Ah yes, you are right, where are my eyes…
P.S. how about combining this mod with zener mod for 2s-3s MT-G2?
Wow, comfy that is a superb job. Very neat soldering.
On the original board the capacitor is attached to the PWM trace and ground. This is the same trace the leg of the FET attaches to. When you moved the cap to the inside you went to the power lead and ground, does that have the same result? Is the capacitor doing the same thing when hooked to these different places?
never mind, I see now that I had the wrong via. Upside down and backwards. :~
Newp! Find the MCU pin marked 'Vcc', follow it down and look for the through-hole right under the fork in the road...
...that goes over to the battery side, to the capacitor. Other end of the cap goes to ground.
The new spot I picked for it goes between the LED+ pad and ground.
Sirius, you can use it with 2S MT-G2. Don't try 3S MT-G2 (this isn't a buck driver).
Also, you need to remember that this will go near direct drive, so it isn't a good setup for those lights that you need or want to limit the current on due to heat or battery life concerns. This is a good solution for when you can or want to run at direct drive current levels. This is a good "hot rod" light driver.
Well, 40% (PWM=100) will limit a triple XPG2 to around 3A, sounds practical to me. :bigsmile:
Awesome guide Comfychair!!!
I’ll have to give this a shot.
Comfy and Richard love to kick ass and chew bubble gum, and they are all out of bubble gum. Thanks for the detailed write-up comfy!
Richard, I don't really have a need for this right now, but I assume this will be your next offering?
Of course. I have just been way too busy. Next week I have a week "off" and will be catching up on everything and posting all of the new "stuff". 
Thanks CC and others. Very nice work.
The place the cap used to be located was on the VCC trace and now it’s on the +Vin trace. It does the same thing, either way? So what, exactly, does the cap do anyway?
It keeps the MCU happy
Master Slave? Possible as well with this? and you still need the zener mod for MT-G2 correct?
To control the DD effect, could you balance it out a little by just not having the high mode, like maybe 200 or 225 instead of 255 for the high mode?
Justin, yes you still need the zener mod to keep the MCU happy. That's all the zener mod effects anyways--the voltage the MCU is seeing.
I don't think you will need to master/slave this setup, the MOSFET is good to 30A+ no problem and is very low resistance. After running an MT-G2 at 10A-11A it didn't get hot at all sitting in the open air. Heat is not an issue with this.
You could definitely run a lower PWM value, which is what Comfy hinted to earlier. You could also run higher-resistance cells. (Like a a Pana B vs. a 20R or PF).
I just take off the components and flux the board, heat each pad, (or several pads at a time), with the soldering tip and wipe with a rag, as I pull the tip away. Cleans off fine. Just an alternative…
Wipe with a dry rag? Or a damp paper towel? I like that clean look, makes it much easier starting over with new components.
Thanks Justin! 
I usually use a paper towel, sometimes damp, sometimes dry, depending on my laziness level. A rag works a little better and dry works fine. You just have to wipe quickly. I place the rag over the tip and wipe in one motion rapidly. Excess flux is cleaned off with alcohol.
The only difference is that it used to be after the polarity protection diode (between diode & MCU), now it's before the diode (between B+ & diode). These little caps aren't polarized, so they won't explode (like the little yellow tantalum caps) if you stick the battery in backwards, so it shouldn't have any effect.
There are other places it can go on the topside and still be after the diode, but they all either interfere with either the driver fitting in the pill, or block a programming clip from attaching to the MCU. It can go from the bottom of R1 to ground, or from the bottom of the diode to ground. Both those would need a notch in the ID of the pill for clearance. It can also go between the lower tabs of the R1 & R2 resistors (since the lower tab of R2 just connects to ground anyway) which takes a little less room, but would still need a notch in the pill. If the driver is still using the 7135s, it can go from MCU pin 8 to the middle ground pin of the 7135 to the left, but that blocks the programming clip.
This type of driver can have the max current limited by using a PWM value of less than 255, but it will always be proportional only, meaning, for example, PWM=200 will give you 80% of a 20R's 8 amps, or 80% of a ICR's 5.5 amps. Changing the cell type or state of charge will change the output current, it is only regulated by the cell's voltage as compared to the LED's Vf, even at less than 100% PWM values.
^^^ But this can be a good thing... the mode spacing will remain the same, even if you change from a 8 amp cell to a 5 amp cell. Other big-current-but-regulated drivers will not maintain a pleasant-looking mode spacing if you use a cell that can't do the full max output on the high mode - mid & high end up looking too similar, and makes the reduced light output more apparent.