How many home roasterists with variable drum speed options vary their drum speed?

Do you do it based on the batch size and keep it fixed throughout the roast?

Do you vary the drum speed during the roast?

I saw the variable grinder thread and this question came to mind, especially since I varied my drum speed during my last roast on the Bullet to start with a lower drum speed and increase it as the roast progressed. I discovered there's a minimum drum speed to use the trier effectively on my roaster.

I don't touch it.

The previous owner of my roaster said 79 rpm worked best for him. The manufacturer recommends 58 rpm for 1kg; "higher RPM if the batch size is smaller; lower RPM for larger batch size." I haven't tried 58 rpm, but I don't have any facing or tipping at 79 rpm, so I'm not touching it. I picked up a surplus digital tachometer so the roaster is adjusted to 79.0 rpm now .

A person over on the North Roasters forum put a plexiglas door on his 3kg and video'd a whole bunch of different loads at different rpm's. The main thing that I drew from the videos was that there is a minimum amount of beans necessary to get a nice-looking bean "waterfall." Strangely relaxing to watch, although perhaps only directly applicable to similarly-sized drums and stirrer/fin designs. http://www.youtube.com/channel/UCO9UQGe ... W8Q/videos (look at the "another look" series, not the "A look into" series).

I don't adjust drum speed during roast, not even the airflow. Drum speed is a compromise between slow enough to get a proper readout from the thermocouple (sensor buried in the bean mass) and fast enough to get sufficient agitation to mix the beans with incoming hot air. This means with smaller batch size, i will reduce drum speed. Adjust it too fast and beans will stick to the wall drum due to inertia and you'll see uneven roast on one side of some beans.

A yet to be proven theory of mine: best drum speed is when the noise/rumbling sound inside the drum is loudest. This mean the beans are well mixed into incoming air, thrown up high by the fins. Any faster then sound becomes softer as the beans start sticking to the wall.

On my 500gram North I leave it at the max 72 rpm (100 on dial) except when doing 300gram samples. My normal charge is 540 grams so that I can net 454 grams or a pound. When I go down to 300, I reduce it to around 70 on the dial which I recall is around 54. No science behind this but the roasts seems to behave more like my full charges. I never change the speed during the roast. I am also not a fan of multiple air changes during the roast. One is standard around 45 sec post FC and then additional if needed. Hope that helps, Michael

Just today, I ran two more test roasts continuing the experiments carried out on this thread:
Conduction/convection ratio roasting experiments

The tests today I wanted to see effects of varying drums speed. On my North 1K, I read a lot of posts and followed some rule of thumb that resulted in a drum speed of 63rpm. I've used this speed for almost all the roasts I have done since I got the machine.
Today, I ran two identical roasts with drum speed of 50rpm and 73 (max) rpm. The bean was the same used with airflow tests, but unfortunately I did not have enough to charge at 32oz each. Instead, the two were both 29oz. I'm not sure this was a significant difference. I used a fixed air of 0.05"wc measured with the magnehelic. This air matches the mid airflow roast used on air tests which I consider the baseline 63rpm roast.

I'm going to spare you the profiles, since they both look pretty similar. And compared with baseline, they match pretty close. But there was a few things I'd like to share.

1) the 50rpm drum speed roast did not ramp quite as quickly with the same ET (exhaust temp). It was close, but required more heat to match BT profiles. The result was 10-20deg hotter drum going into development.
2) the 72rpm drum speed had the opposite effect, but to a lesser extent. It ramped slightly faster and resulted in maybe 5-10 degree cooler drum going into development. Overall, it matched very closely to baseline 63rpm roast. The end result is in a good direction (IMO), since my ET at end of roast almost matched my BT. So, my guess is it may taste better. Time will tell . .

My conclusion right now is 72rpm and 63 isn't much different. But 50rpm is in the wrong direction for my theory that air can fight drum heat (assuming a bad thing). So, I plan to stay away from low drum speed. But will likely continue roasting with fixed drum speed somewhere in the 63 and 72rpm range.

I suppose charge weight would matter. But that is an experiment for another day. . .

I have an R1 and while I do not adjust the drum speed during the roast, I do have profiles that have different drum speeds. It varies more by type of roast than batch size. I have found that my natural processed coffee benefits from an increase in temp and a decrease in drum speed at the start. That has given me the best fruity flavors.

I rarely adjust my speed and never during a roast. I run around 65 RPM for my normal charge, 500-600g but will decrease the speed if I am doing a small charge, say 150-300g. I would guess there are a lot of variables that will play into the best speed. Dimension of the drum, arrangement of the veins in the drum, etc... The speed controller is there, so there is no harm in playing with different settings to see what happens. Too slow or fast and you can get scorching as the beans keep in contact with the drum to long.

Since my bean rotation and heat magnitude and bean height are closely linked as one segment of variables. I have always played around with drum speed in all kinds of ways (but not every) I understand what I am doing, it is the lack of understanding of what each particular bean variety or variation prefers, that is my current limitation.