The S-curve profile


I don’t know much about it except that it is sometimes referenced on forums and such, and one of the top google hits on roasting advice is all about the “S” curve. I surmise that the bottom tail of the “S” is a gentle rise before a steeper rise and then of course the slowing at the end is the top part of the “S.” But that’s just a guess. Anyone know more? For instance, I have often thought the steepest RoR should be right at the very beginning, but there’s probably a lot of flexibilty there. Then again, maybe every profile lags a bit at the very beginning compared to a minute in, so maybe it’s unavoidable. Should we embrace it? Are there benefits to it? Is it even worth talking about since it’s based on bean temp and not input temp?
deven.patel411- you made me think about this profile shape with some of your profiles. Any thoughts?

What have I done wrong?

One of the main things about the S curve is that it is a drum profile. And the dip at the start is all about drum technology. The bit which applies to air/fluid bed roasting is the rise and easing off at the end. You see that on quite a few profiles here.


All that information really has already been written about. Google only works if you can ask the right question. My journey led/went through the back catalogue of’s roasting forums and sweet maria’s library and others (those were just on the top of my head) not focusing on the numbers because temp and duration are specific to each roaster, even drum to drum. We are here to see if this roaster follows the norm or can actually make transferable profiles, which I believe yes it can. Getting back on track…so focusing…on how they roast, what beans like, generalities really that can be extrapolated, compared/contrasted and reformed to fit our roaster.

That is something you should test for yourself rather than taking my word for it. There is no better teacher than experience.


Time will tell.

Correlations my friend, correlations.

I have lots of thoughts, constantly. But non I am willing to share at the moment, some inputs need to be flushed out and put to rest. I need you to come to your own conclusions first before or be proven wrong, I am infallible. Throwing out theories and hypothetical are nice but creating profiles people can use, work with and enjoy are what makes the difference.


Dormouse - that is a highly relevant point that I did not consider while writing. That said, something has to happen at the beginning of every roast - either the steepest RoR or not. It’s surely possible to create this type of profile shape on the ikawa, and I sort of suspect it at play in the Guatemala and Brazil Ikawa profiles. Even the fast hot profiles take at least 35 seconds to get up up to their first point which makes me think there must be at least a few seconds of lagging RoR before the highest RoR. Perhaps I’ll try an experiment like moving the first point on shakiso all the way back to like 4 seconds, and then moving all the other points back the same amount and see what happens. It seems to me that even in the super convective cyclone style roaster there may be some cause for a “soak” period as the drum roasters call it. Perhaps it has to do with managing moisture content in some way - like making it more consistent across all beans…

In general, I’m curious about the earliest stages and their impact. For instance, how might extended yellowing or drying affect the roast? Is it just about setting the “tone,” or the momentum, for a roast and thereby making it easier to control in later stages, or is there something more fundamental and chemical about those early processes? The guy from the Mill City Roasters quoted some study which stated it makes no difference how long you take to get to pale, since there are no chemical reactions yet. Does that ring true to you guys?


I’m a sucker for confusing, impractical, and ultimately useless theories (philosophy major :wink:), so if anyone has any, I’m totally game! We’re all just getting started here, so I say no bad ideas in brainstorming! I don’t mind being misled if I might discover something that doesn’t work - I find that very helpful in the quest to master a complex skill.


I haven’t fully flushed out how pale/yellow has an effect of the ending flavor. I do feel that it does impact cracks especially in the highly convective medium. It is hard to tell for me (and understand) because when changing the time of that 2nd stage of the profile “A” like shortening it. I am not sure if I should keep the ending time the same or decrease the total time. Either way there is an amount of energy in the system that is lost or gained. That fluctuation has to be accounted for and worked around, or not, again not to sure.

What I have flushed out is that initial ramp, my advocacy is that going to high and to fast is going to produce tipping and internal scorching. And there is a sweet spot in the 270F-300F/127C-139C within the first minute or two, so far this has only applied to a 2lbs Barundi, 2lbs Sulawesi,1lb Indian (another 1lb Indian peaberry), 1 Sumatran (5lbs Sumatran Peaberry), 7lbs Guatemalan. Not all weight but a majority was done in the Ikawa. .

Yes the A profile is based on the theories of the S curve profile adapting to the bean and the roaster. And not a sivetz or air/fb roasting profile which I have yet to find one that can taste better the than the S-style. To me the “ln arc” profiles are good for beans that are new and you don’t know to roast, but is hard because even though you can time phase transitions and stop based on sound and smells, its hard to create a specific profile unless you get so familiar starting with that profile that you will know what can correlate to reform into S shape standard. Which is difficult in the beginning but ultimately ends up (for me) as my typical A profile when I go for a concave positive curve.

I am fine with saying that is a good way to look at it because I do not have any information to base my theories on yet. Maybe in a year or so I will.

I mean of course that’s great, but do you see why I’ve been advocating for people to go out and read?? It’s not exactly fun to just put out my hard work, theories and conclusions. Without anybody to discuss/analyze and subjectively retort with, I’m simply explaining why I do things and it’s kind of secluding. For the most part, I would like to discuss and not explain…


Well, sure, but there’s no reason to assume people haven’t already read and researched, especially in this forum! Please point me to anything I haven’t read or watched, And I will happily devour it, but I have read or watched everything I’ve seen you reference already at least once. I’ve read the classic papers enough times to know there’s not much more I can get. The scope of actual science is incredibly small (although growing - but where are the scientific papers on roasting from after 1980!?). The actual research I’ve seen isn’t really interested in producing good tasting coffee, but rather trying to isolate one variable and measure the physical and chemical effects of manipulating it. With good reason, they don’t try offer a subjective flavor or enjoyment evaluation. And that leaves us with experience and educated speculation. But you’re right that up until now we’ve mostly just been putting out explanations rather than having conversations, but we have to start somewhere. On that note, why do you think it matters to have observably distinct stages? I don’t at the moment have any evidence that it matters if the stages of the roast are clear and well-defined. In fact, trying to make that happen, and succeeding, has produced some of my worst tasting cups. I’ve had roasts with no clearly discernible phases - all of them overlapping and sometimes seeming to happen at once, that were positively delicious. Is there any way to keep all else equal, but make the stages more distinct? Sure, just take a profile and stretch it out, but that’s the exact approach that has failed me. It’s also worked sometimes. Maybe distinct stages are just a side effect of some larger approach? Maybe the slow and steady approach you’ve referenced is connected to those clear sensory cues, but do you think those cues are actually integral to the roast? Perhaps they are for the type of roast you prefer, but not necessarily the ones I prefer. These types of subjective back-and-fourths could go on forever even with all types of profiles and coffees to try, but without some underlying theory it would be impossible to really understand those differences. That said, I think a year may be an optimistic estimate for learning anything objectively about coffee! In the meantime, I say speculate and theorize freely. The more info the better, even if it’s mostly wrong. That’s why it’s so useful to have multiple people within the same machine!


I’ve only had my roaster since the start of the month and gone through the supplied coffees on their profiles.
Now it is stage 2 experimenting to find out what happens if… I don’t have enough data to speculate theories
So you guys are way down the line and it is good to read your initial speculations and findings even though at this stage I feel I can’t contribute much to the discussion. Need to bone up on the theory!

Out of interest, approximately how many roasts have you logged to date with the Ikawa?


I mean, that’s exactly why it’s not as easy as just taking a stage and stretching it out you aren’t going to get something that tastes good…like I mentioned before there is still an amount of energy that the beans did not get enough of (if you shorten the roast) or too much of (if you stretch it out) and that inturn affects the momentum of the roast. I believe Rao was the one who said to tack it on at the end, if your cut the previous stage short. But are we going to actually calculate the precise energy in this system to that extent, I can but do that if I wanted to, I suck at math. And if I start listing stuff in Joules I think @Geoff_IKAWA might loose his big beautiful hair.

ok I will compile a list and post it, I’ll put in the read me I haven’t yet continued lol

proprietary patents, people want to keep making money of course. A small amount of information is an understatement. Go to Springer and buy a subscription, or maybe just scroll through the list of books and journal articles. There’s more to asking google about roasting coffee, there are tannin, phenols, acids and lots of specific chemicals that are precisely measured, occurring in succinct order…if you want to see the ocean, I can’t even understand half the stuff without having 9 wiki pages open. IMO not worth it but it’s there.

Because I was struck by lightning and spoke to God, I saw the light at the end of the tunnel. Why do think I’m the half naked mad man screaming that the world is ending.

@pavel will get a kick out of that, I “produced” one that tasted like soy sauce umami bomb with WAY too much salt. I used it in a chicken teriyaki I was making that night. Can’t make an omlette without cracking a few.

Those que’s ARE the roast, I can’t be any more clear without accidentally sounding sarcastic.

I am a hedonist there is so little I don’t prefer. Maybe old folgers that was brewed 2 days ago that my family still thinks is good.

I think when most people think of subjective taste they just haven’t yet discovered it yet, whether it is cigars/tobacco, wine, tea, coffee, beer, liquors. When I say pineapple, you taste pineapple, there is no subjectivity philosophizing what is a pineapple. If you don’t get pineapple and you get something else, that is not wrong or bad, but chances are a lot of people will taste the same thing. Why else is there a purpose to tasting notes, or building a roaster that should roast exactly the same. But there are still an insane amount of variables (like water) so it’s not perfect either.
Even for a drum roaster the consensus is that alot of beans initially wasted. Hence the “paying the tuition” even with people saying you should have this exact percentage here, here and here. So there is a lot of failing regardless especially in the beginning. And I personally didn’t regress back to an easier profile just because it tasted good, I kept going to see what I could get out of it. That is my suggestions, keep chugging through. Distinctive phases are where it’s at. Some people don’t want to scare you away and say here, it’s easy and I’m like whaaaat no it isn’t. Sorry for assuming but I thought everyone who purchased on the kickstarter knew what they were getting into. Roasting coffee IS easy, roasting great tasting coffee is not.

@stephen.pickering21, hey I feel for you, but there are a lot of people just not in it for the team, I don’t believe that it was just Mike, Christo, Pavel and I who got a roaster. It those people who I’m like: hey, what have you been doing? What are your experiences like? How are you roasting?
So I’m not trying to throw any shade your way. I’m glad you have initiated stage 2 and very interested to hear: What are your initial feelings on approaching a roast? What you are roasting? And just your general feel for the roaster.

I shudder to think. Lets just pull the calculator out for a second 5 L b’s @ 50 G’s a pop is 45 roasts so given what I listed so far around 180, I started to count profiles and stopped at 50 but i’d say somewhere around 130 or so after the Great Deletion of October 2017


@jboutte88 Fruit for Thought: A Library of Resources


The type of roasting I’m used to manually is applying a steady heat to 1c then a bit of development 1 to 2 min. So the Brazil profile is the one I’ve more or less been using prior to getting the Ikawa. I found with heat gun that most coffees are ok with this. Prob 75% development. Roasts were generally 11 to 12 min for 80/90g.

So what I’m now trying to understand is the other profiles. I think it will be instructive to see how the beans respond to these other profiles and get a feel for how origin/process tie in to these before starting to hone in and edit profiles with specific goals.

As far a the “S” curve goes the Guat looks closest to this. So i’m asking myself …Why does the Guat fit with this? Is it just this particular Guat? Is it a good starting point for all Guats? Just roasted a Columbian under this profile, not tasted it yet but it didn’t get much into 1c, so maybe extend the roast 30s next time etc…

So I guess I’m trying to dig out some generalities (if they exist) to gain an understanding of good starting points before specifics.


Gee … we really do need to use like 5Kg of the SAME beans … sharing profiles, doing crazy stuff with it … but its just … at least from my POV … necessary to have the same material at least for some time. There has to be a coffee … tha can be sourced from local suppliers on every continent, yet can be tracked to the same source and harvest … because ordering it so that it would travel around the world is stupid. But it would just be so much fun …


Yes. Same beans makes sense if possible


The hardest part for me going from manually adjusting a mini oven to this is trying to set the curve up. I’d say stick with one profile before you start looking elsewhere. A profile can roast any bean, its how you adjust the profile and it’s result is the truth to understanding the roaster. I don’t think the people who made the profiles spent a lot of time making them, they seem really vague but specific to one type of bean, from my point of view. It would be nice to hear from the people who made the profiles, to tell me i’m talking out of my a**.

I think that is the right track of questioning. IMO 30s especially from the browning onward is a LOT of time, try increasing the rate aka the slope of your Browning/MR stage, A faster climb i’ve found will also make louder pops. Did you see when time/temp the beans expanded(loss of chaff too)? And I have a feeling that even though I very much try to get a clear separation of 1st/2nd crack they happen so much quicker than from what I’ve read on the drum forum. Like just around 1min if it’s a good roast and at the same time if it’s not a good roast (IMO). A columbian without much into 1C is going to be not what I think you want your Columbian to be lol. A bit on the tart side? no nuts or chocolate maybe(?). I’m sure everyone has a preference of bean but I believe nobody has a preference different from each other on how to roast it. So if you like acidity your are going to go for those African beans, but not really Columbian. And mixing them (after roasted separately, unless the play nice on the same profile) can produce a nice mix for sure. For instance a clearly well roasted Ethiopian will have all the acidity and fruits a sour head will want, there is no need to try and turn a Mexican bean into a tart bomb when it wants to be a dark plum and chocolate.

Variables like 1) varietal (NOT region, varietal), 2)processing, 3)density ,4) altitude (of where it was grown),and if you want to get more specific, weather/rainfall of the season. Are the real playing rules. Maybe 2 and 3 are switched, not sure. Reading the HB forum you tend too see those “generalities” of what beans are like and do what in the roaster as people discuss their issues, and getting answers from users with literally 1 million times more experience and knowledge than me.

@pavel …that’s a question you should ask @Geoff_IKAWA and @Alex_IKAWA, no person or company is more poised or in a better position to accomplish that. However at 20 pounds, not dollars, for 1lb of green coffee. That is not doable even for the 1%, ok well maybe there are 3 backers that could but really I’m looking for an asking price of $4.50 maybe $5.50 (that’s in dollars) with an included $0.50/lb (labor) fee, and whatever shipping is, so $6/lb which can easily be available for 89-93pt coffee’s and still could be found cheaper in the $5.25-4.25lb range.


You know I thought they migh have a solution for this, but the problem is shipping to some distant places is not a reasonable think to do. (If it was, i can also have a bag of 60kg with ease, split it and send in all directions. But the price for 5kg sent to usa or australia is not sensible.)
I started thinking in another direction. Everyone has some source of,green coffee in reach. And they tend to have some notorious ones, that should be the same around the globe … if there is enough info about origin we shouod be able to be sure we have the same one …


yea that’s def cool too


Looking at the Ikawa beans only the Brazilian is a single varietal. The Guat is a bourbon and caturra mix. The other bean with bourbon is the Burundi with a very different profile.
Generalities might be hard to find…


Now that you mention it, Deven, I do notice some physical changes in the beans as they roast :roll_eyes:! The point I could’ve made clearer is that chasing sensory cues is arbitrary. It is intuitively and sensorily appealing to make the stages distinct, but it’s not chemically necessary to achieve an optimal roast.

So, the question remains, what specifically is it about this type of profile that you like? (Besides the revelations from God that seem to accompany it.) The main thing I’ve noticed from reducing the heat at the beginning is, as you say, needing to add more at the end. Rao generally discourages cranking it at the end to make up for less at the beginning btw, although there are always exceptions in reality. I’d be curious to know more specifically what you consider to be the benefits of this approach - are there any of those distinct notes you’ve been mentioning? Texture, flavor, acidity? Ive been trying S- curves from the beginning, and I’ve noticed some lovely maillard qualities like chocolate from this approach, but the acidity is usually underdeveloped. Perhaps your coffees are low acidity? Mine are extremely acidic, and I simply haven’t been able to break down the acid in those profiles that are gentle at the beginning.

You say not using this approach has led you to defects like tipping and scorching, but it has not usually done so for me. Are you sure the defects are originating from heat effects in that part of the roast? I’ve been able to eliminate those by increasing air speed in one case. Which other profile types have you tried - Andino, shakiso types? Have you roasted the ikawa coffees on the set profiles? Do those have the defects too!?

I don’t think the unique chemical signature of a food or drink is subjective, of course! I wasn’t suggesting that pineapples taste like tomatoes to half of us, just that some people may not like pineapples as much as others. I believe that the vast majority of “quality” in coffee is objective, up until the point it runs into subjective preferences. It will be hard to share useful info unless we can explain what we are getting from our profiles. So for me, the s-curve approach has only worked for coffees with acidity that requires less time and heat to taste not-green. In general the shakiso type approach has given me the best fruit and floral qualities I seek. Which qualities are you seeking in your finished roasts?


Im glad you’re looking for generalities too! I can tell you that the Guat profile is NOT a good starting point for any Guat, although you can get a passable roast from raising or extending pretty much any profile. I believe that the s profile works best when the acidity needs little breakdown/development which is not a feature specific to any origin except maybe lower altitudes. I’ve roasted about 50 pounds so far with well over 100 different profiles!


Now that you mention it, Deven, I do notice some physical changes in the beans as they roast :roll_eyes:! The point I could’ve made clearer is that chasing sensory cues is arbitrary. It is intuitively and sensorily appealing to make the stages distinct, but it’s not chemically necessary to achieve an optimal roast.

Yes, I understand your sarcam, but what I’m getting at is the nuance of the phase changes, how clearly/distinctly do they they change from on to another and how evenly (in bean color) the bean is roasted. Also please don’t forget that yes although we have been roasting and are familiar with the topic, some of us are not so it’s best (for now) to lay out things in clear terms for those lurking.

“but it’s not chemically necessary to achieve an optimal roast”, sure but show me a sources?

Have you tried my “A” profile? And try adjusting it?

1C is around 472*F/7:45 roughly so stop it whenever you feel like, if you like the acid, maybe 8:15-8:30min mark

tasting is believing, that is all.

Forgive me it’s is hard to keep track of who said what but, remember these guys (Rao, Hoos, et al) are talking about a drum roaster that carries a lot more momentum and works like a slingshot rather than a electric car. Its really about the torque and stopping power.

I wouldn’t mention them if they weren’t there. I have no gain from lying to you guys. My motives are selfish, I want a bunch of people who can teach me to do stuff too. In general I pick up things very quickly, try and teach others so they can learn new stuff so I can chill and be lazy. We all have out qualities and attributes.

I’d be very surprised if you used my A profile and described the acidity as underdeveloped. Underdeveloped acid is just going to taste like literal citric and malic (or other) acids, I cook a lot and have both and pure concentrated form, Try and get some if you’ve never tasted them. Developed acid is fruit, fully developed acid is ripe-overripe fruit/cooked fruit, refer to Hoos on that one. (I can’t find my book or i’d quote him)
If you’ve been doing S-curves from the beginning my thinking is that you make think the are “underdeveloped” but you are just roasting them out. Unless the things i stated previously happened.

Going to 400*F in the initial ramp has definitely led to internal scorching and tipping, very noticeable while cupping, and cutting the bean and looking. I don’t know what your profiles look like, nor the beans you used them with. I’m not writing down what I say in stone (my chisel is dull) just my simple observations.

I’ve previously linked to sweet maria’s on cupping protocols and to SCAA’s, that is what I go by, the standard. I also linked one other in Fruit for Thought topic, which I thought was a nice observation of the coffee flavor wheel.

The S shape can be interpreted in a Platonic form, what works for it or not is simply what you can do with it. I look to achieve the most flavor and texture I can possible get out of the cup. I have no preference of tastes. I like a developed cup, with distinguishable flavors. Whatever those flavors the bean has to offer i’d like to capture them. I’m not looking to create doves from my hat, merely maximize what is already there. A man can hit a stone for 30 years, unless he knows how to sculpt he’ll never make the Statue of David.