Tweener roasts

#1: Post by **Almico** » July 13th, 2019, 4:31 pm

I've been suspecting a particular roasting anomaly for a while and think I've narrowed down the issue.

This is a typical medium roast profile:

I actually use this as a template for medium roast Central and South Americas. I let it go to 400* for a full medium, or drop sooner for a lighter roast. When I get lazy I'll use this profile and modify my heat settings when I want something different. I can tell by the BT curve whether I'm getting to 400* in 25% post 1C or 20%.

Anyway...being in the biz, I have the "luxury" of roasting the same coffees over and over again under lots of conditions. I am able to formulate a good game plan to achieve most roasting goals.

The other day I roasted 8# of a fantastic Colombia Yellow Bourbon that I've been working with for a few months. I have roasted dozens of batches, most on the light side. I use the above profile and turn down the heat a bit more before 1C and drop around 18% at 390*. It makes for a very bright and juicy drip. I will also take this coffee to 400* and drop around 20% for a more carmelly flavor profile. It is great both ways.

So I bagged up a few bags for retail and brought in a bucket for brewing at the bar. As my norm, I always taste the first cup of every roast. This time I was stunned to find it almost undrinkable. I tried grinding finer...much worse. I went all the way up to the coarsest setting on my R120 and it was drinkable, but not sweet and still had a hint of methyl alcohol.

Unfortunately, I did not save this roast profile, but I do remember it. I got distracted and did not turn down the pre-1C heat from 30% to 20% where I should have and the BT ran away a little. Not badly, no big RoR hump, but I dropped at the in-between temp of 395* in 20% DTR. I did not think anything of it. It was certainly within the range of what had previously produced great coffee.

But something developed at 395* that was not there at 390 or 400 that was so offensive I dumped 4# of brew coffee and tossed 4 bags in the garbage.

I have not read anything about this phenomenon as yet. I believe Jim touched on something similar in one of the light roast threads.

Anybody have a similar experience?

#2: Post by **EddyQ** » July 13th, 2019, 10:29 pm

I cannot say I ever dumped a batch. But I also don't roast much volume for customers and have avenues for sub-part coffee that mitigate the trash.

Is it possible the beans got a bit scorched? Do they show signs of tipping? A profile like that should produce some nice sweetness if the coffee has some. But you said heat was left on a bit too long when MET and drum temps are likely the hottest.

Other things that seldom amaze me is how off flavors can be just after roast. The next day or so can be so drastic that I wouldn't think it was the same roast. But many beans don't do this. Often right after roast, most beans lack some sweetness and can have a baked and/or more roasty taste. But with rest, these flavors often go away leaving some wonderful sweetness and flavors with no baked flavors. But I recon you have experience that way tops me with this sort of thing. But I thought I would caution you with the chance some rare bean may be messing with you.

July 14th, 2019, 1:24 am

That's the weird thing; nothing drastic changed from my normal roasts.

I roasted a replacement batch tonight. I took this one all the way to 400* for my "medium" roast. (It's actually closer to light than medium on the Agtron scale)

I've gotten pretty good at controlling my roaster. As you can see, I'm 0.7*, total from dry to drop, off my template roast. It's hard to see, but there is a pattern roast behind the current roast.

Even with my evil Sumatra bean, and all that goes on with that roast plan, I'm pretty on the money for replicating roasts:

So this is not a matter of tipping or scorching. It's about a foul chemical compound being non-existant at 390* and 400* drop temp, but present in spades at 395*. All roasts dropped at the same time and with the same 20% development time,

#4: Post by **crunchybean** » July 14th, 2019, 8:37 am

The amazing world of chemistry.

July 14th, 2019, 12:54 pm

I brewed a pot of the above roast this morning and it was delicious. I will try the 390* roast next time.

I feel I should try and duplicate the funky roast, but in order to remove as many variables as possible, I would need to use 8# of coffee. I hate to do that. Maybe it's time to fire up the Cormorant and try a few test roasts and see what I can figure out.

#6: Post by **N3Roaster** » July 14th, 2019, 3:19 pm

I've been talking about this kind of dramatic sensory step change in classes recently and touched on that a bit in the May 13 episode of Coffee and Code (though in this case picking on a raw coffee flavor that's become depressingly common). Some of these, like that raw coffee flavor are highly correlated to end temperature and largely universal regardless of coffee or roasting plan. Those are fantastic to find because once you get good at picking the flavors out in the cup they become reliable places to compare when attempting to match a roasting plan across different machines (I've recently done some work replicating my pre-purchase sample roast from the Diedrich IR-1 to an IKAWA Pro V3 and those sensory cues have been highly useful in fine-tuning that). Others such as the one you've found here, tend to be present only on some coffees and those coffees can often be great at other end points. One of the many reasons I still do a progressive roasting exercise with every coffee that comes in is precisely to find and avoid this sort of unpleasant surprise.

July 17th, 2019, 8:03 am

This one definitely seems to be tied to drop temp. But I have never experienced this before, at least not since getting a handle on controlling RoR. I had been operating under the assumption that all I needed to do was get the RoR right, eliminate the associated roast defects and everything would be fine. Not so. Big time not so.

#8: Post by **mbenedet** » July 17th, 2019, 9:35 am

Hey there, just wanted to ask a follow up. I've read through the posts above but am still confused if one of those curves is the roast in question.

I find it totally believable that there may be flavor dead zones in a coffee's post crack development, but at least for the three curves above, there may be subtle ROR improvements to be made.
The "medium roast" profile has a bit of a flat tail in the ROR at the very end.
The Sumatra has a bit of a crash shortly after first crack, and in my own experiments where I raise gas around first crack without a prior decrease, the flavors are worse. Also the ROR is horizontal for 2 minutes after that crash.
The Colombia is good, but still the ROR is horizontal (if not truly increasing a little) for the last 30 seconds.

July 17th, 2019, 11:02 am

None of the roasts is the roast in question. And the severe taint in the cup is not a result of a small RoR tail. It is not an RoR flavor defect. It was not roasty or baked. It tasted like, instead of milk, someone added rubbing alcohol to the cup.

The Sumatra is not the coffee in question. If you look at the ET curve your will see that the temp is significantly decreased before 1C. This is an extremely crashy coffee. It needs to hit 415* as close to 25% DTR as possible. Without the flat RoR end section, it would not. As you can see, I turned the gas off way before it was dropped. Not much more can be done to improve this curve. BTW, it tastes amazing. I have a 4-star restaurant as a client that picked this coffee out of my entire line up.

This is and old roast, but it shows what that coffee looks like with a reasonable increase in pre-1C temp. The curve above has maximum heat starting 30s before 1C and the RoR still dips a little. I could roast smaller batches and remove that dip completely, but it is not necessary. There is no gain in flavor in exchange for roasting less coffee.

#10: Post by **Chert** » July 17th, 2019, 11:23 pm

Almico wrote: It is great both ways.

I did not know what "tweener" is. Apparently a basketball player able to play two positions well or a person between teens and twenties can be called 'tweener'.

At first I thought the title must refer to a roast that tanks.