Like mentioned earlier, stock driver is well regulated to a level far below what is healthy for two cells. So I decided to try and use that just keep it cheap and simple., and also to see what it was good for. Stock output were no good. Especially not considering the size/capabilities of the light.
I modified the driver to output well above 5A to the emitter outside the light, but it generates a ton of heat, and once it heats up (which happens quickly) output also declines. I decided on 4,8A to be on the "safe side" (added the equivalent to an R080 on top of the stock resistor). I did not consider this safe, but the driver have managed two 20 minutes continuous indoor runs at max output, and about 20 minutes of outdoor use. So it seems good based on limited testing.
The further the driver is pushed the earlier it falls out of regulation. Considering that its regulated down in the 4-5V range, that is no issue. My light is still pushing out the same 4,8A to the emitter (tested with XM-L on cooper) when I had two cells at 4,77v. The warmer the driver gets, the more output it drops. Thanks to all my potting, and lots of solder from driver to pill, Im down to 4,7A @ 30 seconds. If it had no heatsinking it would fall down to 4,5A in less than 10 seconds.
After mod tailcap readings start from just under 3A and goes upwords to 3,4A+ depeding on cells used and how low voltage goes.
I used an XP-G2 S2 2D. The natural high performance choice for de-doming IMO. I also changed to stock centering ring for one that fitted XP-G. That means my centering is perfect. My light can over focus, so I can also get my focus perfect for the highest lux readings.
My light meter is calibrated, by me. To a level well below the highest readings meters on BLF. People can check some of my lux numbers here if you want anything for comparison. But my current calibration is slightly lower now. The other day I compared a light with someone with a high reading lux meter. I had to add 15% to my reading in order to get to that level. That was more than I had estimated and thought. But I don't like being too much above many of the manufacturers that are on the conservative side in comparison to the often seen high BLF numbers. And im not going increase my calibration to a level where Im responsible for pushing (over inflated) numbers higher than anyone else.
As for max potential. I could have pushed the stock driver further, maybe it could actually survive or handle 5,6A? But if its got some ripple, the emitter could be in danger. I also could have rebuilt the pill, which would not have been too much work. Put in a 5Amp FL-2 style driver. I could have squeezed an extra 0,1-0,2A out of that. 5A driver was my original plan when I ordered the light. But settled for slightly less. Since that gave me a simple and cheap mod.
If I had been able to push my S2 2D from 4,8A to 5,6A I would be looking at a 7% lux increase . I know that since Ive measured the light output increase between those currents when collecting data for this thread. That would have resulted in a bump from 457Kcd to 489Kcd with my calibration assuming the driver did not fall quicker due to heat. Probably not worth to push it that far. The light does get very hot after 20 minutes on highest setting indoors. I use the light outdoors, mostly in the cold, so technically I would not mind a hotter light. And I don't really use these types of lights for long. Too narrow beam for me. Its fun though. :)
I hope that was enough details about the mod and makes things clearer? Feel free to ask any questions.
Oh, I can add some more about the light itself. Switch compartment on my light is small. Uses a tiny switch, not room for a large one switch. I plan to improve the switch assembly and swap the switch. If I put in cells that were a bit too long (protected) my light would not even work since the crappy switch assembly did not handle any pressure. One of the O-rings on the head on my light went bad too.. Not too surprising though. I have recently built over 16 UF-T20 lights. Lets just say they were not all the same. Bad switches in some, bad soldering. Some of those even had differences in how far they would focus. I had to find parts from the lights that could overfocus, and swap with lights that did not focus completly in order to get the best out of all the lights.... After playing with so many Uniqefire lights, Im not a fan of their build quality or consistency in parts, lenses or anything. Jax Z1 is in another league build quality wise, but that is not really budget or comparable with the UF1405. Considering the price I paid for the UF-1405. Just above 26$. That is fantastic value for such a capable mod host. Especially since the stock driver could easily be modded to outperform whatever you can make a linear driver do. Very impressed with the overall value considering the price. Although I would not recommend it as a stock light, especially not the unit I got.