This question involves going close to, or into, second crack, so I realize it's a bit different than most of the normal discussion around light roasts, but I've been curious about darker roasts for a while.

Obviously the always declining RoR curve dominates modern roasting discussion, but I'm looking for more discussion on the "S" curve. I've seen the general shape - BT begins to flatten during FC and pitches back up on its way to second crack before slowing back down - but there's no discussion I can find around best practices. How much should you let the RoR slow down into first crack? How fast should you let it climb after first crack? Should you slow it when you reach your drop temp or just let it go? Is there general guidance on expected flavor outcomes of running it one way versus another? These are the types of things that come to mind that make it hard for me to find a starting point.

It would be great if you could post some examples of the graph that you are talking about.

It sounds like what you are describing is that the ROR falls, then increases, then falls again.

If you were stopping in first, with the ROR dropping and then spiking up again, for me this would be a "crash and flick" and my roast would probably have good origin aroma, but with a charred character superimposed on top of it. That's sort of the orthodox view on it at any rate. Sounds like you are saying that you would then keep going from this point. I don't roast dark, so I don't know, but I can't imagine the flicked roast character disappearing with longer roasts. I don't know if dark roast afficionados regard this as a desirable characteristic or what.

You can roast dark without a flick. I'll try and see if I can find a profile. It's all about 'more' heat and the ability to control it.

How to roast darker

I've found that Neal's approach works. I've linked directly to it.

luca wrote:It would be great if you could post some examples of the graph that you are talking about.

One of the few examples I can find is from Neal Wilson (N3Roaster), maybe he'll see this and chime in here



I guess I'm just a skeptic of the universality of the Rao language of predicting off-flavors based on visual crashes and flicks. I'd rather experiment and exhaust different approaches before ruling something out based on conventional wisdom. Though, I probably should set up some more controlled experiments of going dark without flick, I haven't really done that in a while

drgary wrote:How to roast darker

I've found that Neal's approach works. I've linked directly to it.

Yup! Seen his stuff before but I still wish there was more quantified detail. I'd love for him to release his book. In the meantime I need to do some more experimenting. The hard part is not knowing whether I actually have the palate for this level of roast. I'm always doubting whether I roasted it bad or if I just don't have the taste for it.

I roasted this one yesterday in this sort of scheme. 325g of East Timor Ermera from Happy Mug in my Huky. Brewed some up tonight to taste it and... it was drinkable, especially the more that I drank it, but not something I'd want to aim for. I didn't think it was overly bitter but had some of what seemed like bitterness stick on the tongue. Tasted heavy, like it would be a lot to drink a full mug of, could probably use some acidity to make it more drinkable.



I just don't know where to start with manipulating this. So many things I could change to try to get a different outcome

The idea is to preserve flavors and characteristics like sweetness and acidity inside the bean developed earlier while developing the outside of the bean for darker roast flavors, like disillates. Neal says that the rising BT ROR toward the end of the roast cups better in his classes than roasts to the same level with steadily declining BT ROR. We're not talking about a first crack crash that reduces sweetness followed by a flick that creates papery astringency. Consider the BT ROR increase at late stages of a dark roast as not an extreme change. I expect that the optimal curve is tuned by taste and the look of that curve will differ by coffee and the characteristics of your roaster and sensors. Added: I don't think you want to flatten the curve while approaching the end of roast, because that would go against the intention of preserving internal flavors. I would also think that adjustments to achieve increasing ROR would be small.

I typically roast dark since I typically drink espresso. What I've found when roasting dark is to reduce the heat towards the end of first crack and it avoids the "flick" during second crack.

I only take it about 10 seconds into second crack, so that isn't familiar territory to me. There are coffees I sometimes enjoy that are taken darker than my roasts.

There's a lot that really just comes down to the specifics of the coffee and your preferences, but in terms of the question of how much you should let rate decline, I would recommend looking at this more from the perspective of how much time do you want to spend prior to first crack, between cracks, and after 2nd crack. One of these days I need to finish making my roasting plan designer more generally usable as it's possible to go directly from desired range durations to an easy to replicate plan with just a few points on a relaxed cubic spline. That's applicable to S curves, non-ascending rates, strictly declining rates, or plans I'd class as weird things that you probably don't want to do most of the time but every so often you run into a coffee that benefits from it.

As for the book, it's finished aside from another editing pass, but I'm holding off until Typica 2 is out because some of the material is easier to put into practice with better tools which I want to make sure are freely available.