Yup its done internally on the PICs. Can be done on AVR chips but normally the higher end ones.
Guys I just want to say a huge thank you to Mattus, WarHawk, ComfyChair, RMM and RBD for all the work they do, everyone please take a moment to check out my 1k post /1 year GAW thread and help nominate the last 2 winners and check out the custom engraved lights they’ll received.
Special thanks to Calvin at Illumn, Ric at CNQG and Richard at MtnElectronics for helping make the prizes what they are, couldn’t have done it without any of them!
The Medusa PCB works (albeit the version I'm testing required a physical modification that will corrected in the next revision). I'm not sure if people are interested but I'll upload the revised PCB to OSHPark later tonight.
For what it's worth I have not used a 70N02 FET. I have used a IRLR8721PBF. Seems to work fine - has no problems performing the diming or strobe functions of tterev3's firmware.
- Matt
Matt. I say this as a forum friend to save you any more embarrassment here, can you please make sure your soldering gets a little more practice. I'm not saying its bad just not quite up to standard. I'm glad I also told you that you would have to do a (s)light mod to get it working to save you frustration and the substitute fet to use.
Ok The above is BS. Its just my way of saying what an outstanding effort and l have no idea what any of it is. You sir are one amazing talented guy. Do paint fumes help with your creative juices? Can you let us know again what host and batteries you are using here? Cheers. Orsm orsm effort.
Edit. The wheelbarrow was full when I got out there.
I thought you were serious for a second because the wire solder joints are HORRIBLE. I broke my temp controlled iron and have been using a crappy iron from Dick Smith's (borrowed from my father) until I can afford a new station. I might still take this into work tomorrow and clean it up 
It's for a Sky Ray King. I've got the host apart and ready to wire up. I was going to have an 18350 tube machined by sinner so that it'd be a stubby SRK, but I need to stop spending money for a while. Maybe I'll get into the BLF spirit and 'chop' the 18650 tube. Then again if I bugger it up I'll be left with a head and no tube...
You’ll have a spare head for your next tube. 
In the OP you have noted that these boards by Helios are untested.
https://oshpark.com/shared_projects/H63c8d5q
https://oshpark.com/shared_projects/XMtGKfm3
I've made a battery carrier out of these items and they work like a charm. I tested on a run of 5 amps output on 8 x AA NiMh on an MTG-2 was very happy with the way it worked. The only change I would make if any were to be made would be to add a center hole in the center of the positive end the same as the negative one. This is to allow a non conductive rod to be located with tiny nubs on the end that fit in these holes and locate the rod.
Looks good! And it is nice to see that if the carrier is done right, you can run a MT-G2 at 5 amps on a stack of NiMh cells (for half an hour at least)
19 minutes was the max. At 20 minutes it was all over.
That light needs some lower modes then. I really like it, are you building a flashlight around it?
I dont want to build a light with a 20 minute run time on high. I did do a run on medium as well at 2.65 amps and ran about double the time it did on high. I'll try an XML. It should get substantially longer run time but it isn't a MTG-2 if you know what I mean.
Maybe a triple XML for that battery pack?
Speaking of triples… I finally put together one of Scaru’s triple XML’s and nada. They worked before I reflowed them to the board then nothing on the board. Reflowed them to different MCPCB’s and they work fine. Is it possible that I messed up the file transferring the file through my computer before having Oshpark assemble it? And something is wrong internally? I am at a total loss. Any suggestions?
I doubt that you messed up the file. Was it a dead short or open circuit? The diode test mode on your DMM (beep mode) will likely come in handy here.
Do you know the correct polarity to install the emitters in on the PCB? They are not all oriented the same way.
Do you have a pic of the unpopulated board?
Here is a picture of the boards, the dots are supposed to correspond to the dots on the top of the XML. Is the positive and the negative reversed on the mask?
Right now I’m trying to figure out the DMM buzzer…
The dot on XML2 is on the negative side, on XML it's on the positive side. Bond wires are reversed also.
I don't see a dot on the Oshpark board?
Here is a picture of them dots…
I might be slower on my responses shortly…
My buddy just stopped by with some GOOD beer. The DMM showed good continuity for each of the pads.
Did the DMM show continuity between + and - without the LEDs populated?
I just downloaded the files and took a close look in GerbV. Everything looks OK to me on the board. It seems to me that you have polarity/orientation correct.
I’d re-install the LEDs and test them using diode test, which should make them glow.
I always look at the bond leads going to the emitter, XM-L they point at the negative side, on the XM-L2 they point towards the positive side