Effect of Fan Speed


This discussion started in the Ikawa espresso profile topic.

A few people there seemed to think that there are probably a variety of nuanced and interconnected effects from fan speed throughout the roast, and I am Inclined to agree. That said, I’d love to get some ideas going about the more concrete and obvious effects of fan speed as well.

My first guess is that the most basic effect of a good fan speed setting is evenness.

An obvious bad effect from too high a fan speed is losing beans. Some theoretical bad effects are causing unevenness like facing and “aroma stripping.”

Some questions I have include: does the increasing fan speed at the end of a profile like the Andino do anything to help maintain a RoR greater than a decreasing fan speed would? Does a lower fan speed lead to a greater exchange of conductive heat as opposed to convective? Does increased fan speed, generally speaking, increase the effect of the heat being applied - and are there any other forces counteracting that principle?

What have I done wrong?

Good questions! The Andino is the only one with the increasing fan speed after the initial phase. Why?

Another anomaly is the Shakiro natural which has a considerably lower starting fan speed than the others. If this is to slow down the initial ramp why not reduce input temperature? Something particular about a natural?

Presumably convection is higher at the start to get the roast going until bean mass temperature catches up with input temperature then needs to fall to balance with conduction otherwise the roast would cook too quickly. Except Andino?


I’d like to hear what Ikawa thinks, or possible get some of the big names being dropped to chime in to help us beginner roasters out.


Me too - for me its possibly the hardest to understand part of the process, so any help appreciated :slight_smile:

Also, I think humidity of the air plays a role (though we have not much to controll it)


Hey, try and think of the fan as the higher grit or polishing stones for the whetstones. All it’s doing is putting the shine on the bevel and smoothing out the edge for a really sharp knife. Fan will help you perfect a profile but I don’t believe it helps create the profile (besides the beginning and end). The meat is still in the curve. For instance, a decrease in fan speed to first crack with a bounce back to the end of the profile, can help give a quick burst of heat (while the air temp is still hotter than the bean temp) (how do you know bean temp? putting in a probe or guessing, but really in an air roaster the air will always be hotter than the bean temp, it’s pretty rare that it’s not.) Why does the Colombian profile have a rising fan speed? My guess is that those less dense beans are harder to transfer heat too and thus require more air flow = more removal of water vapor (which coalesce on the outside of the bean) which is in tune with my theory that since water is exiting the surface beans cook from the inside out and thus why too high a fan speed at the end leads to internal scorching.

the humidity question was answered in the first paper I posted in Fruit for Thought, the answer is that yes humidity, does effect the temp, not a lot, but a little so that the profile would require minor tweaking (or not, if the user just wants to plug and play),


:slight_smile: well … both true and not … the profile is set by the first stone, no doubt. And the angle of the profile does not change. Yet there is so much difference in what the edge is and how it feels, between like 200 grit, 1000 grit, 30000 grit … that its hard to even compare them. In a way, its a totally diferent edge.

I think its the same with the shape of the profile (what we see - inlet air temperature) and real temperature of the bean. One can be angular, spiky … but the other is allways smooth (and without equipment to measure it unknown :)) Only think I may be sure of, is that there is a gradual transition of some speed, that reacts to the curve shape.

I think the analogy that @Geoff_IKAWA used, with the gass pedal and vehicle speed, is a very good one. And what I do not like at the moment, is not being exactly sure if I am driving a Punto, Porsche, or Veyron or F1 … when we talk about acceleration. I would be guessing we are around the Porche to Veyron teritory, because air roasting and the very small batch may allow for very fast heat transfer (Does not mean we have to go fast, just, if we really push, how fast the bean really absorbs the heat).
But – the temperature curve being the gass pedal is still … quite nice way of thinking about it for me.
Now … what is the airflow … is it in a way just a modulator of the temperature curve? I believe its not … in my mind a low air flow is like a very different engine to high airflow :slight_smile: And probably, as you describe with what is may be doing to the water in the beans, its a lot more … Sure a finer controll than the Temperature, but also quite multidimensional for me now :smiley:


Hey guys, where we stand right now - we’re not sure the exact effects! We have consistently seen that a slower fan speed leads to a lower agtron number (ie: darker roast). However, we’re not sure why. We’ve been doing research into this and have disproved a few theories but haven’t actually concluded why this happens.

Like you guys mentioned, minimum is really a high enough fan speed to keep the beans circulating to create an even roast. Most profiles that IKAWA has developed have a declining fan speed because the coffees become lighter and less dense through the roast so the amount of fan speed need to keep them circulating decreases.

Will be cool to learn more about this as time goes on! I bet we can get to the point where a combination of fan speed and heat after first crack aligns with development and more control.


I feel there needs to be an ability to create gradients within the fan speed. Or rather, more drastic gradients. 8-10 total points should be more than enough. Also if you are doing research can you tell us the air to bean ratio? Volume of air transferred?


That is quite counter intuitive, I would not have expected that. Then again, I’m not really touching the fan speed save for the fact that for a profile I try to keep it as high as possible (for reasons mentioned in this thread) with no bean loss. This usually means that the speed starts at 78% and drops to 60% before cooling (@69%). Note that I roast for espresso so my roastings are longer than usual (varying from 9 to 12 min. Where the 60% is @12min. So about 65% @9min.


Maybe when the fan speed is lower there is more contact between the beans and thus more conductive heat?


So I did a test, looking at how well the beans are being circulated while roasting. with my batch of greens I added 3 dry black beans(legumes), @78% consistent fan speed I could see the black beans being circulated quite well.


An interesting thought, I’ll pass that along!


given that there is a only heat potential between the air and the beans (assuming the beans are all equally heated, they are thus of the same temperature and thus the bean-bean potential is 0) I would say that ‘conductive’ heat between the beans is an unlikely scenario.

However this did give me a thought, but it depends a bit on how the IKAWA machine actually works. I can imagine that the walls of the chamber are actually hotter then the air (edit: I mean:beans) and in that particular case there is indeed more conductive heating; not between the beans but between the beans and the chamber.


I dont think the walls can be hotter than the air, but if it has close to the air temperature than it should be hotten than the beans … :slight_smile:


Yes, I meant to say ‘beans’ instead of ‘air’. Although if the heating element is placed close to the chamber then technically the chamber can be hotter than the air.


Given…what…what magical clone beans do you have, that somehow transfer heat in the exact same way? There is an obvious difference between each individual bean and the way they adsorb and dissipate the energy built up internally. There are a lot of variables that aren’t accounted for. Judging on the face of it (pun) I do. I am a belieber in the fact that the almighty Creator of the IKAWA would not be so silly to allow such heat transeferace unless his evil plan was for obsolescence. (…because the circuits would get fried).
You should run some experiments before your conclusions. Not to sound harsh, but if you did then you would see what I’m talking about. For 1) if your fan speed is 76-60% you are obviously not going to get any fly away beans, not roasting in an almost fully convective heat…what is your moisture loss on the bean? Or how brittle is it to chew? How are you measuring the bean absorption of energy and thus, the degradation of the bean?


ok, obviously I don’t, I could have magical beans though, not something you would know as you’ve never been at my place :smiley: !
Anyway, assuming the beans are not magical, then yes there will be temperature differences between the individual beans. It is however not unreasonable to assume that the average ΔT(bean-bean) is significantly lower then ΔT(bean-air) and Δt(bean-chamber). And I do not have the experimental setup to actually check this but I think its a safe assumption that the q(bean-bean) is negligible in regard to q(bean-chamber) and q(bean-air). Hence my ‘bold’ postulation that conductive heat between the beans is unlikely to be the cause of the beans roasting more.

Δ difference
T temperature
q heat transfer

Now I’m not saying that there are not other variables that come into play, nor am I saying that my explanation is the explanation. I’m assuming we’re exchanging ideas here to explain an phenomenon we do not understand yet.

Now its my turn not to try not to sound too harsh. First of: I did not draw a conclusion! I never said I did. My literal words were: ‘an unlikely scenario’, so please before you’re actually starting to ‘up your tone of writing’ make sure what someone actually said!
Your experiment showed you that beans @78% ‘beans being circulated quite well’. IKAWA’s experiment showed that lowering the fan speed leads (consistently) to darker roasts. I was trying to give an explanation to the latter. Your experiment does not give any conclusive proof for an explanation of IKAWA’s findings nor does it for your own postulation on bean-bean conductance. You make a conjecture (as am I) as an explanation for these findings.
And unless I actually setup a lab to test, there is no experiment I can think of @home that can conclusively give an explanation for these findings (=darker roast @lower fan speed).

Back to the topic.

I’ve thought about it a bit more and another thought that crossed my mind is that when the beans are cooled at a lower fan speed that could explain this phenomena as well…however I think IKAWA was mainly talking about lowering the fan speed during roast, not during the cooling fase…


Given that we can’t conclusively prove anything without an actual lab, BUT that its also an interesting thought experiment, lets try to draw some conclusions :smiley:

Let start with listing all the potential factors (and I invite any reader to add to this list).
‘Lower Agtron’ or ‘Darkening’ or ‘Browning’ factors:
Maillard @150-200C
Caramelization @170-200C
Pyrolysis @220C+

All these browning factors are temperature related, more specifically they concern: q(something-bean) [The heat transference from ‘something’ (hot) to the (cooler) bean].
So which factors are there…:
q(air-bean) (this is the main function of the IKAWA)
q(chamber-bean) (depending on where the heat element is stored T(chamber) could actually be higher than T(air))
q(bean-bean) (given the other two q’s I assume this one negligible)

What also might play a role is the first crack, as the water then vaporizes from the bean (around ~200C). Think of higher humidity, less (or more?) internal heat conduction (k), …?

And then there is of course t(ime). Over time everything turns dark :smiley:

So we have a physics process inducing chemical reactions. My background is in physics not chemistry, so I can’t much tell about the reactions. It could for example be that if you keep the temperature for a long time at 160 you get a ‘better’ maillard reaction leading to a darker roast. However given the fact that pyrolisis of organic material tend to have an end result in black, I think its a safe assumption that the last stage (Pyrolysis) has the biggest impact on the browning of the bean.
Hence I think there is some basis that the browning is mostly influenced by exposing beans to a T > 220 high enough temperature for a long enough period of time.

Given that the premise is to find an explanation for lowering the fan speed leads to a lower agtron number (more browning) we’re looking for a process that increases q (heat transfer) from ‘outside factor’ to the bean and possibly a process that keeps the beans at a high temperature.

Now the theories:

  1. as mentioned: its possible that a lower fan speed leads to more q(bean-chamber), which could explain the browning
  2. lowering the fan speed keeps vaporized water (after first crack) longer in the chamber, resulting in a higher humidity, resulting in a higher k (heat conduction)
  3. lowering the fan speed could delay the first crack (is it? this is actually something you can test!) allowing better k in the bean internally, meaning a ‘hotter’ core meaning that the bean keeps warm longer during the cooling fase.
  4. as mentioned: k(bean-bean) is higher at lower fan speed (because they are longer in contact).

I’m still voting for 1. Not that I’m really convinced that its 1. It’s just that I’m reasonably convinced that its not 2,3, and 4. I have some arguments as to why, but I’ve rambled long enough now…


:slight_smile: cool … (or Hot? :smiley: )
Reading through, I can think of at least some experiment to check at least some factors.

My idea is, to put only as little beans inside that would form one layer both when down and when moving being pushed to the wall etc. Only as many as they dont form piles. And then, I think, independent of the fan speed, they have a simillar contact with the metal (given that even the minimal airflow means they move so they do not stay on one point)
My expectation is, that the trend to roast darker on lower fan speed would be present even in this scenario.

I do think that at least one factor in this trend has to be cooling of the outer layer of the beans by evaporation of water, which should be more present with higher fan speeds.


Well done on a civil response to a hostile post.

I don’t have a degree in science. Any thoughts I postulate are done so in a spirit of friendly co-operation as a fellow Ikawa learner. I don’t expect to be shot down in flames if I’m wrong. I don’t mind being wrong either and am happy to be corrected. But nicely eh?