Designing roast curves for the Behmor 1600 Plus

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
JimF
Posts: 37
Joined: 8 years ago

#1: Post by JimF »

Several books and posts have discussed the effects of the roast curve on the flavor profile of coffee. In all cases that I have seen, these discussions presume that the coffee roaster can control environment temperature (ET) and can measure bean temperature (BT). Coffee roasters designed for non-commercial use such as the Behmor 1600 Plus cannot measure BT without a BT probe modification (such as the FrankenBehmor), and so must make do with designing the roast plan around ET measurements (sensor "B"). This post is a (hopefully humble) attempt to do exactly that: to develop optimized roast plans for the unmodified, Behmor 1600 Plus.

To define the objectives of the roast, let's draw from several sources: Among the best of these is the 2015 Roast-and-Learn-Together diagrams, which characterize roast plans in terms of an idealized range between "Slow Start Fast Finish" (SSFF) and "Fast Start Slow Finish" (FSSF):

Roast and Learn Together - May 2015

Important theoretical contributions that attempt to define the objectives of the roast plan include Rob Hoos' manifesto, which discusses the effects of 'momentum' early in the roast, and the chemical reactions going on inside the bean from Maillard through the end of the roast. Extending the roast time between Maillard and 1C to enhance body would correspond to a SSFF roast plan. An apparently controversial contribution comes from Scott Rao, who advises about commercial roasters but specifically not about home coffee roasters, that the energy at the start of the roast must be relatively high, that the Rate of Rise (RoR) should steadily decline after the initial ramp, and that the time between 1C and drop should be fairly short. Rao's ideas therefore correspond to a FFSF roast style. Efforts to achieve a roast plan with a declining RoR on home roasting equipment capable of measuring both ET and BT is discussed here:

How to achieve a steadily declining ROR (not about the book!)

This doesn't provide specific guidance for stock Behmor owners because BT is not directly measurable, but within our limited ability to measure and control the Behmor, we can define some constraints on what a desirable roast plan will consist of:

- The user manual advises using a 1:45 preheat at full power, and there seems to be agreement that this helps to provide initial momentum to the roast
- There should be sufficient energy in the drying phase (pre-Maillard) to give momentum to the roast. It must also be less than 40F/minute, which is as fast as the Behmor can raise ET at 100% power.
- ET should be monotonically non-decreasing throughout the roast so that the beans are continuously roasted, and not baked
- RoR should decline from MAI through the end of the roast. This is true for the curves promoted by both Rao and shown in Hoos' manifesto, incidentally. Since we cannot measure BT, we will use the RoR of ET rather than the RoR of BT as our RoR measure. In other words, the ET curve should be concave down, and ET_RoR should approach 0 near the end of the roast, just as the ET reaches an optimal maximum temperature for the drop
- The roast curve is feasible to achieve on the Behmor using manual mode controls, possibly by adding foil over the chaff tray, raising the legs, or opening the door at specific times
- Rapid cooling of the beans should begin immediately at the end of the roast.

In many of the SSFF-like roast curves that have been posted for other home roasters, I observe that the RoR declines almost linearly near the end of the roast. Although approximate, if we make it our goal to achieve a linear decline in ET_RoR between MAI and the end of the roast, then ET must be quadratic and concave down above MAI, and in this case there is only a narrow range of ET_RoR at MAI that satisfy all the constraints above and produce curves that have similar curvature to the recommended curves. This restriction on the feasible range of ET_RoR at MAI guided my search for a roast plan that satisfied all the constraints.

There is also the Behmor quirk that the afterburner fan comes on after a few minutes, drawing air out of the chamber and causing ET to drop unless more heat is applied. The time that the afterburner fan comes on is a function of load size. For a 1/2 lb setting, the fan comes on at 5:00 into the roast, and I have found that pressing P5 for 45 seconds at this point will prevent a decline in ET.

After many experiments with an empty drum to try to achieve an ET_RoR SSFF profile that satisfies all of the constraints above, I found that the following very simple roast procedure is reproducible for up to at least 1/4 lbs of beans (untested on larger loads), and produces vastly better bean development and flavor than any other roast procedure that I have found thus far:

1. Load the drum with beans. Insert the chaff tray but keep the drum out. Preheat by pressing 1/4 | Start and then Stop after 1:45. Quickly remove the chaff tray, insert the loaded drum, and replace the chaff tray.
2. For 1/4 lb of beans, press 1/2, P3, C, Start, P4. This places the machine at 75% power in manual mode.
3. At 9:00 on the countdown timer (5:00 into the roast), press P5
4. At 8:15 on the countdown timer (after 45 seconds), press P4
5. At 14:00 * 0.25 = 3:30, press Start to prevent the Behmor from helpfully shutting off and ruining your beans!
6. Press Stop when the roast is at the desired point. Wearing gloves, quickly remove the chaff tray and place on top, remove the drum, and press 1/2 | Cooling to start the cooling cycle. Immediately begin the cooling procedure.


There are many different procedures for cooling the beans, but that's a separate topic.

So, there are quite a few constraints, but it just so happens that 75% power after the recommended pre-heat achieves a "Goldilocks" ET_RoR that produced almost exactly the SSFF roast curve that I was hoping to obtain. I expect that a P5 / P3 plan can be found that achieves a more FSSF-like roast curve, but I'm still experimenting with that. If you know of one that you like that satisfies the constraints above, please post it!

Those of you who have modified the Behmor to measure bean temperature, can you comment on this procedure and suggest any ways to improve it?

ira
Team HB
Posts: 5497
Joined: 16 years ago

#2: Post by ira »

For what it's worth, I current roast 300 grams batchees in FrankenBehmor and this is my routine.

Preheat to 350-380 on the BT probe,
heat to 100%
insert beans
when bean temp starts rising, about 1 - 2 minutes, lower power to 108/127
When it starts to smoke, engage afterburner and raise power to 114/127

Other then some funny stuff right at the end, that is a rather nice looking declining ROR profile

I'm not claiming it's a perfect profile or even close, but so far, the only adjustments I make are between 100/127 and 116/127 and mostly it's just choosing a number right at the start to set the total roast length. 108 is slower, 114 is faster.

That's not what I expected when I built FrankenBehmor, but it's what seemed to make sense once I could see bean temp.

But, I never claim to be a good roaster, I just try to be an observant one.

Ira

Advertisement
JojoS
Posts: 170
Joined: 11 years ago

#3: Post by JojoS »

I am inclined to think that even with pre-heating of the roasting chamber, the Behmor will always be a slow start roasting machine. I believe it is inherent to the design and unless there is a mod that allows the charging of the roaster without much chamber heat loss like what can happen with opening of door and pulling out the chaff tray while heating elements are off and inserting a loaded roasting drum before putting everything back to start the roast, it will remain a slow start roast no matter how quick you can do the process.

I also see no reason why one will bother to pre-heat and not start the roast at full power. My 2 cents.

ira
Team HB
Posts: 5497
Joined: 16 years ago

#4: Post by ira »

I can preheat for as long as I want with no adverse side effects. While lots of heat is lost while going through the insert the drum dance, the chamber starts off way hotter than if you don't preheat and most of the initial heat goes into the beans instead of the metal. FWIW, I always do the insert the drum dance with the elements at 100% to minimize the heat loss and the time it takes to get the elements hot.

Ira

JojoS
Posts: 170
Joined: 11 years ago

#5: Post by JojoS replying to ira »

What you have is not in the same category anymore. My unit won't start if I pre-heat too long. I've considered the idea of delayed charging and installing a toggle switch to stop the motor to facilitate loading of drum while heating elements are in full power but being happy with what I produce with my present set up, I have procrastinated on the the implementation.

JimF (original poster)
Posts: 37
Joined: 8 years ago

#6: Post by JimF (original poster) »

Ira, it's interesting to note that you set power at 108/127 = 85% for what sounds like most of the roast. That's higher than the 75% that I found to be the Goldilocks SSFF setting (but using the stock Behmor I would need to toggle P4/P5 to get an average of 85%, which could be accomplished by toggling P4 for 36 seconds, then P5 for 24 seconds). If you were to set it at 85% after the Behmor-recommended preheat, and just leave it there until drop, in your opinion and experience, would that result in a better roast than if you set it to 75% until drop?

In your modified configuration, do you manually control the afterburner, hence you are seeing smoke at an earlier point in the roast than smoke would be seen in an unmodified Behmor? When I see smoke it is in rolling 2C and time to drop. If I understand correctly, what is your purpose for raising the heat from 85% to 90% at the point where you see smoke, and what ET does this correspond to on your machine?

User avatar
happycat
Posts: 1464
Joined: 11 years ago

#7: Post by happycat »

.An apparently controversial contribution comes from Scott Rao, who advises about commercial roasters but specifically not about home coffee roasters, that the energy at the start of the roast must be relatively high, that the Rate of Rise (RoR) should steadily decline after the initial ramp, and that the time between 1C and drop should be fairly short. Rao's ideas therefore correspond to a FFSF roast style.
I think there's a problem with this statement re the FFSF label?
LMWDP #603

Advertisement
JimF (original poster)
Posts: 37
Joined: 8 years ago

#8: Post by JimF (original poster) »

JojoS wrote:I also see no reason why one will bother to pre-heat and not start the roast at full power. My 2 cents.
My initial attempts were to use P5 after the pre-heat, and then reduce to P4 at some point, and possibly reduce to P3 some time after that. What I found was that the initial ET after the pre-heat is between 138F and 147F, depending on how long it takes me to insert the drum, and that starting off at P5 and then switching to P4 as low as 165F carried so much momentum into the roast that the Maillard to 1C time was always very short. Starting off at P4 slows the roast down dramatically and my perception was that the beans developed a lot more body and flavor.

User avatar
yakster
Supporter ♡
Posts: 7319
Joined: 15 years ago

#9: Post by yakster »

JojoS wrote:What you have is not in the same category anymore. My unit won't start if I pre-heat too long. I've considered the idea of delayed charging and installing a toggle switch to stop the motor to facilitate loading of drum while heating elements are in full power but being happy with what I produce with my present set up, I have procrastinated on the the implementation.
Once I upgraded my old Behmor to a + with the new board, this restriction on pre-heat disappeared.
-Chris

LMWDP # 272

Bunkmil
Posts: 358
Joined: 10 years ago

#10: Post by Bunkmil »

I put the drum in without stopping the roaster (using ove gloves). That allows me to charge at hotter temperature. It also gives me the ability to play with the moment the afterburner starts after the beginning of the roast.

I can charge after the afterburner kicks in so I don't need to bother with ET dropping in the middle of the roast.

I'm still experimenting and haven't found the perfect combination of batch size and heat control so this thread is very interesting!

Post Reply