Roast Profiles (Write-In)

roasting & roastery operations

Roast Profiles (Write-In)

Postby Alistair Durie on Fri Aug 01, 2008 8:09 pm

This letter comes from Bear Braumoeller at Ohio State University:

My question has to do with roasting profiles. I have discovered, as have many others, that changing the shape of the roasting profile has a tremendous impact on the flavor of the coffee that is produced. Given the importance of this part of the roasting process, the dearth of available and readily digestible knowledge on the subject is quite surprising. Most practical references on roasting emphasize the degree of roast rather than the temperature path that is taken to get there. Food science references do exist, but reviews indicate that they are written by and for people with a highly technical background and are not accessible to those lacking detailed backgrounds in chemistry and possibly physics. In between lies little if anything.

What, then, are the general principles behind deciding on a proper roasting curve? What is the most useful way to divide a roasting curve into stages? What reactions take place in each stage? Which flavors (in which beans) are brought out during those reactions? When will they be eclipsed and why? How does the rate of temperature increase have an impact on those reactions? Is it possible to begin to put together a comprehensive graphical representation that would be useful -- perhaps something with temperature on the Y-axis, time on the X-axis, and the various flavors/reactions as overlapping bubbles or areas in X-Y space? (I fear this might end up being nothing more than impressionistic, but it still might be useful.)

OK, sorry, that's more than one question. But it's a complicated topic. :-) One of your members announced a set of experiments to try to answer it a few years ago -- viewtopic.php?f=16&t=715 -- but the results were never posted. I tried to write to him to find out what the results were, but I can't do that without being a member.

Thanks for your time.

Bear Braumoeller
Ohio State University
Alistair Durie
admin
 
Posts: 1002
Joined: Sun Aug 15, 2004 1:59 am
Location: Vancouver Canada
full name: Alistair Durie
company: Elysian Coffee
: www.elysiancoffee.com
: www.coffeed.com

Re: Roast Profiles (Write-In)

Postby Brent on Sun Aug 03, 2008 3:27 pm

Bear wrote:What, then, are the general principles behind deciding on a proper roasting curve? What is the most useful way to divide a roasting curve into stages? What reactions take place in each stage? Which flavors (in which beans) are brought out during those reactions? When will they be eclipsed and why? How does the rate of temperature increase have an impact on those reactions? Is it possible to begin to put together a comprehensive graphical representation that would be useful -- perhaps something with temperature on the Y-axis, time on the X-axis, and the various flavors/reactions as overlapping bubbles or areas in X-Y space? (I fear this might end up being nothing more than impressionistic, but it still might be useful.)


The general principles I would suggest is:

get coffee
roast some, try it
mess around with different roasting profiles
decide which one was nice
stick with it.

I think there is an art to it, just like pulling shots and other steps in the coffee chain, why is it done that way, it is done because it is.

this reminds me of a time in a previous life when I was doing some live mixing, and achieving (to the owner of the audio companys ears) a stunning vocal mix. He took a look at the setup to see what I was doing, and stood flabergasted at how I had setup one bit of gear, and passed the comment "I can't believe the results, it shouldn't be working so well" It worked, because I understood what I was doing, and what I wanted to achieve... which I did.

Coffee isn't really any different - the roasting process is part science, part art, and nailing that art bit is a bit like defining why the mona lisa is famous etc etc

how many roasters (I am sure plenty do) actually keep details of "failed" roast profiles, rather just work out how to roast it proper like and go from there?

just some thoughts
Brent
Brent
 
Posts: 461
Joined: Tue Aug 09, 2005 6:23 pm
Location: New Zealand
full name: Brent
company: .

Re: Roast Profiles (Write-In)

Postby Christopher Schooley on Mon Aug 04, 2008 7:47 am

Hi Bear,
Charting and graphing of roasts is pretty standard in the industry and especially for training. Time is usully on the horizontal axis with Temp on the vert. This is indeed a beyond useful tool, especially if you are graphing thoroughly. Not only should you be marking minute by minute the progress of your roast, but also all adjustments made. When creating or adjusting a roast profile it is usually more complicated than just stretching out or shortening one period of the roast. In a drum style roaster, the amount of airflow through the drum can greatly effect cup character (sure, lengthening or shortening the roast might be a result of this to a degree but perhaps not entirably measurably). Cupping and using your charts as a reference contextualizes your efforts. That being said, perhaps a fun direction for this thread would be for roasters to write in one or a few of their tactics or minor adjustments and what their results are? I know that when you keep too much energy in the drum so that the first crack runs into the second the resulting cup is the equivalent of trying to shove a brick through a pinhole.
http://www.coffeeshrub.com
Armed to the teeth with secret words
Christopher Schooley
 
Posts: 165
Joined: Tue Apr 24, 2007 6:25 am
Location: Fort Collins, Colorado
full name: Christopher Schooley
company: Coffee Shrub
: www.coffeeshrub.com

Re: Roast Profiles (Write-In)

Postby Oliver on Mon Aug 04, 2008 10:52 am

I know that when you keep too much energy in the drum so that the first crack runs into the second the resulting cup is the equivalent of trying to shove a brick through a pinhole.


Can you explain in real tasting terms what happens?
Oliver
 
Posts: 92
Joined: Tue Jun 06, 2006 5:59 am
Location: Olympia
full name: Oliver Stormshak
company: Olympia Coffee Roasting Co.
: www.olympiacoffee.com

Re: Roast Profiles (Write-In)

Postby Christopher Schooley on Mon Aug 04, 2008 11:12 am

I shouldn't say "I know", but rather in my experience rushing the first crack into the second does not fully develope the dynamics of the cup. It results in a front loaded cupping experience with strong acidity and agressive flavors all at the front of the pallete which disapate over the pallete leaving a limp and weak aftertaste and/or no "middle" notes.
http://www.coffeeshrub.com
Armed to the teeth with secret words
Christopher Schooley
 
Posts: 165
Joined: Tue Apr 24, 2007 6:25 am
Location: Fort Collins, Colorado
full name: Christopher Schooley
company: Coffee Shrub
: www.coffeeshrub.com

Re: Roast Profiles (Write-In)

Postby Edwin Martinez on Mon Aug 04, 2008 3:35 pm

Dear Bear,

I am so glad you wrote. I understand you can read, but not write on this forum. I have wondered these same questions myself the last few years. I would like to take 20 pages and attempt to respond, however it would be few facts and mostly speculation based on very limited experience.

I have been working with coffee roasters around the world, some of which I believe are defining the cutting edge of quality in coffee and I find most of them are great managers of their tools. They have good balance of art and science. They can produce great results with out understanding the details behind roast chemistry in a way that would satisfy anyone with a Phd in chemical food science. You don't have to be a good mechanic to be a good driver. You can evaluate how responsive the gas and brakes are and even test drive other tools until you find the match that best yields the results your looking for.

That being said I believe there is a knowledge gap here on a VERY complex subject.

I don't know that this could be put on a single chart. If so I'd love to see it! For every varying result one is after there is also a different path to get there. Few roasters can confidently speak clearly and specifically in plain english about navigating all the chemical processes, let alone list them. Again, I don't think this is a sign of ones ability as a roaster by any means, but there is a big knowledge gap. I think the technicians who really understand this are no where near a roaster, and if they are, their findings are proprietary company information.

Maybe we can collboratively tackle this one question at a time.

I would start with your second and third questions, as "proper roasting curve" will be quite subjective.

2. What is the most useful way to divide a roasting curve into stages?
3. What reactions take place in each stage?
Edwin D. Martinez
Edwin Martinez
 
Posts: 157
Joined: Thu Oct 27, 2005 6:40 pm
Location: Bellingham,WA/Guatemala

Re: Roast Profiles (Write-In)

Postby Jon Brudvig on Tue Aug 05, 2008 6:56 am

Man, this is kind of a scary question to try to answer. So many different schools of thought, with no clearly defined right answer, but here's a rough outline of how we profile:

When we roast samples, there is usually only enough coffee for one or (rarely) two batches. This makes it necessary to use a pretty generic profile, and we use the same one for just about every coffee, with some variation in drop temp.

When we find a sample and buy a coffee we like, we will try multiple profiles and cup them side by side. The profiles vary in total time, curve, drop temp, etc... We then decide on the best profile and try minute variations of that profile until we think we've nailed it. Our production batches are roasted on a Probat L12.

We record our profiles by writing down the bean probe temperature at 30 second intervals, along with any flame or air flow adjustments. We also record the time and temperature at which the first few hints of the first and second crack can be heard.

We look at most of our profiles as having three distinct stages. The first stage is basically warming up / moisture loss. This occurs from the time the beans are charged until the bean probe reads approx. 300F. Throughout this stage, the aromas given off by the beans are mostly green and grassy, with progressively more aroma leaving the beans until its end. A roaster can tell when the next stage is beginning by observing changes in color and aroma, which can indicate that chemical changes are beginning. The most obvious change is the shift in color from green to yellow or tan.

At the beginning of stage 1, the temperature drops dramatically as the beans enter the drum, and it usually takes a couple of minutes before any increase in temp appears on the probe. This varies greatly with batch size. At around 200-220F, the temperature climb becomes evident. From this point until about 290-300, we like to see a 30-40 degree/minute climb, depending on the coffee.

We see the next stage as being 300F to first crack. A lot more is going on here; the aroma shifts from grassy to bread-like, and the color changes from yellow/tan to light brown. Most of our profiles shoot for an increase of 16-20 degrees/minute throughout this stage. First crack can usually be heard at 385-395F bean probe temp.

The third stage is the time between first crack and completion. It is marked by dramatic chemical changes (pyrolysis & maillard), which are evident by the dramatic changes in color and aroma. First crack usually releases some of the most intense aromas of the whole process. The stage between first crack and completion is the most critical stage in the whole process, due to the rate and sensitivity of the reactions taking place.

The temperature climb is very important in this stage, and each 30 sec increment is not going to be the same. Usually it will follow 16/20 degrees/minute. Here is one example, where 0:00 = 392F / beginning of first crack:

0:30 - 400F
1:00 - 408
1:30 - 416
2:00 - 425
2:30 - 434

We would drop most coffees at or before this point, but if allowed to proceed further, second crack would be heard at around 438-448, depending on the bean.

One good piece of information is the total time between first crack and drop or second crack. If this time is too short, several things can happen, including underdevelopment of flavors/aromas and scorching. Underdevelopment is somewhat difficult to describe in tasting terms, but for me, the cup tastes lacking with poor balance, and the individual flavor and aroma components seem undefined and subdued. Scorching is much easier to identify; it typically tastes like burned puffed rice or popcorn. If the time between first crack and drop is too long, the cup can taste "baked out," with poor acidity and and no high notes, and a distinct "baked out" note that I can't describe. If the roast stalls or loses heat at this point, the baked out flavor can be more intense. According to Sweet Marias' website, this could be due to disruption in long-chain polymerization.

Drop temp is a huge part of a profile, and lets the roaster decide how much "roast" flavor they will allow in to the finished product. Here are a few generalizations:

Coffees dropped too light can have overly aggressive acidity, underdeveloped flavors/aromas, poor balance and can lack body.
Coffees dropped too dark can have unpleasant carbon notes, "baked" flavors, too little acidity, too little origin character, poor balance and can lack body.

Its really fun to "trace" the development of certain flavors within a bean through different profiles and drop temps. The currant note in a Kenya seems to change to cedar with more roast, and the blueberry note in a good Harar sometimes only really comes out in a 2-4 degree drop temp range.

I remember when I was first learning how to roast, I really wanted everything to be really straight forward. I wanted to know that to really highlight the cherry note in a good Guat, the temperature climb between first crack and 420F needed to be 9.5 degrees, and that when roasting a Harar the climb between 350 and 385 was critical because that is when the initial reactions that lead to fruity flavors are beginning. Unfortunately, roasting isn't that straight forward. Experimenting with different profiles is the best way to learn what is happening. Tasting multiple roasts with only one variable changed (drop temp, time between first and second, etc..) is a great way to start.

These are all huge generalizations, and I'm sure many other roasters follow very different profiles and have a better understanding of the chemical process. Hopefully they'll chime in.

-Jon
Jon Brudvig
 
Posts: 34
Joined: Wed Feb 06, 2008 6:04 am
Location: Sioux Falls, SD
full name: Jon Brudvig
company: Coffea Roasterie
: www.coffearoasterie.com

Re: Roast Profiles (Write-In)

Postby td on Mon Aug 11, 2008 1:39 pm

Like Edwin I am a little nervous to begin to try and answer the questions posited by Mr. Braumoeller. Also like Edwin I will only attempt to answer a few, but I will say that as a manufacturer of roast control and data-logging system, as well as a roaster myself, I have been struggling with some of these questions. Although not neccessarily for the same reasons. As I am sure that you can imagine.

First let's clear up a little definition problem that has developed as our tools have begun to outpace our roasting language. A Roast Profile is the time and temperature curve ( bean ), from point of equilibrium to drop point. It is possible, in fact, highly likely that many roasters that currently track time and temperature points manually and then draw straight lines between the points are developing multiple profiles without knowing this to be true ( this would be true of excel spradsheets as well ). This is because the actual curve does not follow a linear progression. However, it is true that the more points you track the more accurate a representation of the actual curve you will have. Why is this important? Because as Mr. Braumoeller nicely implies- the shape of the curve affects the taste of the coffee. In other words, the roast profile helps determine the flavor profile.

We have done a little work trying to determine which profile curves fit what coffees, or more explicitly which flavors. And how to draw or suppress flavors through the manipulation of the curve throughout the roast. The most knowlegeable person in this area, I believe, is Kathi Zollman of Coffee Holding. Kathi developed a cupping exercise involving 1 coffee and 3 flights involving 6 different profiles, each flight containing 3 cups and 3 distinct profiles.

Flight 1: light roast, medium roast, dark roast. Same preheats, same point of equilibrium (turning point): Radically different tastes.

Flight 2: all 3 coffees roasted to same drop temperature: 3 different times fast. medium. slow. same pre-heats, same points of equilibrium. Very noticeable differences in taste.

Flight 3: All 3 coffees roasted to same drop temperature and time, same pre-heats, same points of equilibrium, same first crack time and temperatures. Different profile shapes (linear,soft curve, hard curve). Very noticeable differences in the subtlities of the coffee involved.

Note: coffees 2, 5 and 8 were the same coffee, same profile.

Now most all of us know the first 2 flights, and practice our craft accordingly. It is what is occurring in the final flight that Mr. Braumoeller is asking about. With that in mind I would like to venture some opinions on some of these questions:

It appears, from our experiments, that most flavor development occurs before the carmelization of the sugars ( first crack ). Although this sounds counter-intuitive, remember we are not talking about roast level, or the tastes associated directly with roast levels. Only in developing what is inherent in the coffee itself.

The most useful way to divide a roasting curve is- in as small as increments as practicable- focusing on consistency of profile shape before first crack (consistency after 1st crack is important also, just not for the same reasons).

Proper roasting curves should be determined by the resultant flavor profile. In flight 3 above, 2 of the profiles ( 8 & 9) illicit about equal preference among professional roasters that have taken part in this exercise. Different roast profiles, different flavor profiles, different taste preferences. No right or wrong here, just differences. It does appear however, that the most linear of the curves leads to an underveloped, grassy coffee (what this means is that curves relying on consistent rates of rise between point of equilibrium and first crack yield the least disireable flavor results).

We are approaching a revolution in our industry in respects to the control and manipulation of flavor profiles. Like many revolutions it is driven by many factors, including: better and more defined green coffee (micro lots), better tools ( data-loggers and profiling systems ), better communication ( e-mail, web, BBs), and the interest of food scientist and chemists as well as the enquiring minds, and cupping skills, of intrepid roasters. Here is what is still needed for this "revolution" to come to fruition: more and better science, more relevant training excercises, and the development and adoption of new techniques, and finally an increased collaboration between the food science community with the roasters. We need their ability to perform quantitave analysis ( HPLCs, electronic noses, scientific discipline, etc...), and they need our ability do qualitative analysis (cupping, understanding of the market,producton realities, etc...).

So, where do we go from here? As coffee roasting professionals it is our responsibility to keep abreast of trends, digest research and to evaluate tools vis-a-vis our industry and our businesses. One way to do this is to reach out to your local University's food science, packaging science, or any other department and see if anyone is currently working on these issues and could you become involved. While researching the following article for Roast Magazine:
http://www.ambexroasters.com/pages/arti ... art_3.html
I discovered that there are many institutions of higher learning doing work in coffee. And, in fact, have since entered into a research agreement with a major University working on flavor profile issues ( for coffee as well as other products).

For starters did anybody contact Mr. Braumoeller and see what he is up to, who he is? He may well be a PHD food science candidate looking for a research project. What roaster would not want to be involved in that?
-T.D. Davis
td
 
Posts: 134
Joined: Sat Dec 10, 2005 8:22 am
Location: Clearwater, Fl
full name: Terry Davis
company: Ambex
: http://www.ambexroasters.com/

Re: Roast Profiles (Write-In)

Postby Jason Haeger on Mon Sep 01, 2008 8:20 pm

Easy enough to discover who he is:

Bear Braumoeller
Assistant Professor of Political Science

# International Relations


Professor Braumoeller's interests include the sources of war and conflict, international relations theory (in particular, systemic theories of international relations), political methodology (tailoring statistical methodology to fit the particular needs of students of world politics), and Russian foreign affairs (especially the relationship between belief systems and foreign policy behavior). His work has been published in journals such as Political Analysis, International Organization, American Journal of Political Science, and International Studies Quarterly. Professor Braumoeller is on the editorial boards of the American Journal of Political Science and Political Analysis.


As taken from a search that landed me here: http://polisci.osu.edu/faculty/braumoeller/
Jason Haeger
 
Posts: 558
Joined: Fri Feb 10, 2006 1:07 am
Location: Allen, TX
full name: Jason Haeger
company: AJ Coffee Company | EspressoTrainer.com
: http://www.ajcoffeeco.com


Who is online

Users browsing this forum: No registered users and 1 guest