfect sense to me when comparing air temp and internal bean temp in an air roaster, but I don't see how to apply it to a drum roaster. My internal drum probe is halfway to the edge of the drum and so is probably about the average air temperature the beans are experiencing (the beans only touch it when they are falling on it which is about 10% of the time; thus it is primarily tracking air temp). So in some sense that is the average environment temperature. The outside-the-drum temp is irrelevant if it is not making it into the drum, just like the heat of the flame itself. What is obviously happening is there is a temperature gradient inside the drum with the center being somewhat cooler than the sides. What matters is not the temperature outside the drum, but how strong the gradient is across the drum -- a small gradient gives a "sweet" drum roast. If by the edges of the drum it is much hotter than the center then the beans can cook unevenly because they spend lots of time by the edge (I guess -- they aren't there for any long period before moving on to the cooler center and through to the other side; its not clear how much unevenness is added by that). I don't have a very good understanding of the heat flow here I have to say, you have a very hot convecting heat source vs beans a lot cooler and in a vessel only partly open to air circulation (the RK drum skin in area is about 50% open) and spinning rapidly. How much of that convecting heat goes "in" the drum vs "around" it in such a setup? It makes me want to put a probe inside the rotating drum only 1cm from the edge and see what it records in comparison to my existing internal probe.

Scott

Scott
My .02 cents. I don't have much experience with drum roasting but I did consider it when deciding what type of DIY homeroaster to build. In my roaster I measure the ET to control the MET just as it is introduced to the beans. A fan circulates the heat flow with a direct path to the beans with a good STE for replenishment. The main issue being heat circulation. A drum with with 50% opening has 50% blockage of air movement. This area becomes conduction heating including a bit of radiance and the degree of this of course will depend on the metal used and will also depend on the thickness. The open areas of the drum allow the heat flow to enter. That flow will be much hotter upon entering but once in will be quickly absorbed by the beans. It's the MET of the heat flow as it enters the drum that is the concern along with the maximum internal surface temp of the drum, causing either tipping or scorching. To be able to lower the ET outside the drum increasing circulation is the key. either a circulating fan(in what I consider the plenum area of a drum roaster) or I've considered trying to add air scooping fins on the outside of a drum to force more circulation into a drum. Adding plenum circulation can add an issue with chaff control. In my roaster chaff exits the roaster as produced to limit this issue. I also see the rpm of the drum as being a factor too. The higher rpm providing better circulation within the drum but at the same time making it more difficult for heat flow to enter the drum without added plenum circulation pressure.
farm

Ed, I think you gave me a lot more than .02 there, thanks. I did not think about conduction heating from the drum surface itself. Its difficult for me to guess how much that is contributing. The main question is what my MET is and its pretty hard to say without doing some measurements. The temp outside the drum may not be a good guide at all since as you mention the quickly rotating drum can block heat flow from coming in.

In terms of heat circulation, I would say the biggest problem with the drum is the beans are sitting there 3/4ths of the time all piled on each other (I have a 55RPM motor) and only 1/4th of the time dropping. Getting a motor which is fast enough so the beans hardly ever spend any time sitting (say 200RPM) would seem to help things. It would also minimize the conduction heating since they would be in the air most of the time. Of course that could make it even more difficult for the airflow to enter the drum and exacerbate my high MET problem.

Scott

scottfsmith wrote: What matters is not the temperature outside the drum, but how strong the gradient is across the drum -- a small gradient gives a "sweet" drum roast. If by the edges of the drum it is much hotter than the center then the beans can cook unevenly because they spend lots of time by the edge (I guess -- they aren't there for any long period before moving on to the cooler center and through to the other side; its not clear how much unevenness is added by that).

Thanks; this sounds like a much better way of stating Staub's point on METs in drum roasting than my air roasting based description.

It suggests a strategy of having one TC in the bean mass, a few inches, at least, away from the drum surface, and another angled so it almost brushes the drum surface as it rotates away from the heat source (presumably at the bottom).

Finally, a drum roaster seems, in principal, like a very big sauteing pan with a conscientious line cook working it. This means there's similar quality issues: hot spots on the cook surface and anemic agitation will create a bad result. These issues come first, they are purely related to construction quality, and need to be settled prior to any control work. The bean agitation is easy to see. I don't know if an inexpensive IR sensor will help you find drum hot spots or whether it requires thermal imaging, which (afaik) is still too expensive for hobby use.

scottfsmith wrote: In terms of heat circulation, I would say the biggest problem with the drum is the beans are sitting there 3/4ths of the time all piled on each other (I have a 55RPM motor) and only 1/4th of the time dropping. Getting a motor which is fast enough so the beans hardly ever spend any time sitting (say 200RPM) would seem to help things. It would also minimize the conduction heating since they would be in the air most of the time. Of course that could make it even more difficult for the airflow to enter the drum and exacerbate my high MET problem.

Scott

Much faster than 60rpm and centrifugal force starts to take over at 200rpm the beans might be stuck to the drum wall. The new Loring roaster uses a stationary drum with paddles that move the beans inside and avoids the centrifugal force problem.
farm

Doh! make that 88 rpm (according to sdcoffeeroaster).

I did my first 2-thermoprobe-in-drum experiment tonight. I only got good data for the first 5 minutes of the roast; in that period the midpoint probe (in middle of drum r-l and halfway from center rod to drum surface) was around 100F less hot than a probe which was again centered l-r but only 3mm from the edge of the drum (but, still inside it). That was in turn around 200F less hot than the environment probe. The earliest part of the roast has the biggest gap so I think this 5 minutes of data shows the trend.

The data started going wacko after 5 minutes, I'm not sure what happened but the ceramic ends of the probes I had in the grill itself and I think they are getting too hot and melting and messing up the readings. I am going to have to buy some new probes, hopefully with a longer probe rod -- gotta get the ceramic end all the way out of the grill.

Anyway I am starting to get the idea that much of the heat outside is not making it in the drum. Note that this probe by the edge is in the beans ~1/3 of the time which will lower the readings a bit.

Overall I don't know what to make of this; my roasts are now very good. They are running about like the commercial drum roasts, with 1c at ~9 minutes and dump at ~14 minutes. There is no spotting on the beans showing burning. But, there surely is a way to make them better.

Scott

Edit: here is the actual data which I should have included. ET is envt probe. This probe was 2" behind the drum at the same height as the rod; normally I put the probe at the same height as the top of the drum and 2" back from rod (I had to use my normal ET probe in the drum itself since it was a longer one). This ET may have been 50F less than the normal spot. MET is the 3mm from edge probe, inside the drum; MID is the one at midpoint from ctr to edge inside drum. I included column differences so you don't need to compute them. T is time. Note that 1 is actually 1:30 etc (all of by :30) due to a glitch I made.

T-----ET-----ET-MET---MET---MET-MID--MID----ET-MID
1-----500-----254------246-----54-------192-----308
2-----550-----252------298-----68-------230-----320
3-----560-----233------327-----67-------260-----300
4-----600-----245------355-----75-------280-----320
5-----630-----252------378-----78-------300-----330

AVE-----------247.2-------------68.4--------------315.6

Note the MET-MID average of 68.4F is less than the 100F I wrote above; I was just eyeballing the data and it looks like I need my eyeballs checked!

Your "ET" seems like a grill or flame temperature, and isn't a needful part of the picture. Given the sensor melt interfering with the real measures, you may as well ditch it

If your "Mid" is close to bean temperature and your "Met" close to the hottest part of the drum, then the readings look very good as far as they go.