Playing with RoR Curve Shape & Artisan Designer Tool

#1: Post by **Almico** » June 8th, 2018, 12:59 pm

Now that I've given in to the importance of steadily declining RoR, it's time to start playing with different RoR curve shapes and see what effect it has.

If I do a "set it and forget it" roast with my roaster I get a concave curve that looks more or less like this:

The RoR rises quickly in the beginning and then quickly drops off. The slope starts to level out before 1C and by the post 1C development segment it can almost flatten out.

I want to try and get a straighter RoR curve, and then maybe try a convex curve.

I went into Artisan Designer and created a dark roast curve. That's what you see in the background. When setting the Designer milestones I took into account my normal turn around time/temp, 300 dry and 390 1C. I get 2C around 430and 1Ce a bit before. I then moved the points horizontally until the RoR flattened out.

In order for this to happen on my roaster I need a lower charge temp and carefully controlled heat till "RoR turnaround", the upside down TP on the RoR curve. Then I have to continually add heat bit by bit to keep RoR from dropping below the line. It's a pain. But I did a roast of my house dark on Wednesday:

and the results were interesting. I'm brewing it in the bar today and while I thought it tasted a bit flat, every one of my dark roast customers likes it better...much better. I will say that it is totally roast defect free. No astringency or roastiness whatsoever. It's almost too smooth, which is why I thought it flat.

But the point for me is that this was very different from usual roast. Although milestones are basically the same, the path to them changed. I think what has to happen to flatten the RoR is the dry needs to be a bit later Maillard a bit shorter. Basically, the flatter the RoR curve, the flatter the BT curve. This might be a round about way of getting to a slow start, fast finish roast, but nowadays I'm think much more in terms ofRoR curve than BT curve.

In order to keep drop within Raos 25% dev time, and get a flat RoR, I redid the designer curves like this:

There is still much to be done, especially with lighter roasted coffees, but the results are promising.I was hoping that was not the case, because it really is much hard for me to roast like this.I need to be on the heat dial right from TP instead of 1C.

I'm not sure if my natural concave RoR curve is a trait of air roasters more than drum.

June 8th, 2018, 2:28 pm

I just tried a lighter roast curve in designer. You can see where the points needed to be:

And this is the pattern that I can load into the background and follow.

Drop temps can change by smell anywhere along to curve, but this is the target.

#3: Post by **mbenedet** » June 8th, 2018, 2:55 pm

I'm happy to read these posts of your experiments. The only thing I have to add at this point is that Scott himself said it's fine to betray the 20-25% DTR guideline for dark roasts. 30% into second crack is no issue. On my setup, second crack is +40F from the end of first, quite a distance (and time) for the end of a roast. You'd have to be flying at top speed to get there with a 25% DTR.

June 8th, 2018, 3:12 pm

I never worried too much about hitting 30% on dark roasts, but as long as I was designing a theoretical roast from scratch, I figured I try and stay in bounds. It just means targeting DE and 1C a bit later.

I find hitting 2C with gusto makes for a livelier cup anyway.

#5: Post by **mbenedet** » June 8th, 2018, 3:32 pm

I just meant more like you don't have to try to get to second crack within 25%.

Looking at your curve a second time, I'm seeing that your first crack is around 400 and you're calling second crack at 429. What size probes do you have?
Just as an example, on my larger roaster with 1/8" probes, I hit first crack around 380 and second crack around 420. On my smaller roaster with 1/16" probes, first crack is between 380-400, and second crack on my Brazil is around 450. I know every bean and setup is different, but I'm just surprised by how off ours are.

Depending how you're "calling" first crack, your DTR would change. Say for the sake of argument that I'd call it at 9:00 in your curve, then your DTR would actually be ~32%.

June 8th, 2018, 4:27 pm

397* was an anomaly. With that coffee I usually hit 1C between 388 and 392. That's why I set the template at 390. I call 1C when I get the 1st few pops in succession. I call 2nd when I hear the first few crackles.

I use a 3/16" probe. I've tried thinner but they are way too sensitive in my roaster. I don't buy into the thin probe camp. Not with Artisan. Too noisy. My probe is nicely matched to my roaster and gives me excellent feedback to make small tweaks on the fly.

FWIW, I use an air roaster and the probe is buried in the bean mass and isolated from the hopper wall.

June 8th, 2018, 8:01 pm

I used the lighter roast template on a small batch of a nice Colombia microlot coffee. It was hard controlling the small batch, especially starting with a cold roaster, but I think I got it in the ball park. This technique will take lots of practice and experience with each coffee to help predict heat transfer tendencies.

#8: Post by **crunchybean** » June 9th, 2018, 8:42 am

Almico wrote:I used the lighter roast template on a small batch of a nice Colombia microlot coffee. It was hard controlling the small batch, especially starting with a cold roaster, but I think I got it in the ball park. This technique will take lots of practice and experience with each coffee to help predict heat transfer tendencies.

<image>

I have found that similar beans ( 1st)terrior 2nd) varietals) will roast similarly so one profile you have found how to apply the heat to that bean and get the flavor/texture you want, massive variations in profile will not amplify what you are going for and then small tweaks are required. So for two different Ethiopian from Sidamo (different years crop), will roast very similar even though the heirloom varietal composition is random (at least to me). Also I have found that creating very specific profiles will get you very specific flavors, so you will be losing out on "complexity" of having multiple fruits ect, but you may get one note/texture from each category in the beginning middle and end that are very distinct. That is said to share what I have observed. These I am sure, are not new observations, but may help if you find yourself lost. I have not been able to eliminate the baked nose from the hot cup and if you have any ideas on that I'm all ears. On that note, I have not mastered the Rao Method so that advice I am still working on.

June 9th, 2018, 10:08 am

Thanks for sharing this. Very interesting. Just to be clear you are defining light vs dark as your drop temp on your system. Or are you using time after FCs?

It's great that you used Designer.

June 9th, 2018, 10:19 am

Both. To start, lighter roasts are closer to 20% development, under 415* drop temp and mid-2:00s post 1C time. Darker roasts nudge 2C at 430* and I would like to keep them inside 25% dev time.

These are just beginning guidelines to test the RoR curve thing. This is not about flavor nuances; it's about eliminating roast defects and maximizing "dynamics". I'm using coffees I know very well and want to see what happens.

So far I'm finding that coffees I thought had taste defects built in were actually fine and I had only my roasting ignorance to blame.

Designer is a great tool. In order to achieve my desired RoR curve, milestone times and temps had to be adjusted. It basically told me when dry and 1C had to be, given the temps I normally get on my roaster. Very cool.