Let us know what you get and also the f/w version you’re on. Sounds like from @krame’s test that the f/w version has something to do with it as well. This is very interesting stuff… 
1 Like
I have a number of data points that would suggest this to be the case. Plus Aillio support recommending the firmware upgrade as a fix for the wattage drop.
Also three tests that point towards better underbody cooling leads to a stable power draw with the v602 firmware
1 Like
Yea… I’ve been cooling the underbody with a small fan pointing at the left side of the roaster (same side as the plug), but my f/w is a much older version (I think 561 or 567) since I bought mine in Dec 2020. I’ve been hesitant about upgrading the f/w due to so many people having issues with it and having to go into bootloader mode… just didn’t want to deal with a potential time sink to resolve any issues.
1 Like
I feel exactly the same way. I am at v591 and am reluctant to upgrade the f/w.
But… I will probably take the risk and upgrade the firmware after my first test. The benefit may be worth the risk.
you go first 
OK, F2 and D9 with latest production FW (not beta) as well. I didn’t take voltage during the run, but after the IBTS went red I cycled the roaster and that time watched voltage and current as the thermal management kicked in. At the end of that pass I was seeing in the neighborhood of 12.75A@105V for ~1340W, though I sampled it a couple of times at 1320W, and had I left things running, I’m sure I would have seen it continue to track downward.
My AC is far from optimal. It’s a 15A circuit, probably at the end of the branch, and I use a 10AWG 10ft extension cord to situate the roaster out on the patio. My garage has a dedicated 20A circuit so I might have to try this again out there!
| Time | P | Coil RPM | IGBT | PCB | Measured Watts |
|---|---|---|---|---|---|
| 15 | 1 | 2640 | 33 | 28 | 314 |
| 45 | 1 | 2640 | 33 | 28 | 313 |
| 75 | 1 | 2700 | 34 | 28 | 313 |
| 105 | 1 | 2700 | 35 | 29 | 313 |
| 135 | 2 | 2640 | 36 | 29 | 418 |
| 165 | 2 | 2640 | 38 | 29 | 418 |
| 195 | 2 | 2700 | 39 | 29 | 420 |
| 225 | 2 | 2700 | 40 | 29 | 419 |
| 255 | 3 | 2640 | 41 | 29 | 548 |
| 285 | 3 | 2700 | 43 | 29 | 544 |
| 315 | 3 | 2700 | 45 | 30 | 541 |
| 345 | 3 | 2640 | 45 | 30 | 544 |
| 375 | 4 | 2640 | 47 | 30 | 680 |
| 405 | 4 | 2700 | 49 | 30 | 678 |
| 435 | 4 | 2700 | 50 | 30 | 679 |
| 465 | 4 | 2700 | 51 | 30 | 686 |
| 495 | 5 | 2700 | 53 | 31 | 823 |
| 525 | 5 | 2700 | 55 | 31 | 822 |
| 555 | 5 | 2700 | 56 | 31 | 817 |
| 585 | 5 | 2700 | 58 | 31 | 823 |
| 615 | 6 | 2700 | 60 | 31 | 970 |
| 645 | 6 | 2700 | 61 | 32 | 966 |
| 675 | 6 | 3000 | 64 | 32 | 972 |
| 705 | 6 | 3300 | 65 | 32 | 976 |
| 735 | 7 | 3660 | 67 | 32 | 1128 |
| 765 | 7 | 4680 | 68 | 32 | 1132 |
| 795 | 7 | 4680 | 68 | 33 | 1124 |
| 825 | 7 | 4680 | 68 | 33 | 1130 |
| 855 | 8 | 4680 | 68 | 33 | 1281 |
| 885 | 8 | 4740 | 70 | 34 | 1280 |
| 915 | 8 | 4680 | 70 | 34 | 1278 |
| 945 | 8 | 4680 | 71 | 34 | 1275 |
| 975 | 9 | 4680 | 72 | 35 | 1426 |
| 1005 | 9 | 4680 | 74 | 35 | 1428 |
| 1035 | 9 | 4740 | 75 | 35 | 1424 |
| 1065 | 9 | 4680 | 76 | 36 | 1405 |
| 1095 | 9 | 4680 | 77 | 36 | 1399 |
| 1125 | 9 | 4680 | 78 | 37 | 1389 |
| 1155 | 9 | 4680 | 79 | 37 | 1380 |
| 1185 | 9 | 4680 | 80 | 38 | 1371 |
| 1215 | 9 | 4680 | 80 | 38 | 1350 |
I’d put money down that if you try it on the 20a circuit, you’ll see nearly the exact same thing.
I tried on two separate brand new 20a isolated circuits with basically the same result as you.
I use a 10 gauge 25ft extension wire, so that will be the next trial before actually roasting again. I don’t expect to see much change.
1 Like
This is a really interesting thread. Thanks for documenting this.
I don’t have access to the wattmeter, etc, in order to participate, but can add that I’ve been running f/w beta 602 since October 26, 2021, (only 66 roasts since) and had no problems with it.
I’ll try this tonight
The bold chart is without underbody cooling. The overlaid(?) dimmer one is with underbody cooling. Both were P9 F2 D9 plugged directly into an isolated 20a circuit.
These are two empty burns both at P9 F2 D9 with active underbody cooling. One with a 25ft 10/3 extension cord, one plugged directly into the 20a isolated circuit.
EDIT: Both above are on beta 602 firmware
I ran a 500g roast with beans.
Firmware=591
Preheat to 482
P7 1195-1175
then P1=340 and P3=590
0 Minutes. Start roast
P7=1202
1 minute
P9 =1430- 1400
2.25 minutes
P8=1353
3.5 minutes
P7=1200
4.45 minutes
P6=1043
6 Minutes
P5=884
7.9 minutes
P4=742
Drop at 8.13 minutes
This was inconclusive for me. I should have chosen a roast that had a longer P9 to yellow.
P9 was lower than the anticipated 1500 watts, but not significantly.
After the Bullet cooled down, I updated the firmware to 602 without any problems.
I ordered a 120v muffin fan and will bolt it to the inductive circuit board guard. It only draws 10 watts. I like the idea of mounting the fan. Less clutter.
1 Like
It’s a little garbled. Nearly identical roasts, the overlaid and most recent roast was done with F3 vs F2. The only other difference is the firmware and underbody cooling. Almost a 3 minute difference.
FC in 7:40 vs 10:10.
Finally. I can get back to learning how to roast.
4 Likes
Just a random thought:
Would it not be possible to use the cooling tray fan to cool the PCB while roasting?
Roast in B2B mode with the cooling tray fan blowing on the PCB(cooling tray empty of beans of course)[
Cheers!
1 Like
That is an interesting idea.
The cooling tray fan is located to help cool the PCB when it is running. But we don’t really have control of it. We can’t turn it on whenever we want to.
And, when there are freshly dropped smoking beans in it, that smoke gets drawn across the PCB and into the Bullet.
A lot of people put a duct on the exhaust of the cooling tray fan which keeps the fan from blowing on the PCB.
You can control the tray fan in Back2Back roasting mode(just make sure you aren’t cooling beans at that moment), but if your tray fan exit is “plumbed” to a vent then that idea goes out the window sadly. 
I see what you did there
1 Like
Once I get my 3D printers back up and running, I’m going to make a bracket that uses the electronics bay screws to mount a 120mm fan.
1 Like
Yep… I “plumb” that cooling tray fan to a vent hose because I don’t want the exhaust in my house. Interesting idea but I find the cooling tray fan to be quite a bit louder than a small fan.
1 Like