Level: Apprentice

Reading time: 10 min

I have been playing with trying to control my roasts with the Gourmia and was having a lot of trouble getting the temps to hit about 212 at 10 minutes, and then tamping down the temperature for the development phase.  I finally got what I think are better profiles as you can see below.  (or are these good profiles at all?)
Tom deltas.jpg
My deltas are quite a bit off from what you recommend.  My ambient temp is about 75 and my voltage is within normal range.  For Deltas on these I used 335 at start, then 175 at 5 minutes, 100 at 10 minutes, and then about 20 to finish the roast.  If the roasting profiles above are what you think they should look like, why would my temperature settings so much lower? 
Am I trying to get too much control out of this machine?  The chocolate tastes fine but perhaps not with as much complexity as I seem to get from the Behmor—which runs contrary to what I would expect given that I have more control.
As a reminder, these are the kinds of profiles I was getting previously using your Delta recommendations:
tom profiles.jpg
Would love some advice on how I can leave less money on the table.

I am going to talk in a few generalities here and maybe or maybe not give solid answers.  My apologies in advance.

The first thing that comes to mind is that contrary to how you feel, you are doing basically everything correctly.  You notice you are hitting 212 faster than you want (we will talk about that in a moment) and you are decreasing your deltas and the roast is slowing down.  That is 100% exactly what I want you doing.  I DON’T want anyone blinding following my deltas. 

To that end, it sound like it is time I go back and revisit the Gourmia.  You say you are roasting where it is 75 F or so. Now that I think about it I did most of my testing in the winter when my ambient was much closer to 60 F and the beans were 50-55 F.  That is going to make a huge difference in how hot you have to set your roaster and the magnitude of those deltas.

Just the other day I was commenting to Mackenzie around here that the Royal didn’t need nearly the gas flow to hit the profiles and when she roasted she needed to pay attention that the roast didn’t proceed to fast.  When the warehouse was cool, we routinely needed 9.5-10 cubic feet per hour for the first couple roasts to roast 25 lbs with the profile I usually use.  With the warm weather and the ambient temperature and beans being in the high 70s the roaster only required 8-8.5 CFPH.  Since we are talking deltas, in the winter I routinely had the roaster’s air temperatures 30-50 F hotter than I currently do.

I know I am meandering a little but I want to touch briefly on your goal of hitting 212 in 10 minutes.  Why is that a goal?  I’m not saying it is wrong, but I want to make clear that drying time does not inherently matter.  It only matters as a way to gauge how fast you are coming into the Development phase.  When I roasting 25 lbs in the Royal, with 200 lbs of steel around it, momentum is critical.  If I hit 212 F in 8 minutes, I am going to have one hell of a time slowing the ramp down enough to have a reasonable (read 2.5-3.5 minute) Development time.  But the Gourmia, Behmor and any other small roaster isn’t going to have that issue.  The mass is so much less, the momentum is also going to be less and slowing the ramp down is usually no harder than just turning the heat down a bit or even off briefly.

Way back in the dark ages of bean to bar (2003) when I was experimenting with air poppers to roast cocoa I found had reasonable roasts in the 5-6 minute range.  This was because I was only roasting 2-3 ounces and they were dry in 2 minutes.  Yes the Development and Finishing phase are really fast too and one big reason we don’t roast in air poppers, but the point is that under certain cases, 2 minutes can be just fine for a drying phase.  Likewise, you don’t HAVE to be at 10 minutes at 212 on the smaller roasters.

With all that, what I have to say now is


It is also totally possible that you can ignore everything you just read and really there is just no problem. Many people have brought up that depending where you take your measurement from in the Gourmia you can get a pretty different numbers.  I personally take a mental average as I pass the laser pointer of my thermometer over the beans front to back.  For me, (and now I say in winter), those averages correlated well with what I was smelling from the roast so I ran with that.  I fully realize that does not do you much good while you are learning to roast and don’t know what a proper roast smells like.  I have tried to lay it out but really it just takes time and accumulating experience.

It is fully possible that the temperature readings you are taking are not representative of the true average temperature of the roast. You may find that the beans closest to you are more representative of the stage of the roast.  If you remember, I suggest you hold the beans at your EOR temperature for 5 minutes or so.  This is because I know (based on smell and taste) that when the beans initially read 250 F in the Gourmia all of them are not 250 F.  There is no reason not to believe that when your beans initially read 212 F they are not all at 212 F and in fact many are cooler.  That right there is a good reason to take the coolest temperature.

At the end of the day, no matter how much graphs, charts, ramps and deltas I give, roasting is about using all your senses.

And you are doing that.  You mention the roasts are not quite as dynamic as from the Behmor and are writing for help and I will still try and give some suggestions.  Before that though I need to clear up a misconception.  You state this:

The chocolate tastes fine but perhaps not with as much complexity as I seem to get from the Behmor—which runs contrary to what I would expect given that I have more control.

It isn’t contrary at all. I love the Behmor because it does a great job roasting.  It just isn’t transferable to anything else as it is a closed system / black box.  Yes, the Gourmia has more controlbut ALL that means is you have full control to NOT roast well as well as roast better.  It is a double edge sword that cuts both ways.

So after all that, suggestions.

The first is going to be pick one bean so you can do comparisons.  It is really hard (basically impossible) to compare data from different beans.  You need to try and find what correlation you find between changes in profile and resulting flavor.

The next thing is to loose the graphs and use the Profile notation I have developed as close as you can.   Part of the reason is so we can talk easier.  Graphs look nice but are basically useless for critical comparisons of roast segments.  It also gets you thinking about the ramp speed that are just easier to comprehend.  I was going to try and pull the data of your graphs but I just can’t.  It is too overlayed and doesn’t have the needed resolution. The biggest issue is that your graphs increment from 200 to 250 where all the critical reactions are happening.

To that end, I would check your temperatures every 3-4 minutes instead of 5 minutes after the first 5 minutes.  There is no need to go down to every 2 minutes but you do want to try to nail those 212 and 232 points if you can.  Given the wide swings in reality that is 205-215 and 225-235.

This last suggestion is purely based on a gut instinct and that is to take the lowest temperature reading and maybe even discount it 10 F or so trying to account for the knowledge that the inner beans are not quite the temperature of the outer beans.