In this video I'm going to demonstrate a procedure for developing a roast profile. So what is a roast profile anyway? A roast profile is just a record of how a coffee was roasted. Most importantly, that tells you what the bean temperature is over time. How hot is the coffee at any given moment? You can also record other information on that. If you have access to the temperature in other parts of the roaster such as the air temperature the exhaust temperature you can have those in your roast profile as well.
Those are not necessarily things that you'll want to match on future batches. It's not as useful as the bean temperature but it can provide some additional information that helps you match that bean temperature. Similarly, you might have access to the control data such as fuel settings, air flow, there are also roasters that have color meters built into them so you can tell what color the coffee beans are at any given time and that can also be useful if you have it.
So how do you create a roast profile? There are a few different options here. You can record the time and temperature at just a few key points: perhaps just the temperature that you drop the beans at, where it bottoms out, and a few key points such as when chemical reactions start to occur, first crack, second crack, when you're ending. The benefit of that is that you have a very compact record of the roast but you also don't have a lot of detail there. another approach is to record the temperature data at regular intervals. You might record the temperature every 15, 30 seconds.
Do that very regularly throughout the roast. Another option and what we're going to be using here is automated computer logging. With automated computer logging you don't have to worry so much about the mechanics of creating the roast profile. You can focus more on the act of roasting itself. With the software we're using here, all of that roast profile data goes into a database that makes it much easier to keep track of what you're doing over time. It lets you easily pull up the information on any given batch of coffee.
So why use a roast profile? A big reason here is consistency when you have a roast profile you can duplicate that profile creating the same bean temperatures over time as your target profile and that will help immensely in getting the same coffee on every batch you can also have some additional checks on the consistency keeping track of percent weight lost, bean color cup character, things like that. simply using a roast profile gets you consistency now that might be consistently good, it might be consistently awful, but every batch is the same. You develop a roast profile to get quality.
What is the best way to roast this particular coffee? So we're going to start roasting this coffee. What I have here is a very nice coffee from Bolivia. So we enter the green coffee we're using, the weight of that coffee and now we're ready to start our profile development batch. Toward the end of this batch, I'm going to pull several samples and these are going to be identified by the letter B and a number. We can set that up in our data logger. Several minutes into the batch we can see what's going on in our data logger. On the top with your most recent set of measurements. I have two thermocouples on this roaster so I can read the bean temperature and the air temperature I also have the time since the start of the batch. There's also a table view that shows the time, bean temperature, air temperature and control annotations.
I have that set to show one measurement every 30 seconds plus any measurement that's associated with an annotation. That can be adjusted to show more or less data. We also have a graph view of the data as it's coming in. The lower curve here shows the bean temperature and the upper curve shows the air temperature. Since heat transfer is mainly from the hot roasting air into the coffee bean the air temperature is generally going to be higher than the bean temperature. What exactly those temperature measurements are that's going to vary from roaster to roaster and it's going to vary depending on how well the probes are placed, drum geometry, all of that can have an effect on that. Now we're ready to pull some samples. What I'm trying to do is get samples across a range from lighter than I'm looking for through darker than what I'm looking for and with this particular coffee I'm looking for two roasts. I'm looking for one good medium roast something that has a particular flavor characteristic that my customers are accustomed to for this particular coffee and I'm also looking for a darker roast something where you can clearly tell that this is a dark roast coffee but you can also still tell that the coffee comes from Bolivia.
So here I'm marking in the data logging software when I start pulling my samples. and I need enough of each sample so that I can cup these later. You'll note that I'm not particularly concerned with cooling these beans. because in a real batch it does take some time to cool the beans off. With these smaller samples, the cooling can happen much faster. so I'm going to pull all of my samples and when I'm done with that I'll shake the coffee beans and that will get them cooled off in about the same amount of time as a production roast on this roaster. now we can see the data that we collected. We have 12 samples here and the data is saved for future recall. Next, we need to cup all of these samples.
And you'll note that I'm not using a cupping form here. By the time the coffee gets to this point we should already know that the coffee itself meets our quality expectations. We're really just trying to figure out how to roast the coffee. so I'll taste all of these samples that we've pulled. So what I'm doing here is I'm just picking out the samples that I like, what meets the criteria that I'm looking for in terms of a roasted product what I'll usually do here is try all of them once then try them again and pull out the ones that might be good and then as the coffees cool I'll narrow that down further, tasting the ones that were good, see how they act as they cool off and through a process of elimination eventually we'll pick out the coffees that I want to replicate so considering the lighter of the 2 roasts that I'm looking for I like sample 4, I like sample 5 and these are really very close to each other so I'm having a little bit of a difficult time picking between them and this is really just going to come down to how well they perform as they cool off.
and it turns out that sample 5 was really nice and we also have a sample 10 that was that was very nice on the dark side. Now here we found a couple really nice samples that we wanted to duplicate but that doesn't always happen sometimes we'll go through that and find that we need to make some changes to the roast profile a lot of this is based on experience you might know that changing the air flow on your roster does a certain thing and that that might be beneficial for a particular coffee The other major change is adjusting how long you spend in certain temperature ranges.
Having an understanding of what's going on chemically during the roasting process can really help with that if you know where certain chemicals are being synthesized and where those are being broken down that can really help in targeting those changes for a coffee I'm less familiar with I might do multiple batches all up front to see how different parameter adjustments will affect that coffee. So here what I'm doing is defining in our database of roasting information a roasted coffee item We've had this particular mark in the past so I'm just going to take that out of the list of discontinued items and put it back into the list of current items. Now I can enter in the details of the batch. The roasted coffee item, the green coffee that I'm using, how much coffee I'm using for this batch and once again we don't have a target roast profile set up so we'll click the no profile button, of course, we do have that roast profile information.
It's just in a different place so we'll load an additional roast profile as our target pull up the sample batch and load that as our target roast profile and there it is and here what I'm going to be doing is trying to duplicate sample B-5 from that data so we won't be going through all the way. We'll stop it once we're done so here I'm keeping an eye on the data that's coming in, what the current bean temperature is, what I'm shooting for in my target roast profile, trying to keep the controls nice and steady, trying to match up with our target profile as best I can After several minutes this is what we see in the data logger. We have our target roast profile information on the table and on the graph just as before but we also have new columns in the table with our current data as it's coming in. We also have two additional lines on the graph showing the current bean and air temperature.
I'm trying to match the bean temperature as closely as I can The air temperature I don't really care about these were different sized batches as we can see the initial temperature was much hotter those air temperatures are not going to match up. but they do help me make control adjustments to keep the bean temperatures lined up. On this particular roaster, chemical changes aren't happening until the beans hit 300 degrees Fahrenheit so the temperature readings that we're getting before that point don't really matter so much.
That's why I'm not worried about the higher starting temperature or the higher turnaround temp. As the roast progresses though we see that the temperatures line up very nicely with the target profile. So at the end of the roast, we can see that we did a pretty good job. The bean temperatures are lining up very well with the target roast profile we hit the same end temperature at the same time It's not really realistic to expect to get much closer than that.
So now that the batch is done we enter our roasted weight that gets us a percent weight loss and since I'm happy with how we replicated that batch we can go ahead and save that as a target roast profile and we won't have to see the extra data from our initial test batch when we want to duplicate that in the future. Once we're done with this I'll take that coffee and I'll brew it as a customer might That gets me a final approval on that roast profile results from the cuping table don't always translate to a normal brewed coffee so we do want to make that check and make sure that what we're doing really is going to taste good for the customer this is also a good time to record any quality assurance parameters that we might be interested in these are very similar to those checks on consistency that were mentioned earlier We'll record the percent weight loss some places will also record the color of the bean there are machines that will measure the color of a coffee there are also color tiles that are available that are not quite as precise but are much cheaper.
The other thing that we're interested in keeping track of is what does this coffee taste like? Here again we don't necessarily mean a cupping form, we're just interested in what are the important elements of the flavor of this coffee here again we've already determined the quality of the green coffee We're looking for more of a qualitative analysis that we can use for future batches Does this coffee taste like the coffee that I roasted last week last month and so on? and with that, we've developed a target roast profile.
As found on Youtube