Pavio, I am not so young anymore. Most of my life a flashlight got dim as it’s cells grew discharged, so to me that is how it is supposed to work. As my work with the flashlight progresses, I am constantly reminded of how much longer the flashlight will work for me by the ever decreasing output. I can easily judge when to take a break and replace cells. Being very forgetful by nature, all my life, when I am involved in work I have little time sense, so the flashlight itself can help to remind me to take a break, get a drink, food, whatever.
I also do not like pulling more amps out of a nearly dead cell. The regulated drivers do this to maintain brightness as the cell discharges, the lowest current is pulled from the most capable charge level, the highest current is pulled from the least capable charge level. I do not like this. I do not like subjecting this (what I think of as abuse) behavior to my cells. I would rather hammer them down with massive current draw right from the beginning.
And like Jason said, the regulated driver is not capable of maximum output from the beginning. I like Turbo. Not quasi high. Call it a Texas thing. Bigger is better, however you want to look at it my preference lies in the cheaper to make FET drivers. (I can build this for around $5, and I have built hundreds of them for the ~600 lights I have modded as well as for several other folks here.) I have a prototype boost driver or two, and have been told their price level, this is not something I would want to use in quantity, obviously a 5x greater cost would get ineffective very quickly. I don’t mind using them sometimes, but I do find it worrisome the resulting low discharge level if I forget the time it’s run and take the light to the cut-off point. I normally charge my cells when they still have a comparatively high charge, seldom taking a cell down to 3.7V… much less 2.5V! (or even less, as some regulated drivers will do)
I like what I like, I have no need to sell anyone on my preferences or ideas… I use my lights for my own purposes and as I have found over the past 6 years or so, a large part of my purpose is to tweak a light for all it’s worth. I have reason to mod lights for nothing more than the time it kills, the effort it requires, or the final number it can produce… so whether I use the modified light or not isn’t where the relevance lies in the end. This is most likely not how the larger public views things. It is, however, a large part of why I mod. Like asking a mountain climber why they climb mountains and they say “Because I can!”. In great part, for me, it’s like that with flashlights. (What, did you think I really NEED 230 flashlights?)
And yes, if/when I get a D18 I will mod it. I find I don’t really prefer the SST-20 so probably moments after I receive the D18 it will get new emitters… at the very least.
I may even machine a spacer to fit the top end and make it a quad 3V 50.2, under a multi-reflector. This should maintain lumens output and actually keep it sustainable. We'll see...