I have a
Quark 123-2 XP-G R5 that I modded with a Shiningbeam 1.4A driver, but I added an additional 7135 chip to bump the current up to roughly 1.7A. Here are my OTF results for this light, and for an un-modded Quark 123-2 XP-G R5:
_____Quark 123-2 XP-G R5 SB 1.7A mod____
___1 sec______357.6___________
__30 sec______305.5_________
___1 min______297___________
___2 min______292.1_________
___3 min______289.7_________
_____Quark 123-2 XP-G R5______
___1 sec______257___________
__30 sec______241.2_________
___1 min______238.8_________
___2 min______236.4_________
___3 min______235.2_________
As you can see, the 1.7A mod loses 67.9 OTF lumens from 1 sec to 3 min, while the non-modded 123-2 Quark only loses 21.8 OTF lumens.
At 1 sec the 1.7A mod is 100.6 OTF lumens higher than the non-modded Quark 123-2 at 1 sec. At 3 minutes the 1.7A mod is 54.5 OTF higher than the non-modded Quark 123-2, while it's 64.3 OTF higher at 30 sec, so most of the losses are within 30 sec, and it seems to fall fairly steady from there. I would expect the losses to be less with a drive current of 1.4A, but I couldn't say how much unless I tested one. I would also expect the losses to be less with an XP-G R5 with 1.7A drive current in a light that has more mass for heatsinking than the Quark has.
I wish I had one modded with the 1.4A driver without the additional 7135 so we could see how much of an increase there is from the factory Quark driver (roughly 900mA) to the 1.4A SB driver, but I don't.