That’s a good guess.
My driver that does the advertised 9 amps on the left, and 18sixfifty's driver on the right which only does 5.6 amps no matter how many cells you throw at it.
Anyone spot the problem? :facepalm:
-----
One thing I did figure out while running them out in the open, the heat is coming from the FET and/or other stuff on the topside of the upper board, under the choke/inductor/whatever. The part that has zero connection to anything that could keep it cool(ish)...
I honestly don’t have a clue how these things work so you would have to point it out to me. Although from the picture it looks like you stacked resistors or did they do that?
I haven't touched the resistors on either one. There are 3 more pads for more resistors directly opposite those, and both drivers have 3 R200s there. So one has 9 resistors and the other only 6.
Nice to see that these are easily adjustable between 5,6A-9A then depending on the resistors.. :)
Comfychair, would be valuable information if you were able to figure out if less output equals a noticeable change in overheating issues. And if adding some heat sinking will improve it. :)
btw, does this driver have audible PWM, slow mode changing or any other annoying downside beside the overheating issue?
Well WTH? I guess that explains that doesn’t it?
I wonder if MRsDNF's underperforming driver also has only 6 R220s? (edit: R200s, durrr...)
I do hear some squealing but it's from the questionable connections between cells, I've just got them stuck end to end with little magnets that are really too weak for this. I never heard it making noises when it was installed in a proper light (the TR-1200). Mode changes are normal, not like that 3(yeah, right)-18v thing.
The top surfaces of the FET and the two diodes are at the same level so it should be fairly easy to JB Weld a piece of copper plate on there and connect it to somewhere it can dump some of that heat. But I really don't understand what it's doing or trying to do that results in that much waste heat, it gets hot enough after about 45 seconds that it'll sting your fingers pretty damn good. The 6.5A piece doesn't seem to run noticeably less hot than the one doing 9A, you could cook a tiny little egg on either one of them.
will you be giving it another go 18650?
Not for now. If someone had a driver that I knew would work with three batteries I would. The DRY driver with only two batteries would work but it would be only so so in terms of output. I’m already getting quite a bit of battery sag with the 2x and a DRY. I have a bunch of other stuff coming in soon so this one is going on the back burner for a bit.
I have two J-19’s right now so I have an extra host waiting if someone figures out a good driver configuration.
I bet 3 MTG2s in parallel would survive direct drive from 3 cells...
You sure about that??
That is one potentially expensive experiment..
Thanks for the info on the 9A driver.. I might be tempted to buy one.
LOL, no, not sure... 'I bet' as in 'I suspect', not 'I bet' as in 'I'll bet you everything I own'. :)
Of course it depends on the cells used too, voltage sag would be required to keep it alive.
I might try it sometime in the future. Right now I have a bunch of other things coming in so my budget is already shot for aug unless I sell something.
Three parallel MT-G2s direct driven with three cells in series? That is a 6.5V load driven with 12V. About 5V needs to be dropped elsewhere in the path. Most of this will probably be in the cells if the wiring is good. I’d predict something will melt before the emitters fry. If the wiring holds, hope that the emitters split the current evenly 
I got the 9 amp version to run without stepping down until the cells were in danger of venting, though I can't see a practical way to get this kind of cooling in a conventional flashlight...
That's the bottom of the driver board visible, I moved the toroid around from the other side (and that toroid came from an old SRK driver and it's not up to the task, gets too hot - I've built a new one but haven't swapped it yet). The original 'top' of the driver is the side that's against the heatsink.
It ran for about 10 minutes straight, the driver heatsink was cooler than the LED sink, but still pretty damn hot. The big one was too hot to touch, the little one was just uncomfortably hot. Without the driver bolted down it would run a few minutes at most.
I mentioned venting cells... look closely and you can see the cell wrappers shrunk back a ways from the negative ends, they weren't like that before that test. They got that hot. All 4 springs in the battery box also collapsed, and the plastic was creaking and popping before I shut it down.
I know the driver isn't supposed to run from 4 cells according to the specs, but under load the voltage was down around 11v so I guess that's why it lived. After the 10-12 minute run the cells were down to 3.65v each.
Nice! That is interesting and useful info. Thanks for sharing! :)
I added an emitter to my 2x now 3x and ran it with three 18650 2600mAh Samsungs. 20 gauge wire. It runs fine but the turbo on the DRY is really only good for short bursts (30 seconds or so) because the lumens drop down fast.
I have an extra J-19 and I sold a light last night so now I’m debating buying more emitters and trying again. I also broke down last week and bought another shocker ($87) if I’m going to run it with 18650’s anyway that might be the better host.
I’m not sure what is causing the fast drop in lumens. Thermal sag or voltage sag?
crazy, 3x mtg2 in a shocker. that would be a sight to see.
Well so much for the DRY driver. It froze up and is stuck on high. Works fine with 2 king kongs and 2 emitters in parallel though. So I’m back to looking for a driver. Ca sera sera.
Is it because you are trying to pull to high a voltage at the led or something else as this driver works on XML’s ok and the only difference I see is the voltage?