LED Lenser X21... mighty impressive, but can it be improved?

My employer has recently provided me with an LED Lenser X21, due to my job involving out-of-hours visits to remote substation sites.

Now I have to say, although I'm fully aware of it's flaws, to me this is a fantastic piece of kit, I've certainly never encountered a torch/flashlight quite like it! In terms of sheer output it makes my Passat's headlights seem puny in comparison and the focusing function is excellent - lost of nice floody light for use as a worklight, or a viciously intense beam for spotting (and probably blinding!) intruders.

BUT: as most of CPF seems to be aware this light is far from perfect. To start with, and in my eyes the most serious flaw, it will only run on alkaline cells - at least, if we want the LEDs to have any sort of useful lifetime. This results from the use of a rather cheap-and-nasty direct drive system which relies upon the batteries' internal resistance to limit the current to a safe level (NIMH cells having much lower internal resistance which, in theory, could cook the LEDs).

Another consequence of the direct-drive setup is that the light has only two modes - "high" and "low", where "low" is still rather intense. Also, the mode selection is achieved via some sort of two-position push button switch which either completes the circuit directly or via a resistor - the flaw being that the light defaults to "high" before allowing "low" to be selected.

Soooo... I'm now pondering if this can be improved, and if any CPF'ers have attempted such a feat of engineering. I've quickly measured the current drawn from the batteries on each setting (5.6A high, 0.8A low, although I really ought to make a bit more effort and check the actual current and voltage seen at the business end) and am now thinking along the lines of retrofitting a regulated driver board that can effectively turn the light into a regulated, multi-mode affair - four or five modes would be nice, especially if a lower "low" can be achieved, and without having to suffer the flashing/blinking/strobing/SOS modes that blight some of my smaller lights. And, of course, being regulated this would permit the use of a nice set of rechargeable NIMH cells rather than alkalines.

Thoughts, input, advice all welcomed! Anybody who's attempted to do something similar, or can recommend a driver board that would suit my purposes, or guide me on what to look for in a driver board, etc etc (although "chuck it out and buy Light X" isn't a goer; the only reason I have a £240 light in the first place is that work provided it and none of our official suppliers offer anything even remotely comparable to the X21!)

Cheers
Jonny

Just taken a few measurements at the head of the X21, to get an accurate set of current, voltage and power values.

Measurements were taken using a pair of Fluke multimeters: a model 289 measuring current, and a 179 measuring voltage (both meters are part of my test equipment for work and so are kept calibrated).

The test was carried out using the LED Lenser-branded alkaline D cells that were supplied with the light, the batteries having been used to power the X21 for a total of less than 3 minutes since new and showing an open-circuit voltage of 6.14V for the four cells in series.

High Mode
(note: high mode measurement was taken after light had run for ~60 seconds or so as the initial voltage and current decay rapidly before becoming relatively stable at the values below)
I = 4.7A
V = 3.6V
Power = I x V = 16.9W

(initial measurements were I = 5.4A, V = 3.7V so initially power = 20.0W)

Low Mode

I = 0.77A
V = 2.9V
Power = 2.23W

As the 7 LEDs are simply connected in parallel, if we assume that they share the drive current equally then initially they are drawing 5.4/7 = 771mA per LED; after a minute of running this has dropped to 4.7/7 = 671mA. Which brings me to conclude that, to convert this to a regulated light that can still use alkalines as well as NiMH cells, one would require a regulator board that gives a regulated output current of, say, 5A (i.e. just over 700mA per LED) from an input voltage of between 3V and 6V. Seems a little unfeasible; perhaps a better solution would be a single 700mA driver per LED? (there is plenty of spare room inside the X21 head!)

As an aside I can find no conclusive information as to which type of Cree LEDs are used in the X21 but the layout of the die and lead-in wires suggest that it is some variant of the XR-E series. Would be nice to know which bin code the are though as this would give a clearer idea of the current limitations of the LEDs. Can anyone "elinghten" me??

Replace LEDs with XML, and a Li ion battery pack ---- throw king!!!

I've always dreamed about this mod.

Now that would be impressive! Might be difficult to cool though; 7 XM-Ls driven at their maximum 3A (or even at a "mere" 2A) will generate a lot of heat!

To drive them to full potential we would either need a driver that will supply 21A at something around 3.3V, or one that will supply 3A at 23.1V. That's a hair short of 70 watts delivered to the LEDs, with driver losses to add to that. The battery tube might just squeeze four 18650s into its length (a D cell is around 61 or 62mm long) but even four 18650s would be pushed hard with this level of power (3.7V x 4 = 14.8V; if we assume the driver is 80% efficient it would require 69.3W/0.8 = 86.6W input power from the batteries; 86.6W/14.8V = 5.9A).

Would be a hell of a light though - potentially close to 6400lm if driven at 3A per LED!

(figures based on the XM-L T6 datasheet)

Well you might try the high discharge lithium batteries. They can deliver over 10Amp each if I remember correctly. That would be over 140 watts for 4 18650s. A light like that would be mainly for show off so it would be used in small bursts.

I've been using a LED Lenser X21 for some time but find it frustrating that the alkaline batteries drop voltage very rapidly resulting in reduced lumens. Being from an RC background I thought the solution would be to make use of a voltage regulator that I have lying around and put a 3 cell Lipo into the torch. The regulator can handle up to 36v input and puts out a regulated 6 volts at up to around 8 Amps. I made up a tube to replace the 4 cells and put the regulator and Lipo into it, put it in, switched it on and presto - brightness like you cannot believe!!! A problem though... When I initially tested the current draw with the alkalines it was around 3Amps with some cells that probably had 30 minutes on them. So I figured that I would get about a 20 minute run on a 1.4amp lipo. When in use this turned out to be more like 5 minutes! I then measured the current draw and found it was more like 10 amps. (Lipo's can dump power!) No wonder the torch only managed 5 minutes. For fear of blowing the LED's I have stopped using this system and gone back to the alkalines. As I have now read some of the previous comments I realise that the torch is in itself unregulated and relies on the high internal resistance of the alkalines to limit current. I see also that Led Lenser now have an X21R (at huge cost) which puts out 1600 lumens. My first thought is are they still using the same LED's and the same X21 but just driving them at a higher regulated current with the NiMh cells? Probably. Following on from that then is what current can these LED's take? The next step would then be to regulate voltage and current an use the Lipo. The brightness with the Lipo is way better than with the alkalines and rechargeable - so would be the way to go.
Some ideas please from the electronics buffs out there and the source of a regulated Amp/voltage device.

Here a guy in a German forum modded an X21 with 7 XM-Ls each @ 2.5A (5000 Lumens) with a computer-programmable driver (monitors battery voltage etc.). The mod seems like a lot of work though. Comparison beamshots to an ummodified X21 can be found on the second page of that thread. If you want to understand the text just use google translate.