I am looking for a general overview of roasting theory as it involves heat adjustments and first crack. There are three possible times to cut your heat in relation to first crack: just before, at some point after first crack has begun, or at its conclusion.

What are the rationales for cutting the heat at one of these phases as opposed to the others, and of sole importance, what effect do these approaches have on the final cup?

noah wrote:I am looking for a general overview of roasting theory as it involves heat adjustments and first crack. There are three possible times to cut your heat in relation to first crack: just before, at some point after first crack has begun, or at its conclusion.

What are the rationales for cutting the heat at one of these phases as opposed to the others, and of sole importance, what effect do these approaches have on the final cup?

I wish there was a simple straightforward answer to your question, but as you will see this is a little more complicated than that, both in theory and in execution.

Forgetting for the moment the obvious physical differences among roasters, and the fact that coffee beans themselves are quite variable in their roasting behavior, one has a few general things to keep in mind about the roasting process, which include:

(1) As the roasting process unfolds, it requires the most heat input initially in order to get it started, and less heat input per unit of temperature rise later in the roast. People speak about there being a point in the roast at which it becomes "exothermic," e.g. that it creates its own heat, at least to some extent. This exothermic effect does not start, in my own observation, until a significant part of the first crack has already happened.

(2) In a broad sense, the faster you roast, the "brighter" will be the roast product, and the slower you roast, the duller or "flatter" it will be. Neither overly bright nor overly dull are desirable results, but where in the continuum you will want a given coffee to end up will depend in part upon the coffee itself and in part upon its intended usage (for example, to be used in espresso or in brewed coffee).

(3) Roasters themselves are highly variable in their abilities to increase heat (over a set period of time), to retain heat (moreso in drums), and in the ability to input, control, and adjust the temperature by increasing or decreasing the heat inputs and by controlling the ventilation. Not every desired adjustment is possible to accomplish in a given roaster, and the speed with which you can do these adjustments is highly variable. This is to say that you will be limited by the design of the given roaster in the extent and timing of any adjustments you might want to make. At the most adjustable end of the scale would be a small air roaster with a powerful fan and heat source, both of which could be finely adjusted. At the least adjustable end of the scale might be an enormous drum roaster that can roast an entire 132lb bag of beans that is relatively poorly ventilated and lacks fine adjustments on its burners. This is to say that you "can't always get what you want," you might have to settle for dealing with the limitations of your roasting apparatus.

Still, at a minimum, if you really want to be able to roast like a pro would roast, at the very least you want to be able to control the heat input directly and quickly, during the roast, without having to jump through a whole lot of hoops.

Even accepting #3, e.g. the limitations of a given roaster, one also has to consider that during a roasting session the apparatus may change its behavior, so that as it retains more and more heat, it might require less of a heat input, with this changing roast by roast during the roasting session. Drums have this problem and I personally have to deal with this phenomenon during my own roast sessions. The amount of heat input that I have to use during a roast will be considerably less on the 5th batch than it was on the first, even though I preheated the drum before I started using it for the first roast.

Moving onto your questions: In general you will be inputting less heat progressively during the roast. The initial part of first crack is sort of like a hill you have to climb in the middle of the hike, so you might well have to increase heat as 1st crack gets going in order to avoid stalling out the roast. Stalling out the roast is where the temperature of the beans stops rising or even falls over a significant period of time (lets say a minute for the sake of discussion). Once you get through about half or a little more of 1st crack, you will need to significantly diminish your heat input or the roast will "take off by itself," and very rapidly go into second crack. If you do this, the coffee you produce will be of very poor quality, especially for espresso.

So, again getting back to your questions, when you cut the heat input is going to depend to a great extent on your roaster. With my roaster, which has a heavy drum that retains a lot of heat, I need to cut the heat progressively during the roast after first crack starts or at least by halfway through 1st crack. Coffee beans differ in their roasting behavior and in the amount of heat they will need to progress through the roast cycle. In the case of many bean types I need to increase heat after the initial 30 seconds or so of 1st crack, then to cut it back again severely a minute or two later. I do this on the basis of accurate thermometry, in my case a thermocouple probe in the middle of the swirling bean mass in my drum. The increase in heat input as 1st crack starts is done because without it, on my roaster, the roast will stall and the bean temps will fall. I'm trying to have them rise very very slowly at this point, but at least in the case of my roaster, this means I must add heat then. Without thermometry this is harder, but without question you are looking for the popping sound of first crack to continue and not to stop; if it stops right after it starts, you are almost certainly stalling out the roast.

Somewhere about halfway into first is where I really cut my heat input a lot. Once again, I'm relying on thermometry. I then progressively reduce the heat input depending upon what I see on my digital thermometer, with the aim to have about 5 minutes pass between the onset of first crack and the end of the roast, depending to some extent on the degree of roast I'm looking for.

So, in summary, you can't discuss heat adjustments and their effects on the roasting process (or results) without considering the behavior of your roaster, its controllability, the beans you are roasting, and the desired final results. I know this is not the simple direct answer you would probably have liked to read, but there are reasons why it takes a long time to learn how to roast, to get the results you want to get, and the above discussion shows a few of the variables that make this both a difficult thing to learn to do, and a rewarding one, once you have gained some experience in doing it, yourself.

ken

On my roaster, also a heavy ventilated drum, it would appear the beans go exothermic only a short time after the beginning of the first crack. Heat to the roaster is controlled via a environmental temperature probe, if it is above a preset number, the heat is removed. Shortly after first crack starts, I see the environmental temperature quickly spike typically by around 5 degrees C which is a much higher figure than the roasters average hysteresis. What this in effect does, is reduce heat input (as Ken said) quite rapidly. I then need to increase the set point temperature in order to get the heat going back into the drum without letting the temperature decline in the interim. This stops the first crack stalling and lets things progress along nicely. I'm generally also in the vicinity of 5 minutes from the start of first crack to dropping the roast slightly into the second crack, perhaps marginally less if not a full load.

Please bear in mind that although I've done a few hundred roasts in my machine, I am certainly no expert.

Nice post Ken.

Cheers, Chris

Talking about this more abstractly, the beans don't get their heat from the flame or electric element, they get their heat from the air and hot surfaces surrounding the beans -- the environmental temperature (ET).

If there is one thing that somewhat works for all roasters, it's that a good roast requires that the ET follow a certain profile, and stay within certain limits, so the beans don't under dry, over dry, bake, scorch or tip. Basically, this magic profile is a starting temperature of around 325F to 400F, and a ramp up to around 450F to 480F in around 6 to 8 minutes, and holding it steady there to the end of the roast, whenever that may occur.

This ET part is basic roasting chemistry, and the same for all roasters and coffees. But the heat inputs required moment by moment to achieve this ET curve is based on the roaster's thermal characteristics, and is different for every roaster design.

If you have an air roaster with no insulation and no heat storing parts, you will need to add progressively more heat at each stage of the roast, since the ET will be the same as the hottest air temperature blowing in. If the roaster is heavily insulated, has a lot of heat storing mass, or recirculates waste heat, then one needs heavy heat inputs early in the roast to get all the parts preheated, then a gradual reduction in heat as those prior inputs work their way to the beans.

another_jim wrote:If there is one thing that somewhat works for all roasters, it's that a good roast requires that the ET follow a certain profile, and stay within certain limits, so the beans don't under dry, over dry, bake, scorch or tip. Basically, this magic profile is a starting temperature of around 325F to 400F, and a ramp up to around 450F to 480F in around 6 to 8 minutes, and holding it steady there to the end of the roast, whenever that may occur.

Another gem from another_jim

Seriously, that's a concise and clear guideline, thanks Jim. Now:

another_jim wrote:This ET part is basic roasting chemistry

Can you share more insight about this?

another_jim wrote:Basically, this magic profile is a starting temperature of around 325F to 400F, and a ramp up to around 450F to 480F in around 6 to 8 minutes, and holding it steady there to the end of the roast, whenever that may occur.

I'm a bit confused here, and just want to check to make sure you meant what you said. What you are saying is that the ET of 450-480 should be achieved by the 6th to 8th minutes of the roast, not that you crank up the heat towards 450-480 at 6-8 minutes into the roast, right?

Another question. Can I infer then that, hypothetically, one should be able to profile while monitoring ET solely? What then is the practical function of monitoring BT? To back up and confirm ET? Can we prioritize all this so that I can learn to focus and control one aspect at a time (since I am a slow learner...). For example, should I first confidently execute the proper ET profile, then measure the bean temp to align it with the ET profile, and then make it fit within a time scheme? Or should this proceed in a different order?

noah wrote:Another question. Can I infer then that, hypothetically, one should be able to profile while monitoring ET solely? What then is the practical function of monitoring BT? To back up and confirm ET? Can we prioritize all this so that I can learn to focus and control one aspect at a time (since I am a slow learner...). For example, should I first confidently execute the proper ET profile, then measure the bean temp to align it with the ET profile, and then make it fit within a time scheme? Or should this proceed in a different order?

Hi Noah,

I, like you, am a slow learner. I find a lot of these sorts of things (ET, BT, what-have-you) as confusing and more or less useless, unless I happen to have the same sort of equipment as the person describing them, and I believe that this person has really mastered things in a way that they can fully comprehend the (potentially more limited or less easily monitored) equipment that I have.

What you really want is reproducibility; you want to be able to do things in a way that you can follow the impact (in the cup) of what you are able to do with your roaster. Since it is not all that easy to put a temperature probe exactly where you might want to put it, you are left with the situation of trying to find a place to put the probe, that you can put the probe in each time you roast, so you can follow something consistent.

In my opinion, the best place to put a probe if you are only going to have one probe (my situation) is at a fixed point that the beans will be moving around. It doesn't have to be the exact center of the bean mass, but it ought to be the same place or close to the same place each time. If that isn't possible than I'd settle for something like an exhaust temp, but once again you need to put the probe in the same place each time, so that you are getting consistent information. I would try to keep things as simple as possible. Only in the case of a hugely expensive commercial roaster, or some heavily hacked nerd's toy, are you going to be able to monitor multiple temps simultaneously, and even if you could do that at this stage, would you have any idea what to do with all the information?

Once you have consistent information, you will be able to correlate what you can see and hear during the roasting process (e.g. bean appearance, sound of the cracks) with what you measure on your digital thermometer. You probably already have a pretty good handle on what are the variables that you might want to try to control, such as how long it takes to get to first crack, how dark you want to roast the beans, and how much time you want to spend between the onset of 1st crack and the end of the roast.

If you can learn to correlate temperature changes measured from a fixed point, with changes in the other observable factors in roasting (appearance, cracks) with results in the cup, you will then have learned a lot more than most people will ever learn as home roasters, and you will have a degree of control on the end product that will take you about as far as your particular roaster can take you. At that point, if you want more from your roasting experience, be it better results or larger batch sizes, you can take what you have learned about roasting and apply most of it to another roaster. When you go seeking this "other roaster," by this point you will have a pretty good idea of what kinds of roaster upgrades will give you the results you seek, and which won't.

ken

noah wrote:I'm a bit confused here, and just want to check to make sure you meant what you said. What you are saying is that the ET of 450-480 should be achieved by the 6th to 8th minutes of the roast, not that you crank up the heat towards 450-480 at 6-8 minutes into the roast, right?

I don't know how many times I have to link to this. It's a dissertation on coffee roasting. One part of the dissertation is on roast profiles, and deals with the profiles used by the major instititutional roasters (mostly Nestle's, since they paid the Schenker's fellowship).

They have a simple three step profile for ET:
1. they blow the air in at around 350F for two to four minutes,
2. raise it to 475F or so over the next two to four minutes,
3. and leave it there to the end of the roast.

They stop the roast when they hit a certain bean temperature, although the more recent literature has them experimenting with "electronic noses," sensors that monitor the ratio between two combustion products, one whose quantity decreases during the roast, and one whose quantity increases. This is supposed to be more accurate than stopping on either temperature or color, since these are affected by the air movement, roast speed and/or bean type

Basically, if you're not Nestle, and you profile the ET, you also need to measure BT to stop the roast accurately. Thermocouple thermometers cost $10 to $20 now, so it's no big deal festooning your roaster with several. It also gives it some mad scientist chic.

Industrial roasters use high convection rates, so this data is most applicable to air roasters and well ventilated drums. However, the chapter in Illy gives the Reynold's number equation for converting the heat transfer rates, and it's nicely monotonic. This means that for an unventilated drum roaster, you will get the same heat flows and reaction rates except in slow motion. You use the same profile but stretch out the time axis. In other words, an industrial roaster may do a roast in 7 to 10 minutes, and a drum in 11 to 14.

So putting all the research pieces together, here's how you tune the ET timing:
1. Hold the starting temperature until the beans hit 300F. Adjust the starting temperature to get the right amount of drying (higher starting temperatures for faster, less complete drying)
2. Start the ramp when the beans hit 300, finish just before the first crack, when they hit 385 to 390.
3. The final temperature is based on depth of roast and how tough the beans are. Set it so the roast finishes to whatever level you desire in at least 3 minutes for brewing to at most 5 minutes for espresso.

This tuning is not roaster dependent. The first two legs are based on the bean temperature hitting a certain mark, and this can vary by roaster, while the last phase, the roast finish, should be the same for all roasting.

Is this full of crap or is it based on something other than personal experience?

The overwhelming industry consensus, from 3rd Wave roasters to Folger's, is that roast times under 7 minutes or over 14 are not good. Another overwhelming consensus is that environmental temperatures above 500F are not good. A less strong consensus, but one held by all the third wavers, sees roast profiles that are described as Slow Start/Fast Finish as better than Fast Start/Slow Finish profiles.

The easiest way to guarantee Slow Start/Fast Finish roasts is to ramp the ET as I've described, and then to adjust the weight you roast to get seven to fourteen minute profiles. This may mean that the Behmor turns out to be a 200 gram roaster, the Hottop a 175 gram roaster, and the expensive one pound sample roaster actually an expensive half pound sample roaster. So what? You're surprised the manufacturer overstated?

For cupping, the margins are even smaller: For the coffee ratings used at sanctioned auctions or for contract arbitration, the SCAA prescribes roasts times of 8 to 12 minutes, 55-60 Agtron whole bean, and 60 to 65 Agtron ground, no matter what the sample roasting technology. This specified combination of time, interior and exterior bean color can only be achieved with roasts using the parameters I've described. People using unventilated drums for sample roasting just cut the load till they can get the right roast speed. People using air roasters set them up to ramp the supply air temperature. FWIW, my cupping roasts run 9 minutes, my brewing roasts 9 to 10, my espresso roasts 10 to 11.

In other words, this recommendation reflects the current state of the art.

Ken's point, if I understand him correctly, is that there's a hard, a very hard, and an insanely hard part to roasting. The hard part is getting your roaster to work within the parameters described. This will require learning your roaster, perhaps rebuilding it, and perhaps scrapping it and starting over. The harder part is that the parameters I'm describing are very wide; actually determining where you want to be within those parameters for each coffee requires lots of experience. Finally the insanely hard part is hitting that mark precisely at each roast. In actual fact, this could be impossible (say for me using the wrong PID parameters), a piece of cake (for a 20 million dollar industrial roaster) or very tricky and exhausting, say for a full manual roaster.

Great post Jim, concise & specific. Thanks. I'm guessing you are working on that 'How to Profile' article.

another_jim wrote:I don't know how many times I have to link to this. It's a dissertation on coffee roasting. One part of the dissertation is on roast profiles, and deals with the profiles used by the major instititutional roasters (mostly Nestle's, since they paid the Schenker's fellowship).

Probably because it's some guy's dissertation. Probably because it's just under 200 pages long. Probably because it scales heights of scientific geekdom that may well rival that of any of the H-B membership.

I've been reading through it over the last couple of months, and I have to go back to the ever-growing list of notes that I've made to try and integrate it into my own ideas. Between this and Illy's book, it's enough to make my brain pop.

-s.