The S-curve profile


Hey @deven.patel411 :wink: when visiting my parents during christmas, I have shown the profiles I have in the app, including your A profiles, to my father (who is a biochemist, with some deep,understanding of whats really happening ). And I have to say he very much liked thenshape of you A profile … he also does not like going straight for high temperature at the beginning.
He does have some studies about reaction thatnhappen during the roast, so maybe one day I will have some intensive course on the chemistty there ;))


Hey it’s great to hear some positive feedback. And I studied biotechnology and biomedical science for my undergrad. So there is a lot of science behind why I do what I do. But part of science and being a scientist is being able to explain and prove the law so (to others) testing your hypothesis and being able to explain the “why” if it is good or not is really what helps the progression. Darwinism at its finest.


Deven, as much as I’d love to have more theoretical tools in my belt, I’ll be looking forward to trying your profile and variants. I have done roasts with yours but I am on vacation and those samples are waiting for me at home. I’m definitely excited to be trying different stuff! I’m still trying to wrap my head around the difference in approach between you and say, shakiso profile. Did you roast the ikawa beans?


I think you would have a lot to talk about with him :slight_smile: very close fields I guess. :wink:


I guess to put my words where my mouth is, being as clear as possible. The reason why I like the S curve is because its seems more gentle on the beans. When I roasted the Guatemalan huehuetenango, all these really interesting floral and clear candy tangerine came out. While on a profile which shot to 448*F in the first minute and sustained for about two. I got so fixated try all kinds of profiles to capture these in the end cupping, but really I burned through all 5lbs with a great smelling house and no awesome cups. The “S” shape, really was the only cup that was able to pull a lot from all the profile types. The best I got was passion fruit and some chocolaty back on the cool cup, i’d say a 5/10 on sweetness but the key was that there were no defects. I think that is the key for all profiles NO DEFECTS. Honestly, I don’t believe it when some profiles are displayed and nobody says I got this or that. No profile is perfect, And we are contractually obliged to be nice but we can be honest without being mean. So saying,
“well that profile A wasn’t that even in color” I’d be like totally correct, I can see what you are saying…

Seeing clear distinct profile changes, are a big one for me. When I was running through faster profiles, even just finishing up on the last of my Sumatran (shortening it up instead of trying to amplify the chocolate which is what I was trying to save the last 10 roasts for) I wanted to make sure I was absolutely correct when I say it’s harder to SMELL (especially in the Millard/Carmelization phase (2nd to last) the real sweet (drop) spot. I on three of the roasts I totally missed the strawberry poptart and just got heavey vanilla/marshmallow and slight wood, which I’ve tried to avoid and found that if I convert too much over by going too fast too hard, I get lots of mesquite wood in this Sumatran which I know can do better than wood and vanilla with some dry cacao.

So back to the S profile…do you eat some roasted beans after you roast? Do you try one right after the cooling phase is done? Compared to when the roast is completely room temperature? Do you notice how tough the bean is? Is it easier to get more flavor quickly with a more brittle bean than a tough bean? Have you tried a roasted bean that you could taste a distinct flavor with and not something general like “coffee”? Or maybe a distinct flavor to the bean that one might call “origin” flavor?

For fast profiles my conclusion is this. You can take a long profile and shorten it. But going the other way around is much harder. Even a broken clock is right twice a day…

Dropping/cooling is an arbitrary process, nobody anywhere I have ever read has said they intuitively know where the drop temp is off the bat. There is a general idea people shoot for and then tweak and hone in on but…where to end the roast…is up to you. And how to find that end (temp/time) (for me) is throwing in till dark evaluating the bean by smelling and looking at the bean really smells good and marking points and making notes.

A gentle start:
Some beans can tolerate more heat in the beginning, or are tougher to get warm/dry but then they take off like a firework. But How much heat is too much? Each bean has it’s attributes that give an insight on how to approach it. I have no idea what that approach is. I get a ton of weight and just start throwing profiles until something sticks…But when cupping the defect of the roast becomes very apparent. Internal sourcing is kind of hard to see but burnt taste in the cup is very apparent. That’s why you can drink your coffee how you like but the sugar and cream are going to mask the coffee in a way. If you are evaluating a bean/profile you need to cup it. Sometimes I just drink my cupping, when it’s good it’s good. So back to a gentle start…There is a range in the first minute of a profile I like to make my 1st pt., so far and that’s 270-300*F, which helps become the start to get the beans from pale to yellow. I do not believe that there is not any flavor changes that happen from pale to yellow. Can you believe that there isn’t, sure. I believe the former.


I was given a 60g sample of the Burundi MHill that I used the Ikawa set profile for. Doing micro cuppings I determined that it was underdeveloped with tomato throughout and not exiting with time. Could be my water…


Regarding the profiles, I think its not as simple, as modifying one optimal profile for all needs. (As much as I would love it to be like this ;)) I think - just my own intuitive understanding for now - that any shape of profile can be valid for some use … it may be the bean or need to raise some tones in the coffee, that can work with a profile that might be “wrong” but still might bring enough possitive that its negatives might be overpowered for that situation.
Also, and that is something I coming back to, again and again … the profile is not what is happening to the bean, and I am not sure if even a probe in the roasting beans would help much - there is still a hot air circulating through, so I dont think one would be really measuring the bean temperature there.
I think whats is really needed is a good understanding of those cues, colors and odors, and their meaning in terms of temperatures and possibly even chemistry - reactions - connected with those cues. As Deven mentioned somewhere else, there are temperature ranges connected with some reactions, so one might know quite well what temperature is the bean at by understanding the cues and products of those reactions.
But that, thats a long distance run for me.


Hmmm … have to say, I am really interested in how my view will change after some 1000 batches roasted. Or 10k … so much to explore :smiley:


Happy New Year all! I haven’t had a chance to read through everything that’s been discussed through the holidays (and there is a lot written here) but I thought I would quickly put in a quick note here for some newer home roasters.

Regarding the S curve, which is the typical graph you will see if you search “roast profile” on Google, one point that @jboutte88 mentioned and I thought I might help clarify is that shape is normally associated and discussed with BEAN temperature. The IKAWA At Home measures the INLET temperature so the shape of the graph will not always look like an S. Roasting on the IKAWA At Home is fairly unique.

If you’re new to roasting, the bean temperature was traditionally the main or only method of tracing a roast. On all the commercial roasters I’ve used, a temperature probe is placed where it sits in the coffee bed while it’s roasting and measurers the temperature of the beans as they roast. That can produce a graph that looks like this:

On the Loring Smart Roast, a kind of hybrid drum/air roaster, hot air is pushed into the drum rather than the drum heated from below by a flame. The Loring actually measures the temperature of the beans, the inlet temperature, the return air temperature (the temperature of the air exiting the drum) and the temperature of the exhaust from the external burner. I’m using the Loring as an example here because I have graphs from actual roasts that show the INLET and BEAN temperatures.

The INLET temperature graph from Loring roasts overlays similar to this, although can vary:

I’m showing this graph only because it illustrates HOW the S curve can be achieved through heat application. I don’t have a graph for how the bean temperature charts during a roast on the IKAWA (although I am hoping to one day!) so I thought this would be a good starting point.

NB: I have tried roasts on the IKAWA that model the inlet on the Loring and it doesn’t work. Roast come out underdeveloped. This graph of the inlet temp is more flat than an IKAWA roast because the Loring itself retains a lot of heat in the metal and the beans are dropped into the roaster at a very high temperature. The IKAWA begins from room temperature and compared with the Loring (which has thicker, hotter metal surroundings), almost all of the heat the beans will absorb is from the air so is quite different. The IKAWA roasting chamber needs some time to be brought up in temperature.

2 other quick notes about the S curve from the bean temperature (and why it can be misleading to use the graphs that you might find online as a guide).

The first is that the beginning of the graph isn’t accurate to what’s happening to the beans. The temperature of the beans is going up (how could it not), but the readout is dropping in temperature. The bean temperature probe, right before the roast begins, is reading an empty, hot, roaster and measures the “drop” temperature (the temperature that the inside of the roaster is at when the beans are added at the beginning of the roast). Of course, the green beans are room temperature so they cool the temperature probe down.

This brings me to my second point - almost all the bean temperature graphs will all be different because of the probes.

Because the bean probe is sitting in the bed of beans, the probe needs to be robust and strong and is therefore quite thick. This build quality affects how long it takes to read a new temperature - that time it takes depends on the type of probe you’re using. Different probes have different lag times and will totally change the shape of the graph.

Presumably, the graph for INLET and BEAN on the IKAWA At Home might look like this (with a few different INLET shapes shown):

This has become a much longer post than I set out for! Maybe I should put it on the blog. One final thought that may help, regarding green bean temperature and inlet temperature during the roast is to think of driving a car. The speedometer is the bean temperature and the accelerator pedal is the inlet temperature control (our profiles, in this case). The relationship between the speed of the car and how hard you’re pushing on the pedal are not simply this than that. Just because you’re pushing hard on the pedal, doesn’t mean you’re going fast. You can be pushing hard to accelerate quickly but still just be at 30km/h. ¯_(ツ)_/¯ I think that’s relevant here…

Anyway, hope that helps some of you. Would be really great to hear how some of you are building correlations between inlet and bean temp!

What is fast? What is slow? ROR Through the Roast

I’ve found an average difference of 65-100*F difference, is that also your experience?

To use your analogy, you may not see immediate results but that is with drums, the IHR (I’ve found) is quite responsive. If I could call it a car I would call it a new released Honda civic with a v6, or the 240hp ford focus. Enough power to feel like your going somewhere but not enough to loose control. But crashing is always an option lol

The reason the profile A is shaped that way is because this roaster seems like it carries the beans like a puppeteer. So the line is very specific to where I think at that inlet temp is going to be in relation to bean temp. Based on approximation, bean color, and smells. My theory rests on the over arching profiles don’t accurately(to some degree) what temp the beans are transitioning with, aka flying blind.

I’m going to do a ‘mixing’ test with different colored dried beans because I’m very curious now as to how much the walls play in heating up the bean.
If anyone has a thermocouple they can put in the beans (that is safe to do so) and get a bean reading that would be cool.

@Geoff_IKAWA what correlations have you found so far? Also, thanks for the drum synopsis!

Edit: when I say beans I mean actual legumes like black beans and navy beans and not coffee beans.


Exactly - thats what I have been pointing at when I said we dont know much about the actual bean temperature. Its also why I refuse for myself to think about RoR (because only one that can be known is the RoR of inlet air - but it does not say much about how the beans accelerate :)) )

Which also makes me ask another question that I am thinking about very hard – is there another way that the bean temperature can be measured, one that would not need a probe inside the beans? (That would change how they move in the roaster, and also has that lag you mentioned). I have looked at IR Imaging cameras but they mostly can not measure through the glass as the wavelength gets absorbed. But still … having a fast reacting non intrusive measuring device would be great to really see how fast the beans absorb the heat, and also see the points where the reactions get exothermic and so on … so …
Also If I understand it well if the glass was the special kind that allows the wavelength through it could then be measured from outside through the glass. Which may be best to protect the IR camera, and I also plan to use regular RGB camera with a LED lighting to try some magic of postprocessing and averaging and get some analysis of colour changes this way (I do image manipulation and postprocessing, VFX … for living … so … its just easy for me to think about some ways of processing the video to get some insight :)) )


Do you mean a difference between bean and inlet temp? I haven’t had a chance to measure the difference in temperature because we haven’t measured the bean temp on a roast yet. How did you come up with 65-100F?

Yes, compared to drum roasters the IKAWA is very responsive. Because of this I try and imagine what the environmental temperature is for the beans when making adjustments to the curves. Overall, I’ve found that a steady, smooth environmental temperature works best for almost all coffees. Any sudden changes seem to detract from the best possible outcome. A sudden inlet temp may not matter too much for the beans if the environmental temp is changing slowly, smoothly.

@pavel if you can figure out a way to measure bean temp, I’d love to see the graphs! Your idea is beyond my knowledge so I can’t comment but I’m hopeful you’re on to something


Yes I do mean the difference between bean and inlet temp. How I came up with with numbers was based on where color changes occurred compared to the inlet temp readings. The large difference is for two reasons, 1) accurately pinpointing bean temp is impossible with my current set up. 2) A drastic increase or decrease will cause rubber-banding effects so the gap will widen at times.

I disagree, I think environmental temp does not exist in this roaster. Since the beans are in direct contact (and fully encompassed) with the heating sources. The inlet temp is essentially the environmental temp and reading a temp at the top of the cyclone vs the bottom would just be taking the differential of the bean mass. In which case might as well put the prob in the bean mass via a protruding probe (what I would want is a IR sensor) in the bean area.

Spikes have their place, I don’t disregard that. :wink: Sometimes in chemistry in order to break thresholds energy is introduced in the system. However at this time I do not lots of accurate, repeatable info to support this. So I agree…for now. But environmental temp is always going to be smooth, I mean really, air heating a probe, there is no way it’s not going to be smooth…

there are literal tons of ways to measure temp, in the exhaust temp thread I linked a site that had IR temp probes, fixed probes, long probes short probs. Although I like you energy @pavel that science is not new and there quite a lot of info on it…money is the factor…and isn’t the top made of acrylic and not glass?


Thanks, Geoff_Ikawa for the Interesting post! I was definitely surprised to see how mellow the projected theoretical S-curves of the bean temp were with those relatively high inlet temperatures.

Pavel, btw, I appreciate your perspective on the mystery of the bean temp and and RoR. I used to roast with a thermometer sticking into my popper chamber which I could adjust so that it was in the beans or the exhaust. Neither of those really tells you what’s going on inside the bean though as I believe you’ve pointed out. Temperature measurement is always an indirect clue about what’s going on. Still I’d love to get more data too. I wonder if there’s any way to use a laser thermometer? I can’t think of one without a direct path to the beans…kinda wish


It is interesting, isn’t it. The combination of the pre-heated metal (drum, casing, paddles, etc) and the really large batch size (30kg) make it less responsive than you see on the IKAWA. Although, compared to most commercial roasters, the Loring is very responsive.


Ah, its possible, I did not think of it as not being glass but on second look, its not heavy and cold as a glass should be, so … you are probably right.
But acrylic is not transparent for the wavelength thats needed for IR also… so … all the same.
But, but … if its really acrylic, it might be worth a try to buy second cover, make a hole, and find some IR camera that can withstand the contant with the heat :smiley: might work that way.

Sure its all well known to those who work with these instruments, but all very new for me … yet I dont mind to investigate allready known, as I also learn in the process :smiley: So far measuring the colour changes seems more important to me than meaasuring the temperature … and I have an IDEA :smiley: … will test it very soon … hold on tight … :slight_smile:


Sounds good, I think our quest to measure temp is futile (I mean by trying to get the “actual bean temp”) where a temp input is already here, and we can make a coorelation based on that. However if you can get a visual thermograph(?) then I think that will be a very nice insight. To a lot of questions I have.


Hey pavel, sounds very cool, hope you can get that to work! Quick note though - the top is made of glass. I believe the same type used in oven doors. Hopefully that’s the better of the options you need :slight_smile:


Allready working on it… I will post it in my Rwanda thread in a moment!

not in a moment as it seems - it takes about an hour to render the video … :smiley: but soon.

edit: And There It IS :slight_smile:


It appears I’m really behind on my reading. There’s some really really interesting posts here on the forum that I’ve missed so far. Thank you @Geoff_IKAWA for this elaborate and insightful post.