I’ve been making on the fly tweaks to the form of the air box, and have almost nailed down the general design. It is based on older Gothot sample roaster designs that I have seen. I am basically re-creating this from memory and what images I could cull from the interweb. I don’t feel that I need to re-invent this part of the roaster, but am toying with how to build some efficiency into the heating, similiar to how the Kestrel roasters work. That may be phase two though.I need to run through the tests again, but feel that I am fairly close.
Learnings:
Having the air/chaff enter from an overhang (rather than flush with the back wall) creates a nice little dead zone that traps chaff quite well
That if the gap leading to the exist hallway is too large, the fan/impeller will simply pull all the chaff around the corner and out the venting—with lower airflows. The final design will have a slit of a gap.
I may add a fine mesh screen in front of the impeller entrance just to catch any errant flakes, but they could also be caught with t-shaped venting after the impeller (which I should do anyway). A tiny bit of chaff isn’t going to hurt the impeller, after all.
Next I want to repeat this with a larger amount of chaff, and with dry ice (to see what it looks like). and then begin dimensioning this out. I like that it’s really simple and works.After that, I need to add on the air as it comes from the cooling try, which sits on top of the airbox. It’s a bit more tricky, as it will feed both the airbox and the drum—toggling between the two…
Thanksgiving day, Christy drove the girls over to our friends home for dinner while I walked. I was getting a bit squirrelly and the fresh cold air was exactly what I needed. Almost immediately, my mind was engaged with this experimental roaster. I’d been stagnant with any progress since the soapbox derby, and i was really wanting to get back with it. But I was hesitating for some reason.
One of the things that has been nagging at me has been airflow… I had convinced myself that the airflow for roasting 1/2 to 1 pound of coffee wasn’t as important as the drum and bean mass dynamics. Minimal in fact and so I didn’t need to put too much consideration into it… as long as I drew air through the back and out of the front of the drum (the front end was going to be open, like Gothot / Probat and Pinhalenze sample roasters) I believed that I would be able to achieve good control over the roast. But the venting and control that the Gothot has was just sitting there in my mind, picking at my certainty, pointing out that even in small batches, air flow is important.
So I was walking through the neighborhood and re-worked almost everything. Not so much in detail, but mentally re-sketched the entire plan for the roaster. and so now it is more like a grown up roaster, with control over the air flow so that I can control it subtly.
Which brings me to today. I am not trying to re-invent anything, and so built a prototype of the air box based on what I know about the Gothot sample roasters.. an so far just have the chaff collection part built. I should be able to test it in the next couple days and will post images and video of it shortly…
I’m having to admit to myself that being able to build a roaster at the same time as building a derby car is unrealistic. so production is on hold until after August 25th when the Portland Adult Soapbox Derby is being held… or maybe a couple days after that to let us re-coup a little. It at least gives me a little bit of time to accumulate materials.. so I may still make it by the end of September..
Looks like the porcelain enamel I was excited about isn’t all that I thought. Shouldda known. Kinda exactly the opposite of what works for roasting a duck is working against me in the drum design. That and that roasting poultry is a pretty much a linear process in terms of input.. coffee roasting isn’t.
Anyway, I have become (somewhat) initiated into the world of coatings. What’s possible and what’s not. What works well and what doesn’t. Which ones distribute heat, which ones are thermal barriers. Porcelain Enamel, contrary to waht I had read, is a barrier, which is why it works to trap heat (and moisture) in that Le Creuset cast iron dutch oven so well.
So I think I have the right formula for coatings that will distribute the heat efficiently and evenly from the flame to the interior of the drum. It will be a bit tricky to make sure that the ambient temperature around the drum stays stable and responsive to the flame, but I think that with a combination of insulation and coated drum material, I will be in really good shape. So that means I am ready to begin building the drum and the housing. Aluminum and mild steel. I never thought that I would end up with a drum that didn’t have stainless steel as the inner surface of the drum, but here I am, with a design for a solid aluminum drum. Well, it’ll be coated of course but the aluminum will transfer the heat from the heat source to the roast chamber extremely well. Extremely quickly. And because of the coatings, it should be extremely even. Now, all I need to do is build the working prototype.
A couple of tidbits that I just came across… kinda makes me giddy since it may simplify the construction of this drum extensively!
THERMAL CONDUCTIVITY
Porcelain enamel is not a thermal insulator, but it is a relatively good heat conductor when applied in thin coats. Its emissivity characteristics are particularly good. Normally, porcelain enamels are applied so thin there is only a very small temperature gradient through them.
and
Enamel is a relatively good heat conductor when applied in thin coats, however it is not a thermal insulator or diffuser.
Not too worried that doesn’t diffuse heat, that is what the aluminum will be for…
oh, and it’s recyclable (not that ss/aluminum wasn’t, but good to know this is too.)
The past few weeks has been a lesson in the dynamics of falling beans.
I built a couple prototypes for testing how beans will move in the lab sample roaster… This way, I can test, at least visually, how the movement of the beans in the roast chamber will occur. In future iterations of the lab sample roaster, I will be able to exchange barrels out and test different drum configurations and their effect on the development of flavor… But I shouldn’t get ahead of myself… anyway… It’s a beautiful thing.
This is about 0.25 lb of green coffee. As you can see, the vanes are opposing, and in parallel, which will folds the coffee back and forth towards the center of the drum. Three vanes ended up splitting small loads in half which is obviously not a good thing.This is about 1.25 lb of green coffee. To be honest, I liked the way it fell in another test where there were three vanes, but given the problems that were coming up with minimum loads, I think that this is acceptable. The beans stay together well, and though it doesn’t show in this test, the beans integrate/mix within about 45 sec. (That is why there are black beans spread throughout. I painted about 0.25 lb of the beans with black spray paint and laid them in the drum on top of the other beans and timed how fast they integrated. I didn’t think it was necessary to separate them back out…) Anyway, I feel that I have figured out, pretty much as well as I can until I am running real sample roasts what this will look like. It is nice to have the certainty of having seen how the beans will fall. How well they keep together and so on. It’s obviously not a precise science—if it were, roasting wouldn’t be an art—but having gone through this prototyping is giving me both confidence that what I am building will be quality and a much better sense of what is happening inside the drum.This week is my week to draw the mechanicals for this and spec materials… I still need to decide on the construction method. I have a couple options. To slip a stainless steel sleeve inside an aluminum or copper tube, or to press fit machined cooking pots together. I like the latter, but getting pots without the handles, in a reasonable amount of time is my concern. That may actually be a second drum.
I am working my way though the drum dimensions… I figured that it would be easier to build a prototype of the drum than try to envision how the beans are going to fall over the vanes, how many there are, etc… so I built this. My friend Brian (who is helping me build this) squeaked out a little jig that I could put two tire hubs on it—it’s suuper crude, but it works well enough that I can see what is happening inside the drum. The image is just the PVC tubing that I got to simulate the roaster drum before I put in vanes and sealed the back end….
What did I learn? That the depth of the drum is going to be a different ratio to the diameter than I had originally thought—slightly shorter. I need to do more testing around the vanes… I realized (duh) that they need to have a ratio to the depth of the coffee. If the coffee is going to fold correctly, keeping a bean mass, but constantly rotating the position of the coffee beans from being in (brief) contact with the drum wall to positioning it away from the drum wall, I didn’t get it right this time around, but that’s what prototypes are for…
When I am actually roasting, this won’t be the part that gets the most notice, but ya know I am going to completely appreciate it, knowing that the drum won’t be sloshing back and forth. In it’s own, industrial way, I love this thing. It doesn’t have the lower shelf yet—that will be going on when we are placing the motor.
I’m obviously not a metallurgist, nor am I someone who deals with metal piping and tubing on a regular basis… cause I didn’t know the difference till this morning… But now I do. Tubing works off a particular outside diameter.. Pipe presumably works off an interior diameter, though weirdly, it doesn’t necessarily match with the actual ID… hmmmm
Anyway, I got pricing for Stainless and Aluminum, sort of. They didn’t have any of the diameters that I was asking about, which moved me to a slightly different Stainless (316 to 304), so I expect these numbers to shift around.. but I have ballpark numbers so I can begin budgeting. The bummer thing though is that what they said they have available (that I have pricing for) would give me roughly a 6″ dia drum, though I am wanting about 8″.. I will need to shop around some, cause at least I now know it is out there. And too, I did get the name and number of a guy who collects remnants, which should work great cause I don’t really need to buy 16′ of Stainless tubing.
I think that Brian and I have figured out how to build the prototype that will tell me whether or not the idea of clad SS will be a good roasting environment… and for a relatively low amount, be able to test whether I should be buying 1/2, 1/4, or 3/4″ thick outer material. I’m pretty confident that the Stainless should be as thin as I can get it so that it acts as only as a surface, and plays into the thermal properties of the drum wall as little as possible.
There’s no question that having the drum constructed by a good machine shop is the more better of options, but for prototyping this entire idea, I can’t justify the cost without knowing if it is going to work.
Anyway, I have to hand it to Brian, he knows how to build things, and I believe that he is right—that the variation between the way a machinist would create this drum and the way we will, will be minimal and enough that we will be able to know whether it is worth it or not. If it is, and I know how thick the outer wall needs to be, then I’ll think about having it machined.
