Sumatra Organic Roast in the IR2

When a friend gave me this Oganic Sumatra coffee to roast, he pointed out that it was “2nd lot” (not sure what that means precisely) and that some of the chaff that lines the beans was knicked and flaking off,  noting to be careful of this when roasting because these flakes will clog my filter and raise the temperature.

How right he was! The first roast I followed my normal profile for a Full City Roast but first crack occurred at 380 F instead of the usual 360-367 F onboard temperature reading. From there the temperature just kept going up and up and it went straight from first to second without passing Go to collect $200. I stopped the roast as soon as I saw, smelled, and heard the arrival of second crack.

For my next attempt I created a new roast profile with much lower temperatures. Here it is:

1) 355 / 3 min

2) 390 / 2:30

3) 425 / 3:30

4) 400 / 2 min

This worked only slightly better. I did get a nice, long first crack at temperatures between 375 and 380 F. However, temperatures continued to rise and I didn’t get as long of a stretch as I wanted before the beans approached second crack. Still, I managed to get what looks like a nice, fairly even Full City +

I’ll try again today with even lower temperature settings. Something like this, perhaps:

1) 355 / 3

2) 380/ 2

3) 415 / 4

4) 360 / 2

I-Roast 2 Updates

I’m getting a pretty good feel for the IR2 now. I’m starting to get some consistency from the temperatures + settings (it’s also not as cold out) and have been able to get very close to my target roasts most of the time.

Some temperature updates. A 355 F setting for the first 3 minutes should reach a max of 286 F most of the time. A setting of 400 or even 420 for a few minutes after that should bring the temperature up to 335 F, although I suspect that in warmer weather that will be something in the range of 340 to 357 F. Settings of 450 to 460 get me without fail to first crack, hovering for anywhere from 1 to 2:30 or more minutes in the range of 361 to 375 F.  However, it ALWAYS takes the machine a minute to move to a lower fan and therefore reach first-crack temperatures. This is if the previous setting is below 420 F. It may be different if the previous were set at a higher temperature. Once the IR2 does shift to a lower fan speed it takes very little time, generally speaking, to get in the range of first crack.

Here’s a recent profile that worked nearly perfectly for a espresso blend (Sweet Maria’s Liquid Amber Espresso). My target was a Vienna Roast:

Stage 1: 355 for 3 min

Stage 2: 420 for 3:30 min

Stage 3: 460 for 3:30 minutes

Stage 4: 360 for 2 min

Stage 5: 470 for 2:30 min

Total time: 14:30 minutes

The only thing I will do differently next time is shorten the “stalling” Stage 4 to perhaps 1 minute or 1:30 minutes. I still worry a little about having such a lengthy (to my mind) cool down in the middle of the roast. Although there was no problem in reaching a very controlled second crack in Stage 5.

I’m drinking a cup of Costa Rican RIP (roasted in parchment) “Red Honey”. It tastes like soap! I don’t know why. The cup had soap residue? The beans are old? Roasted too dark?

Guatemala Finca La Perla (el quiché)

I roasted this to a City + and again to FC on the same day (2/21) . After resting in glass preserving jars I tried both roasts and to my surprise prefer the darker of the two. The sweetness is more balanced with mild roasty flavors that together give it a wonderfully silky yet vibrant mouthfeel. The medium body seems perfect for these flavors. This is a superb ‘balanced’, middle-spectrum cup: clean, pleasant,  not to this not to that kind of coffee. And there are a lot of flavors to discover in each sip and as the cup cools. Black fruits, especially fig, with notes of toasted grain and microbrew flavors (at about a bock or other medium-to-light darkness ale).  This is one of the more satisfying roasts I’ve had in a long time. It is also about my 3rd of 4th attempt to get this particular bean, the La Perla from Guatemala, to a roast that I feel highlights its best qualities.

IR2 Roast Experiments

Today I will roast a Rwandan decaf. Because it is decaffeinated, I need to consider the effect of not having chaff build-up during the roast. Because it is an african bean, I need to aim at higher temperatures than usual. For a Vienna Roast (at least 20 seconds into second crack), I will try the following:

1) 355/3

2) 400/ 1:30

3) 460/ 4

4) 360/ 3

5) 470 / 3

total roast time: 14:30 (maximum is 15 min on the IR2, so I have 30 seconds to increase or decrease the time of Stage 3).

Another experimental profile I’m planning is one that gets me to first crack more rapidly (more like the air popper). I’ve been reading around and finding comments to the effect that this might develop more of the bright flavors in the cup. I’d like to try a faster first crack with City/City +  and with Full City/+ Roasts as well. Here’s one possibility:

1) 355/3

2) 400/ 1

3) 460 / 3

4) 360 for City/+ and 400 for FC/+, both at 3 min

Panama Carmen Estate 1800+ Meters–lighter roast

I just roasted the Panama Carmen Estate again. The current roast I have of this same bean was roasted darker than I anticipated or wanted. After tasting the malted barley notes that Tom (Sweet Maria’s) describes in his review, together with the overall body and mouthfeel, I felt that I had gone beyond City roasts into Full City territory and I wanted to see if I could tweak the IR2 setting to get a nice City + Roast this time.

I had less beans to roast (that’s the last of them!), about 3/4 cup or less than 145 grams. (Last roast was 165 grams).  Here’s the curve I used:

1) 355/ 3 min

2) 400 / 3:30 min

3) 450 / 3:30

4) 360 / 3:30

I think this curve would have worked absolutely perfectly if I had not added 30 seconds to Stage 3 at the last moment. I wanted to see if I could continue the first crack that had established itself nicely within 1:30 of Stage 3 at around 360 degrees (the temperature seestayed at 360 most of the time but sometimes went down to 356). It didn’t work. Cracking stopped, as far as I could tell, and the temperature went up to 371 and the bean darkened rapidly and a new, burnt smell came up: carmelization. I think with those extra 30 seconds added to Stage 3 I entered Full City territory again!

nonetheless, I”ve put the two roasts side by side and examined the beans and this one (this morning’s) is definitely a touch less smoothed out and it appears to be less expanded. At this stage of a roast it is difficult to talk about lighter or darker as the difference is just so minimal, it seems. To my eye they look about the same in color depth, at any rate. The main difference is in the smoothness.  Once the beans have rested we should be able to taste it in the cup if indeed this is a lighter roast.

2.20.2009

Okay, the beans have rested.

Dry fragrance is nutty and marvelously aromatic. Wet aroma, however, is rather musty, not too pleasant. This is definitely a lighter roast, as there is more sweetness and a bright, faintly metallic note. The malted barley flavor is still there, especially when the cup cools down. When warm, there is a faint hint of milk chocolate in the finish and aftertaste. Very mild bitterness, but not fleeting, and more pronounced in the cold cup than the warm.  On the whole a superb coffee, in my opinion much better roasted to this degree than at the previous roast level. We’ll call this one a City + Roast because I practically stalled the roast after a nice, drawn-out first crack and used the remaining minutes (more than 3:30 minutes) to allow the roast to even itself out. In fact, these beans might even be better with a slightly lighter roast. If I had more I’d try it again but without the extra 30 seconds on Stage 3.

IR2 City + Roast Curve

Here’s an IR2 Light Roast curve that I’ve been very happy with for several roasts now. It does very well for City + to Full City roasts with a variety of different bean types (high altitude, Brazilians, Central American peaberry, African peaberry) by adjusting the amount of time in Stage 3 and at times finishing the roast early.
Stage 1: 355/ 3 min
Stage 2: 400/ 3 min
Stage 3: 460/ 3:30
Stage 4: 400/ 3:30

I usually add 30 to 45 seconds to Stage 3 and get a darker, more even color. I haven’t been able to extend first crack beyond about 2 to 2:30 minutes, however. Stage 4 stays pretty high in temperature, so sometimes it may be necessary to cut that short if you want a lighter roast. This evening, for instance, I roasted about 165 grams of Panama Carmen Estate 1800+ meters. I added approximately 45 seconds to Stage 3 but didn’t hear any additional cracking. I did get evening out of the color and depth and a nicer roast smell in those additional seconds, I believe. First crack was nicely spaced, leisurely, and lasted a good 2 to 2:30 minutes. Temperatures hovered around 361-367 for most of Stage 3 (it takes about a minute to 1:30 to get to that temperature). Adding time to Stage 3 usually means more time at temperatures around 380, however, which won’t help to extend first crack but will smooth out the beans and give more color depth. Stage 4 continues that movement of greater smoothness of the beans, eveness and depth of color, and takes you farther from the grassiness of the lighter roast. By using Stage 4, however, you will proabably always take the roast to at least the edge between a City + and a Full City Roast. If the City + you want is closer to a City Roast, then I suggest not adding time to Stage 3 even if you think you can get an extended first crack, but lowering the temperature setting of Stage 4 to 360 degrees. The extra time should help to diminish any grassiness without, hopefully, darkening the roast beyond a City + Roast. Again, I should state for the record that I’m doing this without a thermocouple with probe to give me a more accurate temperature reading of the bean mass. I’m just using trial and error, sight, sound, smell, and the combination of IR2 setting temperatures with onboard temp. readings that give me a pretty good indication of where I’m at (for instance, first crack almost always occurs with onboard temp readings around 361 F. Although sometimes I get these surprising roasts where first crack starts at 351 or at 385!)

Sumatra Blue Batak peaberry in IR2

I’m drinking a cup of Sumatra Blue Batak peaberry that I roasted yesterday afternoon in the IR2. I probably could have waited a little longer to try this. There wasn’t much pop from the gas when I opened the jar.

Before I say anything about the taste of this particular roast and bean, I should describe what happened with the roast yesterday.  I put in about 165 grams at the following setting:

Stage 1: 355 for 3 min

Stage 2: 400 for 3:30 min

Stage 3: 460 for 3:30 minutes

Stage 4: 400 for 2 min

Stage 5: 360 for 2 min

 

My goal was a nicely drawn-out FC or FC+ Roast. However, similar to other roasts of peaberries in the past, this bean does not produce a lot of large chaff during the roast. What it does produce instead  is a ton of tiny chaff pieces that stick to the top screen of the IR2, allowing the machine to develop higher internal temperatures. At least that is how I explain the higher and faster temperatures reached in yesterday’s roast. First crack began already in Stage 2 with 361. When that happened I knew I needed to reduce the time of Stage 3. I took off 45 seconds but should have taken off much more than that (ideally all of it). If I had done that I would have been able to hold first crack as long as possible and still prevent the roast from rolling right into second crack, which is what happened midway into a shortened Stage 3. Unfortunately, once the beans reached that temperature, not even the decrease in temperature of Stage 4 (400 F) was enough to slow down the second crack. The beans very quickly became dark brown and very oily within seconds, so I stopped the roast. My best guesstimate is that this is a Full City ++ or Light Vienna Roast.

 

Taste. Not all of the sweetness is gone, but much of it is. There’s still some complexity in the cup and the roast flavors are not too dominant. I’m tasting a bit of hide, toffee caramel, BBQ sauce, delicate bittersweetness. Body is pretty big and flavors are slightly assertive. Mouthfeel is gritty, full, and fatty like whole milk. Aftertaste is mostly roasty tones with teh hide, toffee and sauce. It’s a somewhat “heavy” coffee, like a good Bordeaux described as fleshy.

 

Not bad at all, but I don’t want to drink of a lot of this. A bit overwhelming in low-toned flavors. There’s nothing bright or high-toned to this cup. For sure it would be better at Full City Roast.

 

 

City & City + Roast Curve for iR2

Today I tried a new iR2 curve I designed for City and City + Roasts (Cinnamon and Light Brown I believe are the SCAA classifications). My roast is probably a City + Roast. I roasted about 165 grams of Blackburn Estate Grade AA Tanzanian. They produce a lot of large chaff. I got a slow, even first crack that lasted about 2:30 minutes. Roasting in my cold garage may have kept the temperature down lower (and given me a longer first crack) than it would have at room temperature. Or that could be the density and size fo the bean, how much chaff they produce. Hard to say. Here’s the setting:

1) 355 for 3 min

2) 400 for 3:30

3) 460 for 4:30

4) 360 for 3:30

This curve seemed to work very well. I will, however, change Stage 4 next time, as I think 3:30 at 360 doesn’t really do much of anything for the roast. At least not anything different than the machine’s 4 minute cool-down. I have it there because I want to give some extra time after first crack just to make sure that the beans are even, that first crack is fully developed, and take out any grassiness that might be there at such a light roast. I’m not sure this is necessary, however. Nonetheless, in the spirit of experiment, I will change Stage 4 to 400 degrees next time and see if it makes any difference. It may be that if I want a City Roast and not a City + I’ll have to do more than just shorten Stage 3. In other words, it’s possible that at 400 degrees Stage 4 allows the beans to get darker and go beyond a City Roast, perhaps even beyond a City + Roast and all the way to Full City.  We’ll see. Until I get a thermo-probe in there, I will not be able to be too precise in determining the roast level anyway.

IR2 profile experiments

Last night I roasted the same amount of the same beans as the day before. The first two stages of the profile were the same. The second roast, however, did not even reach first crack, whereas the first went a bit too quickly and firmly into second crack.

Why such a drastic difference between the two roasts? Ambient temperature. In the first instance, I roasted outside in the sunshine with an ambient temperature above 72 F. By contrast, last night when I roasted in the unheated garage the ambient temperature was somewhere between 60 and 65 F. Apparently that is enough to have an enormous impact on the roast.

When I roasted outside during the day, the onboard temperature reading at the end of Stage 1 (set at 340) was 325. Last night the roast reached that temperature a full minute later, in Stage 2. The highest onboard temperature, at the end of Stage 2, was 356, whereas for the previous roast the onboard temperature was 385  within 2:30 minutes of Stage 2 and first crack was popping away at a regular, leisurely pace.

I might have learned something very important about roast profiles with the IR2 from this failure. First, I now know from the Hearthware website that the fan for the first three minutes will always be high and the temperature (onboard?) will not exceed 355. Secondly and more significantly, I learned that establishing Stage 2 as the goal for achieving first crack to be followed by a temperature drop in Stage 3 in order to extend the time of the roast (regardless of the roast goal) is simply too risky. Why is it too risky? For one thing because adjustments to the time of the roast cannot be made with Stage 2. (Apparently, increases or decreases in roast time DURING a roast can only be made during Stage 3. However, Hearthware’s instructions on this point are not very clear). If, as happened last night, the temperature simply does not get high enough in Stage 2 for first crack, then Stage 3 will effectively ruin the roast unless Stage 4 is set at a high enough temperature to bring the beans back around to first crack (that would be a dark roast profile, however, stepping in to save a light roast).  The opposite is also true. If first crack moves too quickly in Stage 2 there is also no way to decrease the time and move more quickly into Stage 3 and one may get a darker roast than preferred.

The solution to this problem, so it seems to me, is not to program Stage 2 as the “first crack” stage, but to use Stage 3 always as the stage designed to get the beans into first crack. Stage 2 can function instead as an extended ramp toward first crack. For my next profile experiment I have therefore decided to use a modified version of Sweet Maria’s City/City + IR2 profile. Basically, I’ll take the first two stages of that profile only slightly modified (1 min. less for Stage 1, 1 minute more for Stage 2, just because of the IR2 machines 3 minute fan setting):

Stage 1: 350/3

Stage 2: 400/2

Stage 3 in SW’s Light Roast profile is set at 460 for 4:30. That much time is not really needed, but since in the SW profile this is the last stage, it’s always good to have the extra time programmed just in case. One can always hit the ‘cool’ button and stop the roast when it has reached the desired level. But since I want to use stages beyond this to extend the time from first crack to the end of the roast (if it is to be a lighter roast) or to second crack (for darker roasts), I need to save some time for stages 4 and 5 while still leaving a buffer that I can increase/decrease depending on the state of the roast in Stage 3. If first crack is moving too quickly (“rolling”), I may want to decrease the time of Stage 3. If first crack is slow to appear, I definitely want to increase the time of Stage 3. Consequently, I’ve set Stage 3 at 3:30 minutes.

Stage 4, then, is the temperature drop designed to extend the roast time. I’ve decided to experiment with a less radical drop and go down to 400 instead of 360 for 3:30. If this temperature doesn’t slow down the roast enough, I will lower the temperature further.

Stage 5, then, should be different depending on the desired roast degree. For City, City + and Full City roasts, I probably will go to cool-down before reaching this stage. But if I want a Full City +, Vienna, or other espresso roast, this stage should get the beans into second crack nice and easy and still have enough time to extend that as much as possible, based on the desired darkness. Leaving enough time for a 1:30 buffer (for increases or decreasing time) at Stage 3, I have only 2 minutes left for Stage 5 set at 470 F.

That profile seems like it would work well for darker roasts and lighter roasts. We’ll see. It might not be optimally designed for Full City + and lighter Vienna Roasts, however. In that case it might be better to have a slightly shorter Stage 4 (“the stall”) and slightly longer Stage 5 (as much as 3:30), perhaps even slightly lower temperature at 5 to have a more controlled second crack.

The experiment goes on…. (and I haven’t even added the thermocouple probe yet!)

IR2 profiles: stretching the time between first and second crack

 

I’m still experimenting with my new IR2 (that’s an i-Roast 2). My goal lately has been to get as long of a first crack as possible and to stretch the time between first and second as much as possible. To achieve this, I’ve started paying more attention to the IR2’s fan speed. I’ve noticed that the fan has at least three speeds: high, medium, and low. The fan appears to change speeds based on the temperature settings. I hope that with enough data I will eventually have a good picture of how the fan speed affects the temperature, when the speed changes, and how fast the temperature will subsequently rise or lower based on the fan speed. Of course, there is no way to be too precise about this because it will always also depend on the amount of bean in the roaster, size of bean, amount of chaff, oil residue build-up inside the chamber, ambient temperature, and the roast curve of each setting.

I’m drinking the results of last night’s experiment now, a Mexican organic Nayarit DP roasted far darker than I wished, well into second crack (I’d say dark Vienna Roast). Brewed in the Mokha Pot, this cup has massive body and very strong dark chocolaty flavor, no bitterness or acidity, almost a chicken mole taste. Even at this dark a roast, the chocolate flavor is stronger than the roast flavor and there is some charcoal or ‘burnt’ off-taste. There’s some hint of a more pleasant smokiness, too, though, mostly in the aftertaste. It’s pretty much a one-dimensional cup, however, and that heavy, almost burnt chocolate becomes tiresome quickly. Not a great roast.
Here’s what happened to last night’s roast. I tried a new kind of IR2 setting that I got off a coffee dork email list. It’s a 4 stage setting (I had not tried 4 stages before) that looks like this:

1)   320 for 4 min
2)   460 for 2:30 min
3)   360 for 2 min
4)   470 for 5 min

Stage 1 is obviously a warm-up. Stage 2 is clearly designed to get the beans into first crack. The idea behind this setting is that once in first crack (end of Stage 2), that reaction will be stretched out longer, increasing the time to second crack as much as possible by having the temperature drop in Stage 3. Stage 4 then brings the beans into second crack. Great idea, but a very tricky proposition for the IR2 simply because if something goes wrong one cannot simply hit a button to advance to the next stage. The only control I have once a roast is in process is to hit the “cool” button and stop the roast, or wait until the last stage where I can add more minutes. At this point it should be clear what happened to this roast. Only at the very end of Stage 2 did the temperature get high enough (361) to reach first crack. But then Stage 3 came and the temperature dropped before the beans actually began first crack. That means they “baked” at a lower temperature for a full 2 and half minutes before Stage 4 brought them into first crack followed immediately by second crack. I stopped the roast with 2 minutes left in Stage 4. Total failure.

If I do try it again I’ll add 1 to 2 minutes to Stage 2 and hope that this guarantees first crack gets started before Stage 3. I’ll also drop at least a minute from Stage 4 (I rarely go darker than Vienna Roast) anyway.  Considering the temperatures of my previous roasts, I think I can also risk going with a higher temperature in Stage 1. Looking over my roast logs, I see that settings of 350 for 2 or 3 minutes in Stage 1 typically bring the onboard temperature to around 300. One roast set at 460 for 4 minutes in Stage 1 didn’t even reach first crack (got up to 331). In that particular roast, first crack temperatures were reached another 2:30 minutes later in Stage 2 with an even higher setting of 475. That means I can easily set Stage 1 to 350, 360, or even 400 for 4 minutes and see how well that prepares us for a solid Stage 2 first crack followed by a cool-down before Stage 4 second crack.

Other factors to consider if I tweak this setting and try again: I roasted a full cup of these beans; these particular beans produce a medium amount of chaff; I roasted at night in the garage where the ambient temperature was less than 68 (perhaps 61 or so).