Behmor 1600 coffee roaster - group taste test - Page 2

[Kuban111]

I figured out the C1/C2 myself, since I do roast at home.
But even with the explanation, I have no idea what
Stir Crazy/Turbo Oven means...

I'd be curious to hear how well the Behmor handled chaff
during these roasts. Not sure how much chaff there is
in the Harfusa Yirgacheffe. The main appeal of getting
a roaster like the Behmor for me is that it would catch
chaff, so I could roast indoors.

Hope this helps on the SC/TO explanation

http://homeroast.pbwiki.com/SCTO

Michael.

[CremaKatz]

Hope this helps on the SC/TO explanation

http://homeroast.pbwiki.com/SCTO

Michael.

Thanks. And wow.

[Matthew Brinski]

When is CCC going to open a shop in Dayton? I know someone that would be interested in a GM position there.

After they set up shop in Colorado Springs or Denver first.



.

[majorzx3]

It sounds promising that at least one batch from the Behmor was at least as good as the professional roast.

Is there a chance you guys will experiment with espresso roasts?
that would be most interesting to me....
Thanks, Zed

[mike]

It sounds promising that at least one batch from the Behmor was at least as good as the professional roast.

Is there a chance you guys will experiment with espresso roasts?
that would be most interesting to me....
Thanks, Zed

I won't rule out testing for espresso, but as I've commented on over at CG, there are so many variables with espresso preparation as compared to the careful controls you can do with cupping drip, that the experiments tend to be much less useful.

However, I don't really think it matters to tell you the truth.

From my experience, espresso will be easier than drip, so if we can establish where the shortcomings are, and what the workarounds are for drip, I think we'll be OK.

The nice thing about espresso is that most people are going to take their roasts into second crack, so (a) you will have a good "marker" for hitting the right degree of roast, and (b) that roast level will tend to help cover up any minor issues you might have with your roast.

[lparsons21]

I've never intentionally not taken a roast to at least the first snaps of 2nd crack. Well, not exactly true. I did try it during my early experimentation with the Behmor, but was mostly getting them too light and unsatisfactory. A couple of exceptions, but mostly those were ones that finished earlier than I had planned them to.

My roasts tend to just into 2nd to end of 2nd and then coast. So far, I've not gotten any bad roasts from the Behmor, and many truly wonderful, if not great, roasts. I drink mostly Cafe Cremas and Americanos, with a cap thrown in now and then.

But I can understand that doing a roast that doesn't get to 2nd could be difficult to control on the Behmor. Visibility of the bean is just not good enough during roasting imo, to see the slight developement you would be looking for. One of the things I used to watch on my Gene was the seam of the bean during roast, the color would give me a pretty good indication of what was going on. But with the Behmor you can't see it inside the roaster well enough, and the finished roasts tend to look different at the seam from the Behmor compared to the Gene. Probably because of the radian heaters.

[majorzx3]

Thanks Mike.
there are a few people on this board, that are very skeptical about the Behmor's capabilities. I personally only care about espresso roasting.
I will eventually get into home roasting, but first I want to get a really good baseline with my Brewtus. I'm using a premium quality roaster's blend to "get a taste memory".
The Behmor would fit my lifestyle best among all the other roasters out there, so I'm happy to see some good reports that it might be able to roast at least close to premium quality.
Thanks for you report!
Zed

[mike]

OK, this Friday will be the second taste test.

In the first taste test, we roasted 8oz samples of Ethiopean Yirg with each of the 5 Behmor profiles aiming for a City+ finish. Those samples were then blind taste tested along with a commercial sample from CCC, and the samples ranked in preference. Major conclusions were that (1) the taste preference was closest associated with finish roast degree, (2) getting a consistent finish roast between 1c and 2c with the Behmor was extremely hard, and (3) there appeared to be no correlation between preference (or flaws) and the profile used. Review of the roast logs indicated no overly large differences in roast times, which may have predicted that profile choice would not be a significant effect.

The next proposed taste test to try to identify the modality which causes reports of poor roasts in the Behmor was chosen to have roast load (charge load) as the variable. In order to mitigate final roast level as a variable, it was decided to use one profile (P1) and to stop each batch at the first snaps of 2C - a more easily achievable target. The bean was again the Ethiopean Yirg. Loads of 4oz, 8oz, 12oz and 16oz were roasted.

The beans were roasted last night in 67F ambient conditions. There will be adequate time for the beans to rest before Friday, and they will not stale in that time. Unlike the previous test where profile was altered with no significant corresponding effect on time, this time with roast load being altered there was a very significant difference in time to 1c. I do not have the figures here with me, but the 4oz load was in the 10 minute range, and the 16oz load was near 18 minutes. Both the 12oz and 16oz times to 1c would be considered outside the "normal and accepted" target ranges to 1c. All loads made it to our target roast level, but the 16oz just barely within the 20:30 maximum time.

Dan will again be leading the testing on Friday with a request for the group to simply cup and rate the four samples in order of preference, capturing significant comments (particularly on flaws) if they occur. Because all samples have been coded, Dan will report the results back to me and I will analyze and look for any correlation of roast weight to preference.

Stay tuned for results this weekend or next week. Based on these results we may look a bit further at profile effect on a fixed weight at a fixed roast target level similar to taste test 2, or pursue a different direction in followup tests.

[MellowCat]

Very interesting testing, and producing valuable resulting information - thanks.

Do you think the two day's rest time to Friday is enough for the taste test?
edit - sorry, re-read saw you roasted last night - three days should be ok

[farmroast]

I'm finding that the timing of the thermostat kicking in during the 100% cycles can effect the profile of a particular roast. Especially during and after first crack. Needs to be considered in comparisons and repeatability.

[mike]

I'm finding that the timing of the thermostat kicking in during the 100% cycles can effect the profile of a particular roast. Especially during and after first crack. Needs to be considered in comparisons and repeatability.

I've not noticed a large variation on when the thermostat kicks in, but I'll keep an eye out for it. In the previous and extremely long thread here there was a lot of discussion on heat application. I think the best way to measure this would be to probe the bean mass and see if there are any significant swings. Up to this point it has been my belief that there is sufficient thermal mass in the roaster and beans to smooth out any disruptions, but it has been tremendously hard for anyone to find an adequate measurement point or fully comprehend the effect of the quartz heaters as compared to normal heaters.

Probing the bean mass will ultimately be the best. I did see where somebody was able to make it work with moderate success, but the fabrication was beyond my capabilities.

The one thing I can most definitely say is the repeatability is terrible due to the type of roaster the Behmor is. It reminds me of what my friends tell me about bbq roasting. Nothing was easier than doing a whole roast in a popper with my nose in the chimney, my eyes watching the beans, and a thermocouple stuck in the top of the bean mass.

I am lucky in that 90% of the roasting I do in the Behmor is for espresso. I simply take a reasonable shot at calculating the timing for the heat reduction stage of P2, and follow the roast to the first steady sounds of 2c, then hit cool to almost always hit a perfect light Vienna. Doing my drip loads to City, City+ or FC has been quite substantially harder.

[farmroast]

I've not noticed a large variation on when the thermostat kicks in, but I'll keep an eye out for it. In the previous and extremely long thread here there was a lot of discussion on heat application. I think the best way to measure this would be to probe the bean mass and see if there are any significant swings. Up to this point it has been my belief that there is sufficient thermal mass in the roaster and beans to smooth out any disruptions, but it has been tremendously hard for anyone to find an adequate measurement point or fully comprehend the effect of the quartz heaters as compared to normal heaters.
Probing the bean mass will ultimately be the best.

I do have a bean mass thermocouple in my Behmor. I'm not surprised you had a nice roast on P5 with your 1st tests. Once you get past the first 3 min. with little temp. rise (last 2 roasts .5lb with a Brazil) 130f@3min, the pace picks up nicely and evenly but not the roller coaster ride of the 100% cycle. Did a P1 and a P5 to 7* before 2nd the P1 sucked the oils more to the surface and the roast had a fragrance right away where as the P5 looked much more even(wish I had an Agtron) and had much less fragrance. After two days rest the P5 fragrance was much sweeter and nice all around than the P1, will cup tomorrow. Most roaster are either convection or conduction or some of both but I find the Behmor to be mostly radiance.
farm

[HB]

Dan will again be leading the testing on Friday with a request for the group to simply cup and rate the four samples in order of preference, capturing significant comments (particularly on flaws) if they occur. Because all samples have been coded, Dan will report the results back to me and I will analyze and look for any correlation of roast weight to preference.

The turnout was small last week, so the cupping was me and Counter Culture employees Mark and Nathan. We all agreed the samples were roasted darker than we preferred, but none of the roasts were rejected as was the first test go-around. Sample "B" garnered one 1st place and two 2nd place votes and was the overall winner:

Table:

Sample    Rank
------  -------
  A     2, 4, 1
  B     1, 2, 2
  C     3, 3, 3
  D     4, 1, 4

We had enough left over to run an Agtron:

Table:

Sample  Low/High
------  --------
  A     37 / 48
  B     39 / 52
  C     35 / 47
  D     37 / 48

(The first measurement is prior to grinding). I neglected to collect cupping notes. Positive comments from memory included "good chocolates and high body." Negative comments included "ashy aftertaste" and "the varietal character was lost in the roast."

[DavidMLewis]

We had enough left over to run an Agtron:

Table:

Sample  Low/High
------  --------
  A     37 / 48
  B     39 / 52
  C     35 / 47
  D     37 / 48

(The first measurement is prior to grinding).

Is that the kind of difference, whole bean to ground, that you normally see, or does it reflect a bean that is roasting too quickly at the end and isn't developed in the center?

Best,
David

[farmroast]

Compared to larger drum roasters the Agtron readings seem inline. Wonder if the "ashy" aftertaste is a result of the slow rpms of the drum along with the strength of the radiant heat application. We are used to heavy conduction with variable amounts of convection roasters (various drum roasters),heavy convection roasters (poppers, Sivetzs), and medium convection low conduction (Loring).But the Behmor I find is mostly intermittent radiant with little conduction or convection. It will take some comparison testing as you are trying to be able to understand the effects of this heat application on the beans and in the cup.
farmroast

[mike]

The turnout was small last week, so the cupping was me and Counter Culture employees Mark and Nathan. We all agreed the samples were roasted darker than we preferred, but none of the roasts were rejected as was the first test go-around. Sample "B" garnered one 1st place and two 2nd place votes and was the overall winner:

Sorry, for some reason my notifications from HB have been flakey lately. I will check my coding notes tonight and report back to you all. If the coding means what I think it does, then we finally have some decent correlation.

[mike]

OK, so here are the results.

I'll adopt a simple addition of the 3 preferences, lowest score determines rank

#1: B (12oz load) 1, 2, 2 5 pts [1c 14:00, 2c 18:10] 39 / 52 (13)
#2: A (8oz load) 2, 4, 1 7 pts [1c 11:00, 2c 14:25] 37 / 48 (11)
#3 (tie): C (4oz load) 3, 3, 3 9 pts [1c 8:25, 2c 10:25] 35 / 47 (12)
#3 (tie): D (16oz load) 4, 1, 4 9 pts [1c 16:00, 2c 19:55] 37 / 48 (11)

At least from order of preference, we see some consistency with what has been reported anecdotally: 4oz and 16oz loads are less suitable for this roaster, and can have taste defects. It may be that the 4oz loads roast too fast, and the 16oz loads roast too slow. With only one outlier, both samples were consistently rated in the bottom half of the four samples.

Which leaves the 8 and 12 oz loads. Inversely, with only one outlier, both samples were consistently rated in the top half of the four samples.

I didn't see anything very interesting in the Agtron readings, except that I would have expected more of a spread difference between the shorter and longer roasts. Perhaps this has something to do with the nature of the radiant heaters.

As a point of interpretation, I can say that this Yirg bean was a little challenging, apparently harder than most of the other beans I roast. It may be that other varietals may roast better as an 8oz load, and some at 12oz. Also, we purposely chose a darker roast than normal in order to try to hit a consistent roast finish level. As is evidenced by the Agtron readings, this was achieved, and at least by visible indications was significantly more consistent than taste test 1.

I think this has been able to confirm what has been anecdotally reported: the sweet spot of this roaster is 8-12oz, and that smaller and larger load sizes may be problematic and need to be approached with caution.

It may be worth rerunning this test at a later date with a greater sample size of tasters, but I believe the conclusions are probably solid.

From here it may be interesting to run some 10oz loads through P1, P2 and one of the stepped profiles, and taste test those to see if we can hone in on any significant differences between the profiles.

The turnout was small last week, so the cupping was me and Counter Culture employees Mark and Nathan. We all agreed the samples were roasted darker than we preferred, but none of the roasts were rejected as was the first test go-around. Sample "B" garnered one 1st place and two 2nd place votes and was the overall winner:

Table:

Sample    Rank
------  -------
  A     2, 4, 1
  B     1, 2, 2
  C     3, 3, 3
  D     4, 1, 4

We had enough left over to run an Agtron:

Table:

Sample  Low/High
------  --------
  A     37 / 48
  B     39 / 52
  C     35 / 47
  D     37 / 48

(The first measurement is prior to grinding). I neglected to collect cupping notes. Positive comments from memory included "good chocolates and high body." Negative comments included "ashy aftertaste" and "the varietal character was lost in the roast."

[farmroast]

This roaster really needs a recipe book/advanced guide. Manipulating settings and batch size can produce some outstanding roasts especially at the city level. 1lb. batches have major limitations.

[mike]

OK, so, time for next experiments. This time we'll do 3 profiles against 10oz loads since that appears to be the sweet spot.

I will probably go to CCC this Friday to get three 10oz loads of green. I may just mix it up and get something like a Guat which might take a little darker of a roast level (cool on first ticks of 2c) and still shine.

I am planning on two of the profiles being P1 and P2. I'll roast P1 first, and then set P2 to have the heat reduction start about 30 seconds before 1c time that we'll get from the P1 roast.

That leaves the third profile to choose. I do not have much experience with them and haven't had much luck, so here is where I need you guys if the third profile is going to have any sort of a fair chance.

Which of the remaining three profiles (P3, P4, P5) do you guys want me to use?

What timing do you guys want me to use? (ie. at what time do you want the profile to get to 100% power?) We'll figure out the initial set time to get to 100% by that time, and then add time after starting the roast.

Post your thoughts!

[MellowCat]

Mike - nice work!

Couple comments: not wanting to influence your current choice of bean, but what about a common Brazil bean that many of us use for "bread & butter" espresso and espresso blending? I use far more brazil than any other bean (unless I find a Brazil substitute) because it makes the best 'base' bean for espresso blends, as well as great SO espresso if the bean is good on its own (ie, Daterra, any of the Yellow Bourbons, Poco Fundo, Formosa etc.)

The Brazil's present another challenge with the Behmor, in that they are (mainly) low grown thus soft bean and one must watch out for scorching, ie a gentle heat application. (Not sure if this is too much too soon)

You mention having the power reduction of P2 occurr just *before* the start of first crack. Isn't it better to get first crack off and going before any reduction in heat? Having first crack stall on you, ie it not getting it off to a rolling crack, I thought is one of the big pitfalls.

Of the 3 slow-ramp profiles, Tom at SM prefers P4 over P3, although he doesn't state why. Otherwise I couldn't see much of difference last time I looked. Will make point to do.

Thanks again