chang00 wrote:Rich, please do read the entire posts and not quote out of context.

I read the entire post and my quotation and the following discussion were in context. The suggestions that I respectively did not and did were unnecessary. Keep it above the belt.

Also, I was correct.

Generally speaking, in a solid drum one pass roaster, convection is responsible for about 85% of heat transfer, and conduction from the drum to beans in perhaps about 5%. If you have access, search Dr Henry Schwartzberg's writing regarding this subject and calculation.

I'm familiar with Schwartzberg, and have had some contact with some of the stuff he's written; particularly the article on batch roasting and some of his non-coffee, "food chemistry" articles.

Schwartzberg is all about the chemistry, and my chemistry is limited to what I remember from my brief, pre-med days. I had a great deal more math and physics in school than chemistry; no flair for chemistry then, less than none now. While I can do the math, and have a general grasp of roasting principles, Schwartzberg is otherwise over my head.

Setting aside performing the measurements and turning the Rayleigh number's calculation crank, the "85% convection" figure is available elsewhere, including the Probat website. The Probat article figures the remaining 15% proportion of heat energy transfer as coming via conduction -- whether by contact of drum wall to bean or bean to bean -- rather than the 5% you cited. Probat's number makes more sense to me than yours, since the potential for radiant energy transfer is obviously rather limited in this context -- especially with a solid drum roaster such as mine.

While it's nice to have a general grasp of rough proportions of convection vs contact, when it comes to roasting as a practical art, I'm very interested in the effects that controlling gas and air-flow as they relate to contact and conduction, and more so as to how their manipulation relates to various aspects of the final product. I have no interest in performing the various measurements necessary, or even a vague understanding of how to do them in order to calculate the Rayleigh number. If you gave me the measurements, you probably couldn't pay me enough to turn the crank -- even if I still can, which I very much doubt.

Bottom Line:
The more I roast, the more I appreciate that roasting has the same horseshoes and hand-grenades quality of most cooking. That is, close counts while over-focusing on extreme precision provides no real benefits. Another way roasting is similar to cooking is the importance of being there and being in the moment.

I use more or less heat, more or less time, and more or less air through the succession of intervals between roast eventis, varying them from bean to bean, to do things like profile flavors, develop aroma, lock in fruit notes that otherwise might fade quickly, things like that. In those respects, I don't regard the lack of rotational speed control in my three roasters as much of a handicap. But I'm interested in hearing the hows and whys you use it, at a level of specificity deeper than "better."

By way of a few examples:

• Would 45rpm give you more or less chocolate, and/or more or less acidy fruit than 55rpm?
• Does having speed control make it more likely, less likely, or alter in any way, the likelihood or degree to which you'd manipulate other parameters?
• If so, how so?

Rich

You should roast guided by sense and senses, doing lots of things roughly right; but doing physics that way, not so much. If you make assertions about the effect of drum speed on bean loft, boringly abstract high school crap like measuring speed, force, and loft in actual numbers and units is kind of relevant.

The whole design of the roaster will determine drum speed. If there is useful air flow then design the fins and rpm to create a pattern to utilize it evenly. If not the splay will be designed to better utilize the conduction and radiant in best proportions. With both situations some speed variability can be incorporated to change effects.
I seem to remember Marty posting about drum speed way in the past, somewhere.

The ~15% Probat figure is convective heat transfer of the roaster gas to the drum, based on inlet and outlet gas temperature, and bean to bean. Of the energy received by the coffee beans, ~5% is transferred by conduction from drum to the beans. I stand by my statement.

Additionally, the Probat figure is based likely on a recirculating roaster. When air is heated and burnt, the composition changes, ie, the oxygen is used up and the density changes. Because air is mostly nitrogen, the resulting recirculating air will have different heat capacity, and the numbers will be different than a single pass roaster.

Sorry ... late to the party .. probably a good thing.
I have a Revolution 500 (Dustin's import) which does have a variable drum speed. It roasts better at the higher drum speed .. more beans are tossed about in the air ... at least on my machine.
1) verified by watching the beans through the sight port and trier hole.
2) also suggested by the manufacturer in a discussion I had with them.
3) and turning the revolution rate to just off maximum and noticing a better roast characteristic.
The Rev 500 had 2 sets of fins, solid drum, with a standard perforated back to allow hot air brought in over the burners into and through the chamber.
IMHO I think each roaster has it's own "sweet spot" of rotation based on observation and taste.
Love math and science but with all the variables in roasting (air flow design, drum design, drum rotation speed, fin design, burner design, air speed / damper control) I'm not sure any formula can be successfully applied to all machines.

Interesting thread ..
Steve

another_jim wrote: If you make assertions about the effect of drum speed on bean loft, boringly abstract high school crap like measuring speed, force, and loft in actual numbers and units is kind of relevant.

Looks like Jim put the ball in your court, Henry.

More generally, if enough is known about a given system, modeling can provide some good predictions. The boundaries of "enough," depend on the question.

By the way Jim, I went to and read the thread you linked for AndyS's "physics" in an earlier post. Unless I missed something there was no recognizable physics there from AndyS nor anyone else. For that matter, the word "physics" was not used.

hankua wrote:Rich, you bring good discussion to the table. But I'm still confused on your position regarding drum speed's relationship to conduction. Let's say for example the air flow was 25%, would a change in drum speed affect conduction? Or would conduction stay constant?

Whether rotational speed will effect contact time of some given bean to drum wall is just vector analysis. If you know the drum diameter it's not much of a problem to calculate how long that bean will stay pressed against the drum for any rate of rotation.

The more variable you introduce, the more difficult the problem becomes as it leaves the realm of vector analysis and enters those of stat mech and fluid dynamics. (I don't know about you, but I get headaches and nose bleeds every time I hear, read, write, or say the word "eddy.") The operative fluid dynamics' (ouch) concept here is (I think) the Rayleigh number. For a given situation, it's calculation depends on the measurements for a bunch of different parameters.

Without knowing a great deal more no one can address a figure like 25%. That said, I've got enough experience with typical drum roasters, a sufficient (but rudimentary understanding) of thermodynamics and stat mech, and other blah blah blah about what a great guy I am, to think that in a typical drum roaster you'll a higher ratio of convection to conduction at 25% than you would at 0%.

It's mathematically safe to say that if you vary airflow, while holding everything else constant, the Rayleigh number will vary as well. However, when you change more than one parameter, things become more complicated.

chang00 wrote:The ~15% Probat figure is convective heat transfer of the roaster gas to the drum, based on inlet and outlet gas temperature, and bean to bean.

How do you know that? I'm not challenging whether or not you're right, it's just that I tried researching the assertions made in the article and came up with bupkis.

Worth mentioning -- for the benefit of others reading this -- you have to remember that even if the 85/15 ratio is exactly right for some given roaster, it's only an average reflecting the entire roast and not an exact ratio for any particular time. Those vary depending on a number of roaster settings -- including rotational speed.

chang00 wrote:Of the energy received by the coffee beans, ~5% is transferred by conduction from drum to the beans.

Assuming you're right, and (safely) assuming that there's very little transference by radiance in a solid drum than the remaining, 10% must come from bean to bean conduction (as Probat asserts). Right? Or is there something else?

Whatever the proportion of conduction via drum to bean and via bean to bean, we have to look at the dynamics of bean to bean. Intuitively, it seems to me that's going to be dependent somewhat on how stable the bean mass is at any given point along its journey, and to some extent that's going to depend on rotational velocity as well.

I'm just glad I'm not being paid to come up with the numbers.

JavaMD wrote:verified by watching the beans through the sight port and trier hole

Problematic. Your view of the roaster's interior from the sight glass is limited to the front wall of the roaster. And, I think, mostly what you see through the sight glass with varying speeds is the effects of different trajectories.

More "agitation" in the air as the beans tumble against the glass doesn't necessarily mean more agitation against the drum wall during the time a given bean is in contact with the drum wall, nor during the time the beans spend in contact with one another shielded (to some extent) from the effects of air flow.

Speaking of trajectory depending (to some extent) on rate of rotation... A higher rate of rotation means a higher (along the vertical axis) release point, AND more angular moment. A higher release point means a longer fall, so does the effect of angular moment (the momentum makes it want to fall in an arc, not straight down). A longer fall means more time in the air. More time in the air means more air to bean contact.

Very complicated stuff.

Rich

boar_d_laze wrote:Problematic. Your view of the roaster's interior from the sight glass is limited to the front wall of the roaster. And, I think, mostly what you see through the sight glass with varying speeds is the effects of different trajectories.

What about using some plexiglass temporary substituting the front door as in this video by Casino Mocca?

Stumbled on this thread while reading up on roasting... If I understand this thread:
Higher agitation + higher heat energy entering the system = more even + quicker roast
this is what I knew before reading this thread.

Higher agitation @ same level of heat energy entering the system could result in more evenly roasted beans but a bit slower due to some loss of heat? Is this ultimately what Rich is trying to convey?

This is an interesting subject, however, I am equally confused after reading all the posts!

SomersetDee wrote: This is an interesting subject, however, I am equally confused after reading all the posts!

Go home and roast some coffee. It will help take your mind off it.