I did, but the results were really disappointing..
My TN31 works really well with dedomed XP-G2 R5 1A (from IO) on SinkPAD, providing around 440kcd, but K40 with resistor mod and dedomed XP-G2 R5 1A (from Fasttech) on Noctigon gives only 230kcd. :(
Same light with a dedomed XM-L2 1A gives 365kcd, so the light is not dud.
Now the question is: Did I get bad XP-G2s from Fasttech (tried four of those with very similar results) or is the K40 reflector geometry really so bad for dedomed XP-G2s?
With the results of throw and more lumens that a dedomed XM-L2 will give like on Vinh’s mods, there really is no reason to even use the XP-G2 emitter.
Here’s the results of a K40VN at 671yds. Lots of lumens, more spill and even more throw vs an XP-G2 is why Vinh likes de-doming the XM-L2 and pushing up the amps.
Im rarely impressed with extremly overdriven XP-G2s vs overdriven XM-L2s..
What is the difference in emitter amps? XP-G2 vs XM-L2. The answer might lie there...
Ok. I tested with old XP-G2 on SinkPAD and I'm now convinced that the (huge!) difference between TN31 and K40 is caused by the reflector shape.
K40 reflector has wider flat bottom, which makes it impossible to focus perfectly with XP-G2. :(
One option would be to grind the whole bottom away, but I'll take the easier route and go with de-domed XM-L2s instead..
some reflectors really don’t like to play ball with the xp-g2 
its a shame because I much prefer xp-g2’s to xm-l2’s myself, nice one for trying though, it was a worthwhile experiment