---

Any one interested in posting their BM roast profiles?

---

I've posted some, scattered here and there. I found the key to add lots of energy (heat) as soon as the beans are yellow until just prior to first crack, then back off. I don't regularly go into second crack, but once second crack was reached, the beans could just 'cruise' without heat for a couple of minutes while the temp kept rising slowly.

In general:
4 min to yellow (beans should be yellow by the time they hit 300F, or you're going too fast)
4 min to first crack (first crack regularly was 400F~405F range)
2 min duration of first crack (first crack usually ended around 415~420F)
2 min more to full city + (cusp of second crack, 435~440F)

CHAD

---

That sounds about what I've been working towards. With my hybrid roaster, I have considerably more power than most HG BM's, and think I'm going a little fast on both ends.

My TC is cheap and fluctuates considerably, so my tendency is to err with too much heat as opposed to too little.

On a side note, the Temp monitor I have is the cheap one that SM sells, and it only accurate to the whole degree c or f. It fluctuates as much as +- 5 degrees. Do you think the fluctuation is a faulty TC, or the inaccuracy of the unit itself.

Since Fahrenheit is roughly double Celsius, I might gain some accuracy by switching to Celsius.

---

Does the temperature fluctuate measuring... ambient air temperature? Boiling water? Is this your thermometer:


This one should be accurate. If it is fluctuating, perhaps the probe tip being beat by the stirring action is causing the connection to go open?

CHAD

---

That's the one. The tip is only exposed a quarter inch or so. The only thing that touches it are the beans. Now that I think about it, it doesn't fluctuate in ambient air. It can be kind of flaky if moved around much.

---

If it varies when moved, then it probably has a spot where the wire is almost broken (high-resistance open).

You could buy a replacement K-type probe, even one with a beefy stainless tip. I had the same failure trouble with beaded tip probes... mine didn't go bad right at the tip, but an inch back where it rubbed. There's a thread around here somewhere with pictures of a steel tube running into the roaster to protect the thermocouple. I've searched but can't find it. The tube was part of an ink pen.


CHAD

---

If your temperature is only displayed in whole degrees, then you get more accuracy using Fahrenheit, since it will increment two degrees (roughly) for every one degree Celsius.

---

I remember that thread. I also remember some folks using brass tubing from an RC / Hobby shop. I'll have to see what I can find. I want to still be able to remove the BM pan to dump the beans.

---

I ordered some coffee from SM tonight and went ahead and ordered a new TC at the same time. Here is a video of how much the temp fluctuates when in my kettle.

---

I have to ask: was the probe submerged in the water, and not measuring steam temperature? If so, I agree with you that this won't do for helping you with coffee roasting measurements.

CHAD

---

It was submerged. I'm fairly confident that the issue will be rectified when my new TC arrives next week.

Thanks for the help.

---

Riffing on seedlings: (all temp measurements are BT with 350g loads)
1) as fast as possible to 250F
3) 180-210 seconds to 300F
2) as fast as possible to 1C
4) 6-10 seconds/1F to end, depending on bean and final target temp. Harder beans, lighter roasts go a little slower (closer to 10s/F), darker roasts go a little faster (closer to 6s/F).

Different beans call for different end times; I just finished some Poco de Caldas from Brazil that was great with a 6s/F final ramp ending at 438F, just when 2C started rockin', and am going through a HHT that works well with a 10s/F to 428F, just before the outliers of 2C.

I don't claim that any of this was exhaustive (i.e. "did you do double-blind batch testing at 0.1F increments?!!?!" ), but it tends to work well for me. Also, these are all roasts for espresso, pulled on a lever. So yeah, there's another variable that I haven't accounted for ;)

-s.