Roast level categories: Time vs. Temperature

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
User avatar
[creative nickname]
Supporter ♡

Postby [creative nickname] » Oct 30, 2016, 12:14 pm

So I've been pondering something that I'd like to invite input on. I don't have a color meter so I've originally labelled roast levels primarily by finish temperature. Over time, I realized that this was misleading in that some coffees enter first cracks at higher or lower temperatures, and I get more consistent results by dropping at a particular interval of heat added beyond first-crack-start (let's call this FCS for short). Since then, I've tended to label roasts as follows: City roasts are FCS + 14F-20F, City+ are FCS + 21F-29F, and Full City (which I rarely reach) would be FCS + 30F-40F or so. (I'd normally expect second cracks to begin at around FCS + 36F on my equipment).

Lately I've been increasingly realizing that I don't want to hold ROR constant across finish temperatures. This is probably nothing new for many of you, and I think Jim Schulman encouraged the approach of "Slow and light, versus fast and dark" quite a while ago. But roast color and other visual indicators of roast depth obviously vary with time as well as finish temperature, so that I might have a lot of trouble visually distinguishing a roast that went to FCS + 22F during a 1:45 development window versus one that went to FCS + 15F over a 2:30 window. By temperature I'd rate the former as C+ and the latter as C, whereas if I incorporate finish time I might end up calling them both C/C+ or something else awkward like that.

Obviously, it doesn't matter that much in the end what you label things, and we can always communicate the more detailed info when sharing & discussing profiles to avoid ambiguity. And even more obviously, I suppose I could just pony up several hundred dollars for a color meter to put these sort of doubts behind me. Still, I am curious to hear from others: If you have a color meter, how do you find that its outputs match up with particular time/temperature intervals? Or if you don't have a color meter, do you describe your finish levels based on time, temperature, or some combo of both?
LMWDP #435

btreichel

Postby btreichel » Oct 30, 2016, 5:59 pm

Since I don't even have a temp gauge, I use time after FC, in conjunction with time after end of FC.

User avatar
Boldjava

Postby Boldjava » Oct 30, 2016, 6:11 pm

Most of my drops are based on time into first crack as I tend to mimic profiles or vary based more on time/profile/bean appearance rather than BT.
-----
LMWDP #339

User avatar
another_jim
Team HB

Postby another_jim » Oct 30, 2016, 10:43 pm

You have to distinguish information directly obtainable form the bean (e.g. color, taste, chemical info) from information about the roast, prep, or origin. Ideally, any description of the coffee should be based on what can be verified directly from the coffee, without added information. Right now, the best way to do this is to cup the coffee, think about it, describe it, and based on your experience characterize it. Other standard measurements on the roasted bean itself, like color or density, or even exotic and supposedly complete ones, like MS/GC data, don't do as well. This is true of almost all food and drink, like wine, tea, or other foods; it reflects the miserable state of hard science knowledge about taste.

Unfortunately, experts are very frequently wrong, and always disagree, so cupping info alone is not enough. So we add adjunct information -- where it comes from, the origin stories of the growers, the care taken in prep and shipping, the timeless wisdom of the roasters, and whatever else we can come up with. Among the whatever elses are measurements taken in the roast and other parts of the tree to cup process. When presented with these measures, we automatically assume they mean something in terms of the bean's taste, otherwise, why bother? Unfortunately, the more I know about coffee, the less I think they mean; for me coffee expertize has been an education in skepticism.

Everyone who roasts and cups knows that all the roast variables you mention affect the final taste. We even have some very slight clues in what way they affect it, but with endless exceptions and reservations. So the main point of tracking roast variables is to get reproducible roasts, not to characterize the bean's taste.

My skeptical advice: taste the coffee; keep the reproducible recipes for the roast and other prep factors; forget about everything else.
Jim Schulman

User avatar
MaKoMo

Postby MaKoMo » Oct 31, 2016, 3:26 am

another_jim wrote:So the main point of tracking roast variables is to get reproducible roasts, not to characterize the bean's taste.

I tend to agree.

User avatar
[creative nickname]
Supporter ♡

Postby [creative nickname] » Oct 31, 2016, 3:30 pm

Thanks to all, and especially to Jim for his thoughtful response. Sometimes the best answer is to stop asking a bad question, and this may be one of those times. If there isn't a solid consensus on what the categorical roast level descriptions mean when both segment time and finish temperature are varied independently, it may be best to just retire those terms as likely to cause more confusion than clarity.
LMWDP #435

User avatar
Almico
Supporter ♡

Postby Almico » Oct 31, 2016, 4:55 pm

This topic is very interesting to me, and potentially highlights the limitation of color meters.

Is it possible for two roasts of same coffee to have the same Agtron reading, when one is dropped sooner, but hotter and the other is developed longer, but dropped cooler?

Will the two roasts cup differently?

If the answer to both questions is yes, then Agtron numbers really tell us only color, not taste...right? But we don't drink color, we drink taste. I suppose in the end it's just one more tool in a roasters toolbox when trying to replicate a roast.

Some define the terms City, City+, FC based on the color of the coffee, some use development time to make the distinction. If the above statements are true, then these descriptors are ultimately meaningless. They might be able to get you into the right ball park, but in no way the right section, row or seat.

I agree that cupping ultimately determines the coffee experience, but to be perfectly honest, I can't drink/taste that much coffee!! Whatever I've been able to learn about roasting coffee has come over time and lot's of trial and error. I'm slowly learning what works on certain coffees and what doesn't.

SJM

Postby SJM » Oct 31, 2016, 5:23 pm

Almico wrote: I suppose in the end it's just one more tool in a roasters toolbox when trying to replicate a roast.


I'm hesitant to chime in here because my creds as a roaster are pretty non-existent, BUT I like the Tonino a lot and replicating roasts isn't my concern. The Tonino has given me a way identify my roasts in a way that the terms C, C+, FC, FC+ just never could. To this day I get no picture in my mind's eye when I read those terms. Some of us understand/read different data better than we read other data. If you tell me your roast came in at C+ or FC or ?? I don't get it. If you tell me you got a Tonino reading of 88 or 100 ? That resonates for me. So, I would edit Almico's statement to: "it's just one more tool in a roaster's toolbox."

SAB

Postby SAB » Oct 31, 2016, 6:11 pm

[creative nickname] wrote:So I've been pondering something that I'd like to invite input on. I don't have a color meter so I've originally labelled roast levels primarily by finish temperature. Over time, I realized that this was misleading in that some coffees under first cracks at higher or lower temperatures, and I get more consistent results by dropping at a particular interval of heat added beyond first-crack-start (let's call this FCS for short). Since then, I've tended to label roasts as follows: City roasts are FCS + 14F-20F, City+ are FCS + 21F-29F, and Full City (which I rarely reach) would be FCS + 30F-40F or so. (I'd normally expect second cracks to begin at around FCS + 36F on my equipment).

Lately I've been increasingly realizing that I don't want to hold ROR constant across finish temperatures. This is probably nothing new for many of you, and I think Jim Schulman encouraged the approach of "Slow and light, versus fast and dark" quite a while ago. But roast color and other visual indicators of roast depth obviously vary with time as well as finish temperature, so that I might have a lot of trouble visually distinguishing a roast that went to FCS + 22F during a 1:45 development window versus one that went to FCS + 15F over a 2:30 window. By temperature I'd rate the former as C+ and the latter as C, whereas if I incorporate finish time I might end up calling them both C/C+ or something else awkward like that.


Our ability to accurately describe a roast well and succinctly certainly seems a bit limited, doesn't it? You asked the question in a much more eloquent and effective fashion than I did a year ago in this thread.

What's your assessment of this roast level?

Still, there were some helpful comments there.

My approach at this point is to use time AND temperature. At FCs, I calculate a target time to drop (20-25% development), AND determine a target temperature. Sadly, Im not always able to achieve both simultaneously, so then I prioritize one over the other. If I come in underpowered, and my target temp doesnt get reached in my target time, I'll generally drop at a lower finish temp. If I'm fast, I don't stay the whole time.

Like you, I think the nomenclature for roast level is a function of finish temperature in relation to FCs. How they taste is GREATLY altered by how they got there.

User avatar
another_jim
Team HB

Postby another_jim » Nov 01, 2016, 4:44 am

In terms of getting reproducible roasts; the mega-roasters (Nestle, Maxwell, etc) are increasingly using chemical sensors to end the roast. Obviously, they use super precise profiles; but batch to batch variation on the green beans makes the results not entirely identical. The sensors "smell" the level of two gasses, one that declines as the roast proceeds, and one that increases. The idea is that the changing ratio of the two gasses precisely correlates with degree of roast, more than does temperature, time, weight, or color.

Details have been sketchy, since everything about it except the bare idea is a trade secret. So I have no idea which gasses, what levels, etc. are used. I also have the impression that it's still a work in progress; so if you happen to know how to design highly specific real time chemical vapor sensors; you can get a cushy job with the big coffee roasters.
Jim Schulman