Effect of Fan Speed


this certainly is turning into a ‘hot topic’ :wink:

that fully applies to me as well!

similar to what? I thought I understood what you are writing…but after reading it twice more, I’m not so sure anymore, could you elaborate?

good point, the sudden evaporation requires energy which would be drawn from the surrounding effectively cooling the bean instantly

erm…why? Even if there is no fan speed at all (which would lead to a very uneven roast) at a certain point in time when enough heat has been applied to the bean the bean will burst open from the internal pressure expunging the overheated water from the bean. ‘Fanning’ that process will ‘waft’ away the vapour faster so it doesnt ‘stick’ around but I don’t see how this would speed up the ‘expunging’ proces. But perhaps I’m misunderstanding/interpretating your words?


Me too :slight_smile: just a humble amateur here … (proud to be …amateur :smiley: )

Similar contact with metal parts between high and low airspeeds, since there always will be some contact and no layer of beans can get between …

Hmm …I am thinking along the line of, if I want something to dry I may blow on it (or use a fan) to dry it faster. Thus it seems to me, that more airflow with the same temperature should be drying it more/faster … so also cooling the outside of the bean more …


Well, as I’ve said before, I’ve come to assume, the bean temp to be about 60-100F colder than the inlet temperature. Based on sensory ques and correlations. Maybe I should say that in the beginning it is some where around 100F difference but towards the end of the roast it is closer to 60*F possibly less(?).

At a certain point (1C) then any heat in the system becomes the most influential for the final result.

At that time beans could at least be (relatively the same temperature. I do not know which would dissipate heat faster (get colder), my guess would be the beans.

I feel that the drum has less contact area than the bean to bean ratio. So at that point the influence of one bean to another is more significant. I believe low fan speed correlates to a darker roast because of the heat being transferred from the metal to the beans and from bean to bean as the air is stirring the beans, is simply to slow down and allow the beans to conduct heat from the drum to the bean mass. And increases the time of contact. Also high fan speed decreases the pressure in the system allowing for lots of volatile aromatics to vaporize. So lowering the fan also benefits the beans by not dehydrating(? the oils?) and overheating the beans at their most delicate point.

too true, my apologies

Well you should ask me then because that is/was my conclusion being retested.

You can do more than you think… You don’t need fancy equipment. Set up an experiment and run it. I simply took a recipe/profile I created which I was satisfied with the tastes/ending porosity (brittleness…sry I’m still on my 1st cup of the day) (but whatever tastes good works too). Then started tweaking one variable, looking at how it altered the beans. Limiting the variables I could by keeping a room thermometer/hygrometer, using the same weight, etc.

Hey look, In the world of science, telling someone you believe they are wrong is not “hostile”, experiments, conclusions anything and everything is under a microscope, for analysis, critiquing and rebuttals.

You should search for my posts and read them as I have already laid out A LOT (not fruit for thought) of information that I am just ending up repeating myself…


Hey you are absolutely correct, and I am sorry. I believe I sound more hostile than when I’m talking to someone in person. I can say the same thing but online it just comes off as wrong…


I hope this time I don’t come off as harsh or offending. I am simply stating what I have already come to conclude. By all means, test you theories. And test against mine, we are here to figure out how this roaster works. I sound like a jerk online, I really am sorry, its just the way it comes off. Please read this in the nicest way possible. I am going to be telling you I believe you are wrong, please do not think I am attacking you.

I think:
Dry @150 - 200 (the green to yellow is just the break down of chlorophyll (more like bore-aphyll)
Millard/enzymaic browning : 200- ?
Caramelization : (really should check out: Fruit for Thought: A Library of Resources) Joann Change video on sugar (the 2nd youtube one), it’s so good…
Pyrolysis(1st crack): dependent on the profile, water loss, momentum

All of these things are dependent on speed of heat being introduced, temp range, duration.

It’s all one of the same.

My advice:

Experimentation is fun, but also tests our own boundaries.

This is why I stressed soooo long ago, buy 5lbs (of one type of bean) of cheap(er) beans, in the 85-89pt range and try everything out. Start broad, fluctuate profiles drastically, don’t roast more than a batch or two before tasting and tossing. CUPPING IS LIFE. Store coffee’s in containers made for coffee (super critical, ultra important). Don’t change grinder settings. Failed profiles can be used as fertilizer or used to season a new grinder burrs. Underdeveloped can be roasted in the oven to *$ level and saved in the freezer for family and friends who enjoy coffee flavored cream. That’s what I do. My personal experience: I roasted like 15lbs of coffee was at my wits end, tired, fatigued, lost. Reading those things in Fruit for Thought really helped me, also for a while I took a break and switched over to tea for a while. I’ve been experimenting since I first received my roasted early July. And only got satisfying results in November(?).

Good luck on your experiments, I look forward to hearing your results. If you have any questions, I am here. Again, apologies for being offensive.


few… where to start…

@deven.patel411: I’m not offended, actually I enjoy a heated discussion as it shows some passion in the topic…as long as strong opinions are backed by strong arguments and relevant facts! Given that there’s little fact to go by I’ve been careful in my choice of words using words like ‘think’ ‘opinion’ and ‘unlikely scenario’ :wink:. You’re a bit less careful with these choice of words, making your tone a bit harsh. Again I don’t mind, and I also don’t hesitate to point this out if you put words in my mouth (or keyboard :thinking:).
By the way, you can edit your posts, so that you don’t have to create new posts after your initial one…

I just got the numbers from the interweb (again I’m physicist not a chemist) so…you could be completely right…or not…or a half…?

Very useful, thank you, liked and bookmarked, I’ll start digging!

One thing I don’t get though… you said:

but, I can’t find the spot where you’re actually telling me I’m wrong? Granted you think the temperatures differ and find the caramelization process questionable, but these are merely the background processes for which we’re trying to find the reason why ‘browning’ increases @lower fan speed?

not from my perspective :grinning:

Good advice, thank you!

Good point! so…the experiment would be 2 roasts with different fan speeds with a few beans in the roaster both times. Assuming that a lower fan speed does lead to a lower Agtron no in this experiment…does that exclusively prove that its the contact with the chamber?? I just realized that I don’t have the equipment to actually measure the Agtron number and I’m unsure whether its visible to the naked eye. Suggestions?

oof that’s a tricky one, pavel also made a remark in this regard. Always need to be careful as this can be counter intuitive. You could be right though, its just that this is not so obvious as it appears. Let me explain:
Cooling your tea by blowing over it works, because when you blow over it you ‘remove’ the hot vapour hanging above the tea. This gives room to excited water molucules on the surface to detach themselves from the surface and start ‘vapourizing’. When they do they decrease the energy of the ‘fluid-tea-system’ and thus cooling down the tea.
In IKAWA’s case, you could argue the same is happening (@deven.patel411: note that a fan in the chamber increases the pressure, not decreases it! If the pressure in the chamber would decrease then external air would be sucked in, not internal air blown out, is is the case with IKAWA!), however in this case the excitation of molecules continues to happen by the external heat source applied to it. Translation back to tea: your heating your tea with a hair-dryer by blowing on the surface of it. So on the one hand you’re heating it up by applying heat and on the other hand your cooling it by blowing away all the excited molucules. Question then is: does the tea heat up faster at low fan speed or high? And back to IKAWA: beans are NOT solely fluid, they’re ~80% solid and ~20% fluid (technically its >20%, but we’re only interested in the part that vapourizes, which is ~20%). So yes you could be completely right…or not ?!?


Yes, I understand. But I dont think about cooling by blowing. I think about drying by blowing on it - and I think from the experience with hairdrier, it works woth cold air, and works with warm or hot air (maybe faster with the hot one :)) ) … but if I want to dry the hair faster I use more airspeed on the hairdryer, not more heat … so this leads me to my suspicion that it might work the same here. With the same temperature the faster air speed might dry the beans faster, cooling them faster on the outside.

I dont think its that well visible to the naked eye - when its brown small differences are not so obvious. But my not yet fully developed process of recording a calibrated video under constant lighting, and analyzing data from that, should be better than naked eyes (though not as precise as colorimeter …)

Hmm … I have noticed this being mentioned a few times, and I am completely confused by it actually. Could someone with good understanding of it explain more?
I can imagine both … the pressure going up by blowing into the system with some resistance along the way, and also going down around the beans as the air has to travel faster around and through them then in the parts not blocked by them.


well, in case of drying your hair:

  • more airspeed will ‘catch’ the vapour faster and allowing more water molecules to vapourize, that and it simply blows away the water
  • hotter means that the watermolecules on the surface get more excited which makes them vapourize faster.
    In short: both increase the drying.
    Anyway, you could be right, its hard to tell.

Well its a bit counterinuitive as you’re inclined to think if the air ‘blows away’ that means that the pressure get lower, right?! (the correct answer is: you can’t know!)

Point is ‘airflow’ and ‘pressure’ are related but basically entirely different concepts. For example I can increase the airpressure inside a cylinder by simply pushing a (airtight) piston into it, there’s no airflow going on. Same way if I pull the piston outwards the inside pressure decreases.
Now let put a plug in the cylinder and push the piston down. That means that the pressure increases. Now i pull the plug and the pressure will thus equalize with the surroundings, effectively meaning: it blows outwards. Should I pull the piston - thus decrease pressure - then pull the plug - the air flows inwards.
In short: if there is an inequality in airpressure between two systems (e.g. a chamber and the outside world) and NO boundaries (e.g. a plug) to keep the inequality in check, then both systems equalize.
Note that ‘pressure’ is caused by two factors:

  1. the amount of molecules (the more there are the higher the pressure)
  2. the temperature of the molecules (temperature is basically a way to measure how fast molecules are moving, the faster they move, the higher the temp, the more they ‘bounce against the walls’ thus applying more outward force, thus causing a higher pressure).

Airflow is the process of air going from a high pressure area to a low pressure area. Similarly: fill one room with water (high pressure), leave a connecting room empty (low pressure), open the door between them: water flows from the ‘high pressure’ room to the ‘low pressure’ room. Now if your standing in the empty room, the water is ‘blown’ in your face when it equalizes. Would you be swimming in the water room, you would be ‘sucked’ into the empty room.

Now if you hold you hand above the exhaust of the IKAWA does it ‘blow’ or does it ‘suck’ ?


So am I, I have experimented enough to feel confident in my claims.

Not entirely wrong, and good of you to do that. Check out: https://ac.els-cdn.com/S2211601X11001131/1-s2.0-S2211601X11001131-main.pdf?_tid=2a300c8e-fd24-11e7-8506-00000aab0f02&acdnat=1516371922_b74bc0904a7b7f1c76f04b243b482be5, also found in Fruit for Thought

Browning occurs at lower fans speed at the end of the roast because at this point the element with the most thermal mass are the beans themselves and also have the most contact surface. By increasing the porosity (air within the bean and bean size), decreasing the capillary water and total water content of the bean they have attained enough energy and thermal potential at this stage (post pyrolysis) to become exothermic . The roasted bean is able to conduct heat more easily, without stripping away the volatiles as a high fan speed would do and would require more more heat in air than necessary. Also with the lack of humidity because at that inlet temperature water (I believe) would not act as a good conductor as before.

nope nope nope

Also partyly but no, as the higher temperature increases the excitement of the water, it would increase the pressure, however since the air is constantly being moved out the excited water molecules or humidity given off only acts a conductor. And the pressure remains the same because…open system.

My answer is: that the use of the fan acts differently at different parts of the roast. Adjusting the fan, whether increasing it or decreasing it affects the system. @pavel didn’t like my analogy last time :wink: but I’ll say it again. The fan speed is like a high grit polishing stone, it puts the final edge on and creates the finish. It accents the profile but meat of the profile is still in the temperature curve.


I’m truly perplexed at the stuff people choose to study! Interesting read though.

yes, that was my initial thought as well. But then I realised that this ‘only’ applies to the cooling fase (assuming the temperatured is NOT lowered during the development stage). I then assumed that ‘lower fan speed’ only applied to the roasting fase of the bean and not the cooling fase. In which case the fan speeds in the last stage are the same and form thus no explanation for this phenomenon. I should have mentioned this though!

:slightly_smiling_face: As I’ve been active on forums for a long time and like this time, when I know I’m the expert on a specific subject (to be clear ‘expert’ only applies to the ‘physics’ definitely ‘‘not’’ to the roasting!) I can get lost in a heated discussion concerning a derailed sub-topic. And I’ve learned that there is little merit in that. Hence: let’s agree to disagree on this one!

Can’t argue with that!


True as it maybe be that i am only using anecdotal evidence, however the results for me are clearly the results and I am usually looking for an explanation for why the beans resulted in it’s particular way.

This is confusing to me ( I had to read it a couple times over). Although we don’t have a bean probe we can safely assume that the bean mass temp is well below the inlet temp.(Yes?) So, even if you have a negative slope profile, the bean mass temp is still increasing unless you dip the profile below the visual temperature ques.
The reason for a lowered cooling fan speed is so the beans don’t slip out prematurely. You should adjust the fan in the cool down phase so that you can maximize the speed without beans overflowing. This this phase the higher the fan speed the better. The general reduction in fan speed during the Millard/development phases are to compensate for the moisture/weight loss so not only the beans don’t slip out but also to increase contact to each other.

You are saying that fan speed increases pressure in the chamber? This is has me curious.
https://www.captiveaire.com/manuals/airsystemdesign/designairsystems.htm, lists as fan speed increases in a system the pressure goes down. (See chapter 2 “air systems” )
Micheal Sivetz, the inventor of fluidized coffee bed roasting, no where mentions in Coffee Technology, pressure being increased though fan speed. He also used both periods of open circuit and closed circuit fluidized bed system, though he does note somethings of interest when talking about drying and talks quite extensively of pressure in the system. But it would be weird if he did because of…2) 1st law of thermodynamics…And this is an open system, air is not be recirculated so we can’t even say the density of the air is really altered in that fashion.(compression) And lastly, I do not recommend doing this but a while ago I was curious about this same question, so I put my hand over the vent on the roaster and the beans stopped rotating (before the roaster was hot, of course, and only for half a second, i don’t know how that impacts the roaster nor do i want to damage mine or your roaster in any way possible). physically it would make sense that the pressure is increased if all variables remained the same, maybe because of the cyclonic nature of the roaster also adds to it. However it is really hard to keep florals in a roast/coffee. And the temperature behind/around the fan is altering the air density drastically.

If that doesn’t satisfy you, then I guess we agree to disagree. No worries


that was written a bit quickly. A typical IKAWA fan has 3 stages:

  1. roasting fase (so thats ramping up T; first crack and development)
  2. cooling fase (so the heating element turns OFF and the fan turns UP)
  3. (after pressing the button): ejecting the bean -fase.

reflects on 2. and thus NOT on the roasting fase. IKAWA’s findings that the agtrom no is lower @lower fan speed is most likely by changing the profile during the roasting fase NOT the cooling fase. HTH.

If I have two (equally sized) chambers connected by a fan and I increase the fan speed, then indeed yes the pressure goes DOWN in one chamber, but goes UP in the other (by the same number)
I think that’s the part that is confusing you. It depends on which chamber we are talking about whether the pressure goes up or down. Now to recognize the difference between the two you can drill a hole in both chambers; where the air is sucked in, that chamber has a lower pressure and where the air is blown out, that chamber has a higher pressure.


Edit: in my previous post I said Micheal Sivetz, et al, did not mention an increase in the pressure in the system. That is false, he does talk about the increase in vapor pressure during the drying phase. Coffee Technology, Vol. I, “Drying Process” p. 119 - 126). It is also important to note that he is talking about drying wet bulbs coffee however I feel this is important information that can also give insight to how fan speed effects moisture loss during roasting (to a lesser extent) but helps understand the concept.

I dont see how this applies to what is being discussed? My original theory is talking about during the roast. It what we have been talking about the entire time, please stay on topic, I would not want at all this thread to be closed because of derailment. The OP asked:

Which I was not quoted, but the original post in question was one I made: Hacienda Esmeralda Gesha, I am not sure if that is the specific post but I’ve said it many times both here and on the kickstarter comments section…Ikawa’s experiment was also to confirm/deny my assertions. Whether the actual physio-kinetic theory is disputed, nobody has given me a a legitimate, scientific or colloquial explanation of why I am in/correct.

To get this back on track:

I just tasted a roast of my Ethiopia DP, I did one roast at A) 81% straight through B) 78% throughout C) 93%-60%-83% start, middle, end, respectively D) 95%-83%-67% start, @6:48, end E) 86% to 400-420*F (gold), decrease to 71%.

All using the same temp curve.

Results using V60, 18g, 240ml, 2:30+/-5s, @202*F (in terms of accenting blueberry flavor, chocolate and body):
A) Lot’s of blueberry, juicy (mango juice/maltodextrin type viscosity), transitioned to third day with a more aromatic blueberry, some malty flavor
B)retesting because of storage flavor. 1st day was much more blueberry than “A” in the roasted bean taste, cup was a slightly sharper blueberry, less blueberry muffin/dehydrated blueberry. More “juice” texture.
C) roast failed, way to hard, i didn’t even hear a crack, roasted bean taste/texture was “teeth shattering” and terrible.
D) Note reads: lowering fan speed did not help bean color, slow down before 1st decreases crack sound.
also no flavors worth note. Roast failed.
E) Roasted was the most blueberry flavor, I didn’t drink this cup right after roast because I thought after my errands I was going to be back home and make a second cup to keep me up all night but I ended going to a friends to help fix a car and that took the whole weekend. So I’m drinking this now and 3 days after roast (4 days total), the blueberry is faint to non existent, however the molasses has dominated the cup. Also to note, this profile and others with high fan speed for a long time, in the beginning, seems to be correlating a lot of fines that are clogging, slowing down my extraction in the V60

So far the flavors/texture I’ve been getting from this bean as matched to roasted notes so I feel confident in testing different fan speeds with the profile. Using the fan profile “E” has now done well with two different beans Ethiopia Sidamo DP, and Sumatra Raja Batak Peaberry. I would have like to test it on other coffee’s but I was not able to roast them to either match tasting notes for find other tasting notes I found to be acceptable. My next tests after I finish these coffee’s are to test a 81% to gold and then decrease to 71% and another with the same format but just starting at 78% to gold.


Ethiopia Yirgachefe Aricha Natural

you’re quite right that we’re derailing, which was one of the reasons I (tried to) cut of the pressure discussion, but it appears you’ve seen the light!

I’m wondering if @jboutte88 gets this far in reading through our ramblings :wink:. Should you (@jboutte88) read it I would suggest to read devens’ last post (the one above this one) for his great assessment on the effect/affect of fan speed on the andino profile.

I have to note though, that I’m quite surprised that the fan speed has such a big influence on the roasting. Granted if you lower the fan speed such that the beans don’t move anymore, you will most likely fail the roast, but anything beyond that I would have expected to have a (perhaps noticeable but) marginable effect.


Great work Deven. I think we should try to get to doing some serious testing and research, I still feel like there is much more potential that we are not aware of at the moment.
I will do the series of test roast at different temperatures as you suggested, as soon as I get a grip of some cheaper beans, and fix my way of shooting and calibrating to record the colours.
Btw - if you were interested in doing some colour recordings, the app I use on ios is FilmicPro …an amazing app for shooting video with very professional capabilities (showing highlihghted edges where you have it sharp like a pro camera, overwxposure zebras etc … a lot of very interesting options)
I plan to make a simple rig to have constant conditions height etc, and also get my iphone a bit distance from the hot glass as the led overheated last time … and also eliminate other light than the one from the led. Regarding calibration I am thinking of using some patterns printable on a laser printer, as the black should be a +/- simillar and also a paper white printer paper is quite simillar … so it would be easy for anyone to print it and get +/- calibrated in the brightness range … the colours would be constant only on the same phone, but that does not matter too much I think, does not matter if the yellow has a bit more R or G you still see where it happened. And being able to add temperature to those colours will be great :smiley:


Another correction: Fan profile “E” does well initially but moves very quickly (day 2 and day 3) to Millard flavors, so good if you like the Millard flavor, bad for keeping those lighter notes we associate with light roasts.


Very interesting! I will also get the team doing some more experimentation here on fan speeds. I’m still getting up to speed on the conversation and what you’re discussing. Is it regarding the cooldown fan speed or fan speed during the roast?


I think mostly we discuss the fan speed during the roast, but also Deven mentioned some effects during the cooling phase.


Ok that’s what I thought, and that’s when we have found a difference. The profile that deven shared had what looked like an adjusted cooling phase fan speed.
I’m looking forward to gaining more knowledge of how the fan speed affects the roast and how we can use it to dial in a roast better


Haha, yes rsegers, I’ve been following, and I did find Devens post useful! I’ve been working a lot on Ethiopian naturals too. Here’s my new favorite approach to fan speed on them.