HomeRoast Digest

Topic: A twisted tale of three roast profiles - long (9 msgs / 397 lines)
1) From: Jim Schulman
Hi All,
This is long, but I hope interesting. I also posted it on ac; but it's really 
more appropriate here. Use a fixed pitch font to read the tables.
Jim Schulman
Quick riddle: Suppose one roasts the same beans, on the same roaster, for the 
same time, finishing at the same degree of roast. Would they taste the same? 
Not necessarily. The roasts can have different profiles; that is, while they 
start and end at the same time and temperature, the curve of the temperature 
rise during the roast can be different for each.
I wanted to find out how differently such roasts taste. I also wanted to put my 
new roasting modifications (toys?) through their paces. The set up is an FR+ 
with a variac on the heater, constant fan, and a thermocouple measuring exhaust 
air. The measurement seems accurate enough, the first crack running from 370F 
to 410F, and the first pops of the second coming at 435F to 445F, with a 
rolling second 5F to 10F higher. Roast batches were 85 grams each, this batch 
size scorches and roasts unevenly if begun under full heat, but develops evenly 
if the heat is ramped up slowly.
I picked three profiles:
The first is a dead straight line of 20F per minute rise from 330F to 450F, 
which I label "straight". Straight line profiles are advocated by, among 
others, Rusty Staub. Barry's published roast curves are fairly close to a 
straight line; (hey, he's roasting 20 odd pounds, not 3 ounces, so cut him some 
slack ;). Commercial drum roasters in general seem to have fairly straight 
profiles, especially when compared to airroasters
The third profile is one I usually use, and is called the "Blackbear Profile" 
on the roasters list, since the Blackbear Roasters in Vermont (famous for their 
banned charbucks blend) offer a good description on how to pull it off. I'm 
sure there are many other roasters who do something similar. I label it 
"convex" because the roast starts out with a steep 35F per minute spurt to 400F 
then flattens out to 10f per minute. An airroaster coming to a stall at the 
beginning of the second crack would have a profile similar to this.
The second profile is intermediate between straight and convex, labelled 
"medium;" it approximates the medium convex profiles published for airroasters 
like the HWP. I use it as a control. If the coffee from this roast tastes 
intermediate to the other two, the effect of straight versus curved profiles 
should be easy to systematize. If it doesn't; coffee roasting gets more 
My first atempt went sour (literally), since the straight profile required a 
10F - 15F higher temperature to show the same degree of roast as the convex 
one. This could be an artefact of my temperature measurement technique (exhaust 
air), but I don't think so. The more convex a profile is, the higher the 
average roasting temperature. I think it makes sense that the degree of roast 
depends not just on finishing temperature, but also on the average roasting 
temperature (how roasted would a bean be if it were raised instantly to the 
finishing temperature?).
The Three Test Profiles
Time (Minutes)   Straight   Medium    Convex
Start to 0        4 to 6 min warmup to 330F
         0          330F      330F      330F
         1          350F      360F      370F
         2          370F      385F      400F
         3          390F      400F      410F
         4          410F      415F      420F
         5          430F      430F      430F
Target   6          450F      445F      440F
Narino   5:45       445F      442F      435F
Harar    6:10       452F      455F      458F      
Average (4-10)      390F      395F      400F
The warmups to 330F, (faint grassy odor, but just prior to any discernable 
roast odor) took about 4 minutes for the Harrar, and 5 1/2 minutes for the 
wetter Narino, and were simply done by setting the roaster to 85 volts, then 
100 volts. The timing was dictated by my plan to start the business end of each 
roast with a very even tan color, thereby ensuring an even roast throughout. 
(OK, I was real impressed by how even Z&D is early in the roast)
The actual roasts were always within 3F of target during the first 5 minutes of 
the profile. The final minute of the roast, above 430F, was done by ear, eye 
and nose to get to the very beginning of a rolling second by minute 6; with the 
target and actual times shown on the table. The Narino behaved as I expected, 
with lower finishes for the more convex roasts, but at a slightly lower crack 
temperature than my previous effort with some blends. The Harrar was a 
surprise, getting to a rolling second at around 450 no matter what the profile. 
On the convex ones, I overshot, since I had to crank the voltage to catch up. 
But there were no visible differences in the degree of roast,probably because 
the temperature overshoot only lasted a few seconds.
The straight profile requires a gradual ramp up from 105 to 112 volts; while 
the convex profiles require an initial shot of 115 volts to initiate the steep 
climb, then 105 volts to brake it, then slowly back up to around 112 volts to 
I just got a new order in from SM, and decided to use two of these for the 
I picked the MAO Harar to see what happens to the origin flavors, since they 
are nice and complicated according to Tom's rather enthusiastic notes; "The 
aromatics are more vivid than any other [Harar sample]. The dried apricot 
fruitiness is there but varies from cup to cup (as is true with all natural 
dry-processed coffees from Africa-Arabia). It finishes with a light chocolate 
flavor. There's even a bit of jasmine in there. But this lot (3174) is mostly 
about a raw honey flavor. The aromas as you grind the coffee are downright 
addictive. It can literally fill a house!" (BTW: The current SM Harar, also 
highly rated, is not from MAO)
I also picked the current Columbian Narino ("Holland Vintage"), because the 
last lot (Narino del Abuela) had a powerful caramel/malt flavor which, if this 
one has it, will make it easy to guage the effect of the profiles on roast 
Since my speculation was that the straight profile would favor origin flavors, 
and the convex profile, roast flavors, I blind cupped so as not to fool myself. 
I tasted each roast in a regular cupping, and as espresso. For the espresso, I 
scored only the idiosyncratic factors that don't show up in cupping: crema, 
mouthfeel (of emulsified oils as opposed to just body), balance (of sweet, 
bitter and acidic), and aftertaste (over a longer time period than cupping 
finish). I cupped the Narino roasts first, after 24 hours rest, then the Harar 
roasts, after 48 hours. 
The Narino was surprising in two ways. First, this isn't the Abuela; no caramel 
or malt, just the basic coffee nut roast flavor with some hints of milk 
chocolate, allspice, and a non-descript, maybe mango or melon like fruitiness. 
In short, a nice, but nothing to write home about, coffee. Second, the roasts 
were a lot more varied than I expected. I'm typing up my tasting comments prior 
to peeking at which was which.
Sample B was the star of the cupping with the best developed mango and milk 
chocolate notes, along with a subtle hint of allspice, a nice buttery body, and 
a warm all-round pleasantness (hence the +1 correction). Sample A was a generic 
forgettable coffee with a very fugitive hint of chocolate. Sample C was a 
diminished caricature of B, with the same notes, slightly weakened, and with a 
rough, unpleasant edge to every category, which got it a -2 in the correction.
Category:                A                       B                      C
Cupping:  aroma/10       5                       7                      6
        acidity/10       3                       5                      6
         flavor/10       3                       7                      5
           body/10       5                       7                      6
         finish/10       4                       7                      6
      correction/5       0                       1                     -2
    (add 50) TOTAL      70                      84                     77
Espresso: crema/10       6.5                     7                      6.5
      mouthfeel/10       7                       8                      9
        balance/10       9                       7.5                    9.5
     aftertaste/10       6                       7.5                    7.5            
I remixed the samples prior to pulling shots, so my cupping impressions 
wouldn't bias the tasting. The crema was a mediocre thin cinnamon on all three. 
The nice flavor of B showed up again in the finish, while A was just as blah as 
an espresso as a regular cup. The surprise was that the best espresso was C, 
with an excellent oily mouthfeel and nearly perfect taste balance. B was just 
too sour for espresso, which spoiled the aftertaste, and even affected the 
mouthfeel, which didn't seem oily enough.
Now the unmasking; drum roll please: A is the convex profile; B is the medium 
profile; and C is the straight profile. This is contrary to expectation. I 
expected the concave profile to at least do better on body, and perhaps on 
roast flavors. Instead, the slightly curved medium profile seems to optimize 
almost every aspect of the this coffee, with the straight profile comong close, 
but showing rougher edges. If this result holds up (that intermediate curved 
profiles are better than extremes), roasting becomes much more tricky.
Ah, coffee worth drinking! In this case, two roasts B & C were very good and 
quite close, whereas roast A was just a shadow of the others. The tastes of B & 
C were a medium dry chocolate and that heady Yemen and Ethiopian apricot 
schnaps flavor I'm so addicted to. No sign of blueberries though. It was 
strongest in C early in the tasting, but strengthened in B as the cup cooled. 
Body was medium, and acidity was all fruit, no sourness. There was the usual 
Yem-Eth minor barnyard note in the nose of B & C, but only a whisper. This is a 
fine Harar, especially in the very pure chocolate roast flavor which doesn't 
have the normal admixture of dry leathery notes.
Category:                 A                       B                     C
Cupping:  aroma/10        6                       8                     9
        acidity/10        4                       6                     6
         flavor/10        7                       8                     9
           body/10        5                       6                     5
         finish/10        6                      10                     9
      correction/5        0                       0                     0
    (add 50) TOTAL       78                      88                    88             
Espresso: crema/10        8.5                     8.5                   8.5
      mouthfeel/10        7                       6.5                   7
        balance/10        9                       8.5                   9.5
     aftertaste/10        7.5                     6.5                   8
The espresso pulls were a bit of a disappointment, with the flavors coming out 
slightly more muted and harsh than in the cupping. I may have pulled the shots 
too hot. B was sourish, A was a tad bitter, and C had a nearly perfect balance 
and the best finish. There wasn't much to chose from in the crema and 
OK, the unveiling: A was the medium profile, B was the convex profile, C was 
the straight profile. OK, I admit it, there's no sense to it.
Conclusions? You've got to be kidding; to me these results make no sense at 
all. The roasts are entirely different, yet there's no pattern I can discern to 
the differences. If this holds up, picking a roast profile becomes an exercise 
in voodoo - with each coffee reacting differently, and intermediate profiles 
not necessarily having an intermediate taste.
I was particularly careful, both in the roasting and cupping, so I don't think 
these results are because I screwed up somehow. 
I would certainly be interested to hear from others with setups suitable for  
vaying profiles. Maybe more results will shed some light on this.
homeroast mailing listhttp://lists.sweetmarias.com/mailman/listinfo/homeroast

2) From: Mike McGinness
From: "Jim Schulman" 
Great work Jim. I've read it through half a dozen times and am still not
ready to really comment. It is nuts isn't it! Attempting to find the *best*
roast for a given bean let alone a few dozen (or more:) different beans. I
was thinking just last night about how Dietrich teaches going through
dialing in a SINGLE BEAN over an 8 hour time period of work. And what a
disadvantage we are at as homeroasters with virtually unlimited variety of
greens and very limited time. Sometimes too many choices. I *think* I've
gotten Kona, or I should say some Kona, dialed in very well. Yet even just
Konas vary quite a bit depending on crop elevation, early crop vs. late
crop, processing moisture content etc.
And to top it off I'm battling either a minor bout of food poisoning or
intestinal flu. Haven't had an Americano or shot or Cafe' Crema all day and
it sucks big time!!!!
MM;-) aka Kona Konnaisseur miKe mcKoffee
Dual Variable Transformer Rosto Roasting
Rocky grindin' - Miss Silvia brewin'
homeroast mailing listhttp://lists.sweetmarias.com/mailman/listinfo/homeroast

3) From: John Abbott
Oh BUMMER!  Sorry to hear that Mike - we've all been there!  Its bad enough
having your stomach at sea, but its adding insult to injury to lose your
coffee taste.  Too bad they don't make a Norton Anti-virus for us humanoid
types :o)
Here's hoping that you spring from the bed in the morning and that Kona
calls your name REAL LOUD!!  And hoping that Debbie doesn't share the bug
with you.
John - hoisting a crema in your honor!

4) From: Jim Schulman
On 7 Feb 2003 at 19:14, Mike McGinness wrote:
I'm hoping to discover some sort of simple rule of 
what profile is best for what coffee. If there is 
a rule, I'm not even close to finding it. 
Anyway, it's the best of all possible worlds: I 
get a meaty puzzle to solve, and also get to enjoy 
drinking the experiments!
Anyways, I've ordered a small Variac for the fan, 
the full-fan roasts are more even due to the 
better start, but they all seem a bit thinner 
bodied and less developed on the roast flavors 
than my previous, slow fan roasts. Having yet 
another variable will keep me busy for the next 
decade or so!
Sorry to hear it. Hope you'll be back to 24 hour 
briskets and terrific shots asap.
Jim Schulman
homeroast mailing listhttp://lists.sweetmarias.com/mailman/listinfo/homeroast

5) From: Ben Treichel
Very interesting.
Jim Schulman wrote:
homeroast mailing listhttp://lists.sweetmarias.com/mailman/listinfo/homeroast

6) From: Rick Farris
I'm sure you have wideband, but did it ever occur to you that for folks who
still have a dial-up connection, that it's a real slap in the face when you
include 31KB of quoted message in order to add 2 words?  Why not reply
directly to the sender, or else trim like I'm about to?
-- Rick

7) From: Mike McGinness
From: "Jim Schulman" 
Good luck! When you find *it* let me know:)
You know, I only roasted a very short period of time with the fan full
voltage heater variable. Put together the transformer boost wired dimmer
controller and used it ever since to keep the beans slowed down as they dry.
I find I'm backing of the fan on the Rosto well before 1st crack even
starts, beginning around my 350f. (1st crack about 400f my temps) I'm
targeting 350f in 5min then 10f per min from then on most roasts. Darker
roasts ('bout 450-455f going into 2nd) I've gone to upping the ramp speed
to 20f per min from the 450f point on. Not sure how effective yet, goal
is/was to maintain more brightness while going darker.
Thanks, I did not put the Brisket on tonight. It'll get at least another day
of rub rest... hopefully I'll feel up the Q journey tomorrow. And an AM
MM;-) aka Kona Konnaisseur miKe mcKoffee
Dual Variable Transformer Rosto Roasting
Rocky grindin' - Miss Silvia brewin'
homeroast mailing listhttp://lists.sweetmarias.com/mailman/listinfo/homeroast

8) From: Mike McGinness
From: "John Abbott" 
Thanks, I can almost taste it in my mind, but just not the same:-(
MM;-) aka Kona Konnaisseur miKe mcKoffee
Dual Variable Transformer Rosto Roasting
Rocky grindin' - Miss Silvia brewin'
homeroast mailing listhttp://lists.sweetmarias.com/mailman/listinfo/homeroast

9) From: Ben Treichel
Sorry, latest count about a T1 equivalent.
Rick Farris wrote:
homeroast mailing listhttp://lists.sweetmarias.com/mailman/listinfo/homeroast

HomeRoast Digest