Dieter01 wrote:Typically 80% of the heat or more in large commercial roaster. I don't think you can easily conjure up a golden "ratio rule".

I also have read in a few different documents, that heat transfer in drum roasting was indeed estimated to be in the order of 75% through convection. This of course implies that the beans are properly agitated and that the air flow is appropriate to sustain such a transfer and that the drum size and batch size are of the appropriate ratio.

At the risk of contradicting someone else's statement, the percentage of the heat transfer through radiation from the drum surface is probably much less than through conduction. Again this would depend on the amount of surface contact and radiating surface. Each roaster will be different.

In the interest of adding to the knowledge data bank; the Cafemino electric has a 4 inch radius drum that turns at a very sluggish 28.8 RPM. For the record this is way too slow and I have ordered sets of replacement sprockets to change the speed ratio. I will experiment 40, 60 and 90 RPM. I already pretty much know that the 90 RPM will not work with the internal heat element, but it might work with only the external heaters that I have installed.

Hi.

Is there any info on rpms and radius for drum roasters? The larger the radius the faster the outer edge (if same rpms). For example, if the rpms are too high, the beans could end up stuck against the edge (like a washer machine).

Cheers

Arpi wrote:Is there any info on rpms and radius for drum roasters?

The maximum drum speed is, in rotation per minute, (30/Pi)*(g/r)^1/2. Where g = gravitational acceleration (about 9.81 meter/sec), r = the radius of the drum in meter. This is the speed at which the beans are kept in contact with the drum (force of gravity = centrifugal force ).

There is a paper by Henry G. Schwartzberg, (http://www.nt.ntnu.no/users/skoge/prost...s/600b.pdf) (original link by Jim Schulman, thank you again) in which it is stated that optimal should be between 0.9 and 0.75 of that speed.

Personally, I think that at 0.9 of max. the beans are being tossed around a lot yet the remain in contact with the drum a long time! But I will experiment and see.

Whale wrote:The maximum drum speed is, in rotation per minute, (30/Pi)*(g/r)^1/2. Where g = gravitational acceleration (about 9.81 meter/sec), r = the radius of the drum in meter. This is the speed at which the beans are kept in contact with the drum (force of gravity = centrifugal force ).

Thanks. I guess that max speed would for the beans closest to the wall. The other beans (closer to the center) would have a lesser radius. In the paper they mention that the fall of the bean heats up with the hot air. If we assume that there is a fixed amount of hot air, then the emptier the drum (less bean volume), the higher contact of the beans with hot air.

Cheers

The size of the fins, number of fins, size of drum and how much you load the drum will tell you what % of the beans are in contact with the drum at any given time. Although there hasn't been extensive study most equations assume that coffee beans are comparable to wood with regards to heat transfer. Using a few simple equations you can get a ballpark idea of conduction heat transfer.

For convection you need to know how many % of the beans are in the air at any given time. This, the surface area of these beans and the air temperature (which will drop from intake to exhaust postition depending on bean mass and bean temperature) will tell you how effectively you are transferring heat via convection.

In a few weeks when I am back home I will put up some equations if you are interested.

With all due respect, the part abut the wood is nonsense, and the rest of the discussion is making some false assumptions.

-- The beans contact the drum not surface to surface, but point to point. Any prolonged contact results in scorching, not conductive heat transfer.

-- The drum is a radiant surface, and heat transfers heat to the beans by radiation. The rate depends entirely on the drum temperature.

-- The convection heat transfer is complicated, since both the airflow speed through the drum, and the tumble rate of the beans in the drum, affects the rate of heat transfer. However, the higher the ratio of air to beans, the more convection. So a relatively low load has more convection heat transfer for the same airflow and heat settings than a fully charged drum. Which is why low loads roast faster.

-- Finally some drums draw all their ventilation air through the heating section, while others have means to draw in unheated air. In case the ventilation uses tepid or unheated air, so that the air being introduced is close to the beans' own temperature, all the heat gain has to be by radiation.

All this information is available in the technical literatue from Sivetz on. The first industrial convection roaster using recirculated air, the Burns Thermalo, dates from the late 1930s. So do the first solid drums with single pass hot air ventialtion, the Probat and Gothot. Fluid bed and continuous flow roasters were added in the 1950s (and mostly dropped by the 1980s), and since then the basic technology of roasting has been mature Modern roasters have better controls and efficiencies; but the basic designs remain the same.

Thanks for clarifying, it was never my intention to say that bean contact was surface to surface.

I don't think the comparison of heat transfer coefficient with wood is totally off though. Of course its not entirely accurate either, as the coefficient varies some from tree to tree (ex: balsa has an across-the-grain heat transfer coefficient of 0.055, white pine 0.12, and oak 0.17). I expect the same to be true when comparing a robusta with a Kenya AA, the coefficients will be different but probably not too far off from the values above. Does Sivetz specify?

Dieter, there is no conduction, and these heat transfer constants are meaningless. The air to bean heat transfer constant is a logarithmic function of airspeed, and is given in Illy and and the other references. I am not an engineer, I do not know the physics of airflow and convection, so I can't really comment on which wood chip would heat up as fast a s a coffee bean if thrown into a roaster.

The point I'm trying to make is much simpler. The more beans you throw in a given drum, the hotter you need to run the roaster, the more air you need to push through it, or the slower the roast time. For any given roaster, there is a range of loads that performs well, and going lower or higher degrades the roast. I sincerely doubt that with all the variations in airflow, thermal mass, and heating capacity in drums, this ratio is a constant.

another_jim wrote:With all due respect, the part abut the wood is nonsense, and the rest of the discussion is making some false assumptions.

-- The beans contact the drum not surface to surface, but point to point. Any prolonged contact results in scorching, not conductive heat transfer.

-- The drum is a radiant surface, and heat transfers heat to the beans by radiation. The rate depends entirely on the drum temperature.

With the same level of respect, I do not accept the absolute that is implied it the above statements.

The conduction from the hot surface of the drum to the cooler surface of the beans is not to be dismissed. The prolonged contact will only result in scorching when the drum temperature is much higher than the beans resulting in the heat being transferred from the drum to be higher than the bean capacity to absorb and distribute.

For my part, I load the beans at fairly low drum temperature. It is a pain because I must let the drum cool before doing another batch. This low initial drum temperature limits the temperature differential between the beans and the drum. As I have written before the Electric Cafemino (as delivered) has a very sluggish rotation speed and thus keeps the beans in contact (point contact to a degree) for a significant amount of time. This when properly handled does not result in scorching or tipping. This is true even when using external heat instead of the internal radiant heaters.

The fact that there would be scorching is an indication that at the same drum temperature there is much more potential for conduction transfer than radiative transfer.

There is a significant amount of heat being transfered to the beans through conduction in my roaster at least.

I am not disputing that there is truth in the statements above but I do not accept that it is the absolute truth. But feel free to disagree.

Respectfully.

Sylvain