Roasting Techniques. What Every Drum Roaster Should Know: Roasting to Target Your Time, Not Temperature.

Why a bean temperature probe isn’t required…

With the RK method of roasting, or rather any manual method of coffee roasting, there is a certain art to the process. A certain understanding is necessary. Many new RK roasters, or any drum roaster, is intimidated by the process of both roasting without seeing the coffee, or knowing what temperature the beans are in reality.

I want to make the case here that “knowing” the temperature of the beans and seeing the beans roasting really isn’t necessary at all. One of the things we have wanted to engineer is an in-drum bean probe that is economical and durable. This is something we may well do one day, but it presents some unique challenges as it needs to be inside the rotating drum, and be able to withstand those temperatures and display that data on the outside. While we may tackle this at some point, let’s look at why it’s really not necessary in the first place. Let me start with this question:

I’m used to roasting for this many minutes in X Roaster, how come the time on the RK Roaster is longer/shorter?

It’s important to understand that roasting coffee is a ratio of volume of coffee to heat input. Let’s assume you are roasting on a small batch roaster and are roasting 500g in a nearly full container. In order to fully saturate that amount of coffee, being somewhat confined and bunched together, with enough heat to get it through second crack, it takes 16 minutes of heat at your temperature. If you tried only 100g of coffee I would take much less time to roast and therefore saturate that lower volume of coffee as it will be spread thin within your container.

6LB RK Drum 

The same principle applies in a larger RK drum.  500g in a 6LB drum would roast in about 6 minutes as it would be very thin within the drum and would require very little heat energy input into the system to complete.  Conversely, a fully loaded drum (6LB) at approx 83% volume (you need space to mix the coffee) will take 20 minutes to fully saturate the coffee with enough heat energy to bring it to 2nd crack.  This is our target roast time.

Have a look at the following roast table.

6LB Drum

Finished Roast Volume Green Load %Load Temp Est. 1st Crack Est. 2nd Crack
1LB 3.2 OZ
8 mins
10 mins
2LB 6.4 OZ
12 mins
14 mins
3LB 9.6 OZ
14 mins
16 mins
4LB 12.8 OZ
16 mins
18 mins
18 mins
19 mins
7Lb 3.2 OZ
19 mins
21 mins

You will notice that the time increases on each respective drum as the load increases.  This is that amount of time needed to saturate that prescribed volume of coffee to bring it to 2nd crack.  Some worry about “baking” the coffee, yet this won’t occur until you exceed these roast times by 30-40% generally speaking.

Why isn’t it necessary to know the exact temperature of the beans?

Let’s discuss this:

While knowing bean temperature would certainly be ideal. Currently, it’s not a very convenient thing to accomplish in a spinning drum.  It’s, however, not cost effective related to its benefit as it would be about $200 USD in equipment to pull it off. 


This is more important for Cacao roasting as you don’t have the 1st and 2nd crack markers to gauge by.

For now we feel very comfortable with our solution.  Let me explain it and hopefully it will make sense. Ideally of course, it would be preferable, as you would suggest to have your probe in your beans to know what temperature your beans are “feeling”.  Can’t disagree with you here.  That would be ideal.   Because we can’t easily get a temperature probe into a rotating drum at 52RPM without all kinds of complicated things, we have to settle for having a temperature probe right in the chamber, as close to bean level as possible.  Clearly the coffee will be sitting closer to the heat than the probe a few inches away.  Okay, fine.  So let’s say, for the sake of the discussion,  the barbecue is running at 500 degrees F within it’s chamber as registered on the lid-mounted probe.  We can assume, whether it is true or not,  that the coffee is running slightly higher.  Let’s say it’s at 530 F.  So is this a problem?  The coffee is hotter than we think it is!  I say it is not.  Here is why.

Coffee roasting is a process that follows a time and temperature curve, based on time and temperature variables. The more heat the faster the process goes, the less, the slower, it’s a simple operation. A bit like boiling water. You inject enough heat into water, it will eventually have enough heat energy input into it where it begins to boil, convert to a gas and boil away right? Coffee is the same principle. It is essentially a seed with water in it. Between 9% and 14% ideally inside the bean (this is why your old green has a weak sounding 1st crack, as it’s water has evaporated over time sitting in storage and makes the bean “pop” less as its increasingly limited internal water produces much less force escaping from the bean and producing a smaller pop.)

Now, if we inject enough heat into the bean, the water inside will begin to boil, the sugars will begin to caramelize, and will enter first crack. (the evaporation point of the water, the sound of the water escaping is first crack). Eventually as the bean caramelizes, it will eventually fracture its outer shell, (2nd crack). All of these can happen at a relatively predictable, rate (provided that heat is constant) once the coffee has received enough heat energy. Just like, once all the water has reached 212F, it will begin to boil and evaporate every time from our pot of water. We can predict it. If we run the same amount of heat into our pot of water in repeat experiments, we can pretty well set a watch to it and within 15 seconds or so, predict boiling every time.

So let’s put that to action.  When roasting on the RK Drum, (or with any manual coffee roasting method) you always want to roast your coffee towards your prescribed roasting time referenced in the chart above.  In some systems, it is computer controlled so you never think about this.  With manual roasting, we do want to consider what is happening in the system.

So let’s take this new knowledge to an example. Let’s say you want (per the above chart) to roast 6LB of green coffee to produce your 5LB (2.5KG). Our estimated temperature will be between 550 and 575 F.  That is the temperature of the chamber, not necessarily the beans as you might suggest.  But, at this estimated temperature, (and whatever temperature the beans happen to be at) it will cause the water to boil and go into 1st crack at around 18 minutes, and then enter 2nd at 19-20 minutes.  (We know this to be generally true from repeated experimentation).

Now, there will be variables, your barbecue might be different. Your temperature probe might be closer (and therefore feel hotter) to the flame (or it could be farther away) than our test probe from which we created the above roast charts, so your numbers will be different than the chart. But, the chart is close within a close range (1 or 2 minutes). It is close enough. So let’s look at one final piece and put it all together

Coffee roasting for this particular profile will be a linear operation, and something that we can control and predict. You inject more heat it roasts faster. You inject less heat it roasts slower. We can adjust our heat to target our time to get that perfect coffee.

In our example, we are targeting an 18 minute first crack. Lets say we choose 560F for our roast. We start our roast. We want to make sure we are at 560 and holding at no later than 4 minutes after the start of the roast. (Always remember to preheat your drum before putting in the green coffee). This should, produce a 1st crack at around 18 minutes, more or less, generally speaking. If we are at 17 minutes and we see no steam/smoke (telling us we are approaching 1st crack). This tells us what? That we aren’t close to first crack right? We are looking for an 18 minute first right? So what must we do? Accelerate the roast. Inject more heat. More heat means faster roast. So we bump up the heat from 560 to 580 or 590. The exact temperature isn’t important, but we know we need more to get back on our time profile since we’re behind. We are going to try and get the roast to fire first crack at 18 minutes. We adjust our heat accordingly. Do we care what the beans feel? Not really. We have a relative reading from right next to them, from the chamber probe. We have a relative idea of what’s happening in the bean mass, they are “on target” or they are roasting too slowly, or too quickly based on time and 1st/2nd crack indicators. We are roasting to target our time, not our temperature.

So now we adjust the heat to 590F. This accelerates the roast and we begin to enter first crack quickly now and we achieve first by 18:15 (approx) because we injected more heat. If we keep this trend we will be assured of a 2nd within 2 more minutes which keeps us on our time profile.

Conversely, if we are seeing signs of 1st crack at 16 minutes, that tells us what? We are too hot. We need to back down to 530 degrees (approx) to slow things down to stay on our target time. So, you should start with the chart on the website as a starting point and you adjust your temperature to achieve your time profile and you will come out with the same roast every time (for each coffee) without seeing the coffee, or knowing what the beans are feeling. You can set your watch to it and you can even predict 1st or 2nd crack by your clock, to within about 15-20 seconds of accuracy. This is what I mean by adjusting your temperature to achieve your roast time. It works every time.

Keep in mind, your roast times and temperatures will not exactly match the chart, nor should you worry about trying to get exact. Just get close. Within a minute or two of the times and there is nothing to worry about. The temperature is far less important, as any coffee will generally react the same way within a close margin.

Sure, one coffee will roast faster or slower depending on it’s internal water, mass and composition, but within reason, most coffees will perform close to the same as the next coffee. In one case, we had one roaster who was only getting consistent, predictable roasts at 750 degrees. This is far, far above the chart predictions. In his case, his thermometer could have been off. It could have been very close to the flame. Whatever the reason, the fact remains the same. If you put X amount of heat into a pot of water, it will boil at X point in time, every time, given the external environment (the air’s temperature and pressure around the pot) remains the same.

Thus in the same vein, coffee will enter 1st when you have finally achieved enough heat to get it there, and same for 2nd.  Coffee doesn’t care what temperature it is at.  You can pour heat into it so fast that it enters first in 2 minutes.  Or you can drag it out to 40 minutes.  The question is, what is the ideal roast time for coffee in general.  Longer than our prescribed time, we start getting into the “baked” taste range.  The shorter we go, we get generally brighter coffee.  In any coffee, eventually the internal water will boil and you will get your 1st and then your 2nd.  This happens regardless if you know your bean temperature or not, or if your thermometer is reading 1200 degrees sitting in the fire.  The temperature is merely a reference point for us to have a consistent indicator of how much heat we are putting into the system.  Where we place our thermometer will give us a consistent indicator for our own roaster.  So, in all cases, roast to target your time and you will get excellent roasts every time, regardless of what your thermometer says, or how much it matches up with the charts, because your coffee really doesn’t care what the thermometer says.  It only cares how much heat it’s receiving and how fast.

coffee beans with smoke coming out.

To do this, it is important to recognize the signs of 1st and second crack.  1st will start showing steam/smoke about 1-2 minutes before 1st crack ( a bit like fog or steam).  This is a good indicator that 1st is imminent.  Use this as your guide.  2nd will show heavier wispier smoke (more like cigarette smoke wisps) about 1 minute before 2nd.  Use these indicators to plan your 1st and 2nd crack arrival on time for your roast volume.   Note that the smaller volumes of coffee, the corresponding smoke will be less as there is less material in the drum to convert.   If anyone has any difficulty with this, we will assist you on the phone and walk you through the roast while you do it side by side.


 If anyone has any difficulty with this, we will assist you on the phone and walk you through the roast while you do it side by side.

Your grill will certainly be different to some degree more or less, but these charts work for most people.  In any case, adjust your temperatures to achieve these times and write down what is working for you.  That becomes your roast profile for your roaster.  Use that from now on, just utilize our charts as a starting point to get you there.

Understanding these techniques will allow you to produce, repeatable, consistent, uniform coffee time after time, without seeing or knowing the temperature of the beans. Feel free to email or call with any questions at all about these techniques and concepts. Having ultimate control over temperature ant time, gives us far more interesting and often better roasts that what you can get from commercial $30,000 machines. On another note, you will finally understand and comprehend the actual art of roasting coffee as opposed to pushing a computerized button and walking away! Enjoy and Happy Roasting!