I just ordered $80 more stuff from DigiKey an hour ago!
Edit: Voltage readings I’m taking with my DMM, amperage with my new clamp meter.
Have they fulfilled? Cancel.
wouldn’t you know it, they got in a big rush on this one.
I don’t know what I’m missing here. The MT-G2 is cranking, having to wear really dark sunglasses to try and deal with these measurements. What’s up?
We can still use these drivers, it’s just not ideal. Please don’t get all upset over this man! Give me a second to post my thoughts on how to deal with the situation.
The XM-L’s were also cranking before all those others fried. Stay tuned.
With 2 brand new Efest 35A fully charged, 3.83A at the emitter.
3.53A at the tail, as it were.
Want me to turn it on high and leave it? See how long it goes? Run over it with a herd of pack mules? Plug it into the wall outlet?
6.56V at the emitter on high. 8.14V on the cells during run on high.
I can hook up MATT’s emitters to this thing! 
<—-singing while working, “Angie baby, you’re a …special lady…living in a world of make believe…well maybe…”
So I guess Mattaus agrees with my conclusions on why these are popping. Based on that, these may not work reliably for the intended purpose, driving an MT-G2 at 4A. I don’t know that for certain, but it looks like that’s the case. We are exceeding the absolute maximum wattage rating for the chip by maybe 10%, depending on how much of the buck circuit heat losses happen in the chip vs the inductor (anybody stop me if I’m wrong).
For MT-G2 use these things should do about 22W at the absolute maximum rating, I think 18W would be conservative and not a bad thing. 18W is not a very high drive current for an MT-G2 for us but might be OK for some of our applications. A second board can easily be slaved in order to double the current, I don’t think there’s an issue with that. This gives 36W which ends up being a 5-6A drive current. Therefore all LED2001’s we purchased could still end up getting used w/out issue. They may not fit in a 17mm cavity doubled up, but these should be perfectly usable.
It's still not doing what it was intended for. The datasheet example is designed to handle 4A, but then again their calculated efficiencies may not line up with real world use. Then again (again?) that example is available as a demo board for purchase. Some drivers popping, and others lasting the distance may also come down manufacturing differences (hence why a range of ratings is usually given rather than a single specific value).
Regardless, my offer stands. If you want money back, please PM me.
Dale - what emitters? Did those 3 XPG2s and that optic arrive? Let me know what I need to give you to get them shipped over here.
- Matt
but it was still making 3.49A at 3 minutes run time.
I can’t speak for the rest, but for me your work laying out PCBs is more valuable than refunds on parts that still operate. Everybody knew that this was an unfinished design when they ordered parts.
I recommend everyone put testing on hold till Mattaus can re-design the board with a bit better heatsinking (the vias and moving the ATtiny) then try a few that way
Try to see if limiting current down to 3A MAX (for a XM-L) [as a temporary reprieve]…this way not pushing the LED2001’s too hard and to see one in action
Has anyone contacted the designer of the chip yet…do they have tech support for this type of thing
Gotta love R&D…or not
Started at 3.62A, still making a very solid stable 3.44A after 5 minutes.
I have this wedge of copper on the regulator right? So I strapped the driver to my 3x3x3 block of aluminum with the copper wedge flat against it, the big copper chunk housing the MT-G2 on top of this block. The copper under the emitter is hot, and rightly so. But it’s working. The copper under the MT-G2 is for my bullet light build, it’s 35mm dia. x 25mm high. With 27 cubic inches of aluminum under the copper, I think it’s heat sinked.
Good work so far.
You crack me up Matt. I think the efficiency numbers will improve with some changes so I wouldn’t be to bothered just yet. As far as I can tell, only wight has had a clue in your absence and I think we’ve only scratched the surface. Now that Dale has one working, let’s try to come up with improvements. Heat is always a bugger. I know you have a big chunk of copper on there Dale but consider the direct thermal path of a noctogon vs a star with a dielectric layer. Heat sinking that solder pad will be the equivalent of eliminating that dielectric layer and we all know what that did for output.
8 minutes combined on high at this point, reckon these 2 cells will go some more? 
8 minutes in one run, after a 5 minute and 3 minute previous run.
Start 3.66A
at 1 minute 3.35A
at 2 minutes 3.26A
at 3 minutes 3.19A
at 4 minutes 3.11A battery at 7.36V at 4:32 while running
at 5 minutes 3.04A
at 6 minutes 2.96A
at 7 minutes 2.91A emitter Vf at 6.26V while running
at 8 minutes 2.85A
This is all on the same initial charge on the cells, much as you’d use a light out in the field. A combined 16 minutes on high.
Batteries at 7.53V now at rest.
Dale. If I was rich or my wife got that fifth job I want her to but she refuses I would buy you a power supply. Nice effort.
Dale, does your DMM have a thermocouple? It might be more accurate on small parts than an ir gun.