Protocols For Experimental & Trial Roasting Unfamiliar Beans

Discuss roast levels and profiles for espresso, equipment for roasting coffee.

#1: Post by RAILhead »

Hey folks

Just curious to hear how you all approach roasting unfamiliar beans. I know for general cupping, one should roast light to maximize the bean's natural flavors — but outside of that, what do you do? On our 1kg MCR, our general first roast is:

400° charge temp
Dry: 5 mins or less
Mid: 3 min
Dev: around 2 min or until end of FC

So let's say you do that on an Indonesian Sulawesi and it's an excellent cup — but now you want to know what else you can do with it. Where do you guys start? Maybe a little more Mid (3:30 or so) and stretch Dev to right at SC start?

Our conundrum is that, as we learn this new 1kg roaster, we need to work on full batches as we familiarize ourselves with it — and we're trying to maximize the use of our greens as best we can. We know we're going to waste some/R&D some, but we're just trying to be good stewards.

So how do you guys explore unfamiliar beans to see what all flavors and notes you can squeak out? What do you do to make yourself feel you've done all you can to get the best out of the bean?

Thanks for any input — y'all are awesome!


#2: Post by N3Roaster »

The light roast for cupping (when used as part of a sample roasting protocol; it's important to distinguish between this and other uses of cupping methods in training, product development, and QA functions) isn't about maximizing the bean's natural flavors. That's a common myth. Rather, it's about not masking defects and getting enough information about potential flavors to make smart buying decisions. In truth, as you roast a coffee darker, you're going to diminish some flavors while further developing others and the decision that you make about where to end the roast comes down to the specific flavor results you want to achieve, which should vary depending on the coffee and on the preferences of the person who will end up drinking it.

If memory serves, you should have a large enough trier that for full batches you should be able to pull a bunch of samples (enough to get 1 cupping bowl each, you'll probably need a few trier fulls and it takes some practice to get the coordination down) during a single roast and thus get the full range from very light to very dark to taste. That's the fastest way to gain sensory experience with new coffees. Only after you've decided on roughly where your degree of roast should end up should you worry much about the timing of how you get there. In other words, the decision on degree of roast is going to be responsible for the vast majority of the differences in flavors you can get out of the coffee and the timing of how you get there, while still tasteable, is responsible for substantially less of the potential flavor variation.

In the case of coffee from Sulawesi, these are often quite boring at light or medium roasts, but you push them out to around a 30 on the Agtron Gourmet scale (substantially beyond SC) and they can develop neat sweet spice flavors along with a good thick body.
★ Helpful


#3: Post by RAILhead »

Thanks for the clarity — yes, cupping roasts to not hide defects.

And also, I never thought about pulling from the trier for different roast levels throughout the roast...

User avatar
Supporter ♡

#4: Post by yakster »

There's a good Neal Wilson @N3Roaster video showing this.

LMWDP # 272

User avatar
Supporter ♡

#5: Post by Almico »

Pulling the trier that much on a small roaster will really disturb the roast. Sucking that much cool air into a small roaster is probably not a good idea.

I would see just how small a batch you can roast and still get good profile data. I can easily roast 1# of coffee on my 10# roaster and 1/2# with some care. If you can't roast 1/2# batches or less with good data, you might want to consider moving the probes.


#6: Post by N3Roaster »

The cool air thing can be an issue on some machines, especially on smaller machines, but at 1Kg and larger that concern is often overstated. Especially if you can back off on the airflow, it's often not the big deal you might expect (small inlet hopefully outside of the primary airflow, covered by a much larger flowing mass of coffee, you can use a strip of paper or a lighter to get a rough sense of how much air movement we're really talking about on your specific machine and if it's a recent enough MCR you might have a gauge and fine control over that). Given the value of the exercise, I wouldn't preemptively discourage someone from trying and making that determination for themselves.

The video yakster mentions is probably this one: ... a78155139b and the technique is also used in the Roast Profile Development videos at: ... ing/videos (Disclosure, I'm the same Neal/N3Roaster who made those and the videos are also available and easy to find on YouTube if you'd rather watch them with ads)
★ Helpful

User avatar
Supporter ♡

#7: Post by Almico »

Pulling the trier on my 5kg roaster, even for 5 seconds, causes a significant dip in my ET and even translates to the BT dropping a bit. Pulling it over and over again changes my roast curves dramatically. The good news is it comes in handy to help prevent a flick at the end of an over-energized dark roast.

FWIW, my roaster has damper operated airflow and I keep it at 50% open which results in max heat transfer on my roaster and sucks enough air at the trier to bend a lighter flame, but not extinguish it. I do not find it necessary to change airflow at all during a roast in order to produce great coffee.

As long as the OP has logging software, they can see if using the trier causes a downward enough turn in the RoR to effect the reliability of the subsequent sample pulls. If not, great. But I have $5 that says it will.


#8: Post by false1001 »

Going to have to agree with Almico here. If you find pulling your trier doesn't affect your temp then you're probably not utilizing airflow correctly, or something else is off. At the very least you'll have a decent change in humidity within the roaster, which might not immediately result in temperature fluctuations but will definitely change the roast.

I generally like to hit beans with a lot of heat early and keep it relatively hot throughout. I'd rather race to FC in 7 minutes than try to prematurely fiddle with things and end up with a prolonged roasty result that tells me nothing about the bean character. I've found the acidity you get in those fast and bright roasts can be very illuminating as well. There's a very obvious difference between a roast that's way too acidic but has potential, and an irredeemable battery acid roast. The trick here is to ensure you're fully developing the beans, which can be difficult on more traditional drum roasters (my roaster's main source of heat is convection). I try to shoot for 14-18% dev, anything more and you risk hiding defects. I generally ignore my curves on these roasts, and while I do try to stay constantly declining my RoR ends up looking a lot more like a slightly tilted horizontal line than a Rao curve. A fantastic side benefit of hitting these cupping roasts with so much heat is you really get to see the thermal momentum of the beans.


#9: Post by badperson »

I don't have an official roasting routine I do, but with a new bean I'll want to start out with basic Scott Rao type curve; descending delta BT, hot start and keep DTR in the 25% or below range.

Then I'll try doing a really fast one, hit hard with heat for a short roast with more acidity and then stretch it out a bit more. The batch size, charge temp, drop temp, weight loss% and time are key variables I usually compare, and I typically try to play around with those.

Earlier this year I got in the habit of a good cupping routine, which did more for my roasts than anything.


#10: Post by N3Roaster »

It's very easy to imagine theoretical reasons why an exercise like what I've described won't work (and different ones every time this topic comes up!), but have those who are objecting actually tried doing the exercise as shown and did you not get a sense for how the coffee used changes as it progresses from a lighter to a darker roast? I never see the objection, "I tried it and because of factors x, y, z that I couldn't figure out how to mitigate I couldn't learn from the experience." That's kind of the point here: to get a good overview of what the general ballpark potentials are for a coffee you're not familiar with, and to explicitly try things that might be outside of your usual comfort zone. It's not the thing that you'd do for your day to day roasting once you've figured out where you want to take that coffee, it's probably not the best use of coffee for the typical home roaster operating under the constraints of not having a ton of the same coffee (and I'll admit that a lot of home roasting equipment in use doesn't make this practical which in no way says they're not making delicious coffee, though plenty of people here do have machines large enough to try this, including the person who first asked the question), and it's not even the last step in a product development process, but this is an exercise that I've repeated both in classes I've taught and as a starting point for product development on (back of envelope estimating) a few thousand different lots of coffee over the decades on a variety of machines (though mostly split between a couple of Diedrich roasters in my shop, a 1Kg and a 12Kg). I can only speak to my own experience here, but this has never failed to be an informative exercise for those involved.

As an aside, clearly a trier you can't use is indicative of a design flaw rather than not having a problem using the trier indicative of operator error?
★ Helpful