Still plugging away until I find that firesale Diedrich IR-2 on craigslist Seattle.

Anyway, I finally tried another profile for 14 roasts and after 109 at the same one. After some initial excitement, I'm back to my standby - 1lb/P1/A.

Here's what I tried for 14 roasts:
- 320g (a couple were 280g)
- 1lb/P2/C
- Machine still modded (disabled chamber thermocouple, 1/2 covered chimney, no power to AB)
- Behmor thing from Ira to follow the profile visually.
- Roasts are stopped by hitting "COOL" and opening the door 1" at the top for 1 minute. After 1:00, I open it all of the way. I pull it and can touch the drum with bare hands at 3:00 in.

Since I've effectively got a brick on the pedal, I had thought that using a modulated profile would allow me to extend the times between 1st and 2nd as they had gotten quite a bit shorter. I'd lost about a minute from when 1st got rolling to when 2nd got rolling. (From a span of 3:00 to 2:00 or so. Feedback on these times is welcome.) I did not take the time (being honest) seriously taste-test this change, I was directing it from what I had read and learned from some previous experiments. (I have thought about cracking the door to bleed off heat but I need a repeatable method for this. Maybe a screw through the door into the jam would allow me to make consistent yet incremental changes.)

Results:
It did draw out the times from rolling first crack until the end of the roast. Normally, this time window is 1:45 to 2:30 long. With P2, I saw from 5:20 to 7:15 spans. So I learned that I could do this but it was ultimately not that rewarding for the following reasons.

1) Since it cuts power to the elements at a fixed time of the cycle, it's basically cutting the power at random times when compared to my stage of roast. Looking at my notes, I had a band that was 2:30 wide where 1st crack would get rolling. Since the power always cuts at 8:36 remaining, I could be 1:00 into 1st or 1:30 from it when the power cuts. My dream had been to try and hit that power cut right at 1st crack. I know there are other things to change but I'm not sure you could keep your sanity and do it with some sort of matrix of ambient temps, bean masses and voltages.

2) It became quite difficult to audibly recognize the stages of roast. The transitions between phases became really long and tapered. I would have to stand with my ear 6" from the machine trying to catch the end of 1st and then the beginning of 2nd, sometimes for a couple minutes each. This may be my inexperience and the fact that I can't reliably use the window for monitoring. It's all based on sound for me.

3) In order to get to an acceptable level of roast for me (Full City +), I had to take it....get this.... 85 seconds into 2nd crack. Not kidding. I tried 30, then 40, then 60, then 70, then 80 in the process. The shorter times were quite light. I'm not sure that this is categorically negative but I've not read a lot that supports needing that much time in 2nd.

3) The roast times were getting really long. This profile is 21:30 to start and I had some where I had to add a full minute to get that level of roast I wanted. That's 22:30, the machine's maximum roast length and seemingly bordering on where roasts head to "baked". I did not notice a deteriorating taste in the cup but it did not seem like a good direction to continue heading.

The variability of the roasting times compounded by/contributed to/magnified by the lack of control continues to challenge me. Whether it's the starting temps of the beans/box/room, the voltage or who knows what - I still can roast to my liking on the machine BUT every single roast takes micromanagement. The P2 profile had some potential but again the lack of flexibility during the roast renders it more of a complication than an asset.

Full manual control over the heating elements seems to be the only logical next step. I'd really like to be able to step it down when I want it to AND a percentage that I want.

djmonkeyhater wrote: The P2 profile had some potential but again the lack of flexibility during the roast renders it more of a complication than an asset.

Full manual control over the heating elements seems to be the only logical next step. I'd really like to be able to step it down when I want it to AND a percentage that I want.

I've come to the very same conclusions, after really trying over the past year. And using P1 as the only viable program still produces roasts that I wouldn't give away to friends. I've moved back the Gene cafe for the time being, which is no shabby piece of hardware to say the least.

Thanks for the detailed post, there's lots of us wanting the Behmor to be something more. (Perhaps like hands-on analogue heat control like the Gene!)

Ira,

I've been trying that off and on although not scientifically. I did a couple roasts where I tried to get the thermocouple to display a set temperature (100-130 degrees) but it was very difficult to hit it each time. On a couple of occasions, I did a 2 minute warm-up, essentially independent of a temperature target. I'll have to puzzle through a comparative data tool to understand if there are any linear relationships. For me, it was getting to be too hard to keep track of it all.

All in all, my frustration has abated a bit since the Behmor outputted 2kg last night (not a small undertaking) in 320g IN/260g OUT loads. With the 1lb/P1/A program and all of my mods, I am getting very consistent roasts that I like to drink but I'd like to experiment a bit with stretching 1st to 2nd. That profile is 18:30 long and I hit COOL near 2:00 +/- 0:20 remaining. I realize I don't have a ton of room to stretch before the roast gets overly long. Shrinking the charge would be a countermeasure to this.

My Auber Instruments PID Temperature Controller With 30 Steps Ramp/Soak showed up today. Now the goal is to take control of the elements with an SSR and see if I can program it for a "ramp" through the drying phase, to hold a temp until 1st, drop it a TBD % to 2nd and then pin it for the finish line. I'm an intermediate amateur electrician with limited PLC/digital control experience so don't look for it tomorrow.

Wes

mellowcat wrote:And using P1 as the only viable program still produces roasts that I wouldn't give away to friends.

For all of my grousing about the Behmor, I've been able to get some very good roasts out of it. (That's my opinion and I'm not a cupping Jedi but..... I think I can tell terrible.) I'll send some to the curious if requested.

In my experience, the two simplest ways to open up the opportunities with a Behmor are to:
- Stop relying on a profile without manual intervention
- Stop relying on the cooling cycle with a closed door.

Caveats:
1) I have NO concerns about smoke and smell.
2) I have NO concerns about killing/damaging the little guy in the process of using him waaaay outside his intended use. Especially the afterburner.
3) I might have a fire extinguisher, two EMT's, a Nomex suit, a lawyer, a NASA helmet and/or be a certifiably crazy reckless thrill seeker on a deserted rocky outcropping when I roast. So only listen to me if you have/are those things as well or can deal with what might happen if you don't/aren't. I'm just another guy on the internet with an odd username after all.
4) My voltage is 120 without heat and 117 with it. This can vary 1-2%.

Safety First....
- If there is a chaff fire, I might know someone that just pulls the collector out to stop it. Of course he isn't roasting on or near much flammable, especially his floor covering. Chaff will burn itself out in less than a minute I've heard, he just doesn't like the way it makes the beans smell. 10+ of them haven't affected his roaster.
- If there is a bean fire, you are on your own. I've never heard of this. Probably unplug it and order another machine online while it is cooling. Keep the burnt one for parts.

My recipe for an espresso roast:
- Greens in the 280-320g range. I weigh every load in and out to the gram.
- 1lb/P1/A

I follow and write down the following things and the time they occur as it displays on the screen. All are done audibly. I only use the window to see when the elements cycle.
-- 1P: The first or second major pop that I can hear. This means it is time to pay attention. For my setup, this is with 6:00 to 5:00 remaining.
-- 1R: 1st is rolling/sustained and noisy. I think it's about 30 seconds later.
-- 1E: 1st ends... mostly. For me, it's never totally silent for more than a couple seconds so it's a judgment call. I have found it tricky to repeatably note the end of 1st.

YOU ARE RISKING THINGS IMPORTANT TO YOU IF YOU READ THE FOLLOWING THREE SENTENCES!!!!
-- 2R: 2nd gets going. This is faster to ramp up than 1st. I'll get ready to open the door in EXACTLY 30 seconds. My experience has been that you can't miss this by more or less than 5 seconds without changing the results. Maybe 2:30 left.

- Hit cool and open the door. The more the better. I used the cooling program twice until I figured it was totally cooking the beans for a couple minutes after killing the roast. The drum should be cool to the touch in 3:00.

- Weight for me is reliably 225-230g for a 280g charge and 260-265g for a 320g.

dj - thanks for the detailed instructions - i'm still optimistic about getting some great roasts out of the behmor.

Part of my frustration - and presumably yours - is the inability to temper the heat during the P1 roast, maybe your larger load addresses that. But if your first crack comes on fast and hard, what do you do then? (to temper the heat) Only thing we have is to open to the door to let off heat. How much heat does that let off? One inch or four inches? And for how long?

What about the opposite, first crack is feeble? And then you see the elements start to cycle, entirely out of your control. I'm sure you've come up against this.

Also I've not yet done voltage measurements, but I could argue that why shouldn't the roaster perform acceptably under my 'normal' north american voltage supply.

thanks again for the info

I have the PID kit - ramp/soak controller, Thermocouple and an SSR. I've been closely and digitally monitoring the temperature of my beer for a while now.

Question #1 - Does anyone have an opinion on where to put the thermocouple in the roasting chamber?

There is a huge chunk of the machine that is covered by the chaff screen when it is in so it either will be pretty high in the machine or pretty far back. I could cut up the screen but.... will it be better?? Anyway, does it really matter as long as it's consistently placed/operated? I will have to "calibrate" the ramping numbers to make the roaster perform as I want it so the temperature value is seemingly more important as a relative value than it is as an absolute.

I think that I really need to get something into the bean mass for any sort of data useful outside of this project.

WES

Wes
I had planned to meet up with the person I lent my Behmor to and get it back last weekend,but he got a flu bug. It cost me $50. to have the new hub made and attached to my drum as shown earlier in this thread that made it possible to get a decent BT reading. I plan on using a bead tip thermocouple and fishing it around to different spots to find an appropriate spot for a ET/MET reading. My thought is a small diameter, fast reacting probe shot through the non-control side with the tip close to the drum on the heating element side. If the chaff tray is in the way I will drill a small hole through the tray and just pull both probes before removing the drum and tray. A bead tip stuck out of the side where the present thermostat is might be an ok location but until I do some testing I'm not sure.
farm

farm,

Thanks for the tips.

I'm wondering where to place the thermocouple in the roasting chamber or if placement really matters relative to the type of heat in the Behmor.

The TC for the bean mass will have to wait a bit. I have a couple hollow bolts that I was going to mess with - they are like this: http://tinyurl.com/8ykknw but I'm not sure if the I.D. is big enough. Your install looks pretty clean - I'll try to live up to that standard!

WES