How are you doing this? if you were trying to do it based on pictures of the MTN Max, just a tip (you may have already noticed) the lp boards are 2-layer, but the hp boards are 4-layer.
Well actually I just talked to Richard about this project and it might not be something that is best not taken on at this time as I only have a few more days to work on it and it sounds like it could be bigger then that.
Far as overhead voltage I think it would be fine, the losses in the buck circuit should be made up for by less losses in the rest of the circuit due to less current flowing through it. Might not be quite as good at 100% as an FET but should be far better everywhere else. For an XHP I think it would be better as the hard part with them is not given them too much current at 4s voltages.
A properly designed buck driver I think would be superior to an FET, the issue is proper design. Apparently according to Richard buck drivers involve magic to get right.
I am waiting to hear back from him but I might have spoken too early on this project, I let the relative ease of the FET drivers go to my head lol.
Yeah I was wondering your experience level but didn’t want to offend. A buck circuit is out of my league. I know Richard worked a very long time on the Mtn-max, and other members have tried and failed before. Some of that I’m sure was because they were trying to get it inside 17mm though.
It depends on the 12V used. XHP’s appear to sit around 14-15Vf which gives a reasonable amount of overhead. The issue is popping LED’s due to too much current with an FET. I have a dead xhp50 and xhp70 that can attest to this.
Yeah, no offense taken. I simply have a reasonably good grasp on electronics. I am by no means an expert in the field.
When I get something in my head to try, I tend to go after it, particularly if I can learn something from the experience. I have already learned a lot of about the inner workings of buck converters I never knew before that was worth the time invested if nothing else.
Mostly though I just hate knowing that something can be done better then I am currently doing it, it really itches me something silly and sooner or later I have to scratch it.
Well after talking to Richard I think the best idea is to wait to work on this until I have a lot more time to mess with it.
He did give some good suggestions such as using the LM3409 instead of the MAX1620 buck converter.
I do still think that a buck driver is the best option for 4s voltages but for now I am going to focus on polishing up the FET 46mm version in this thread: Texas Avenger "TA" Driver series - Triple channel + Bistro or Narsil + Clicky or E-switch - The Ultimate open source driver!
It should work fine for the XHP50/70. I think a quad XHP50 Q8 with an FET driver will still be something to behold. Just don’t think that an XHP35 will get along well with the FET.
So thanks for your time, sorry to get ya’ll hopes up. I just think my limited time is better spent on other projects at this point.
I am real curious where the possibility of a triple channel Q8 driver goes.
Nothing wrong with diving in and learning as you go. Gives me a chance to learn with you. Your buck concept is definitely worth exploring and I hope it can eventually be brought to reality.
I certainly appreciate all the stuff you’ve been up here to TA. The hobby has already benefited greatly from your contributions. Sure hoping Tom E adds the third channel to his FW. I have lots of momentary lights that I would like put your latest DD driver in.
Yeah, I am not giving up on it entirely, just going to finish up some other projects first before I dive into this one. I can live without an XHP35 SRK for a little longer….I really can….
I see china lights with what looks like super simple buck drivers in SRK and even when modded to 15A+ they still work fine, so I know it is not impossible by any means, Just something that would take more time and prototypes then I can afford right now. At $15 a prototype it could get expensive since it is unlikely that the first try would nail it according to RMM.
Thanks for the kind words, I have got a lot out of this site so I figured the least I could do is be open source with some driver designs. I am just glad that others find them useful.
If anyone can think of another size for the TA driver series please do let me know, in that thread naturally.
Because of the prior conversation regarding XHP-35 emitters, I thought I’d add this…
This afternoon I modified an LD-2 driver and put it in a Convoy L2 with 2 cell tube, 4 of the 26350 half cells to an XHP-35 E4 1D emitter on a Cutter 20mm star stacked on a second copper star. Results? The L-2 XHP-35 E4 1D, 3.13A at 2942.85 lumens with the half cells at 4.21V!
That’s the most amps and most lumens I’ve seen one make! New PB! Where’s my Gold Medal? 
Seriously though, it’s quite intense and it’s not even an HI, dome on HD. 170.25Kcd for some 824M throw. Learn something new every day… forget several old things in a matter of minutes so it’s a backwards curve. 
Edit: In retrospect the dome on is probably why it’s making 2900 lumens, all my other built lights using the XHP-35 are HI variants.
Very interesting, although this makes a lot more sense based on what cree rates them. This is right in line with what I expected them to make actually.
That said you might be onto something on the Dome possibly being to blame. Although it is the same bin it is possible that it still produced more lumens. Odd it is this much more though.
Has anyone ever tried a V2 XP-L dome on vs a hi version to see if they line up lumen wise? I think they would as my HI’s line up with ~28% loss in lumens you would expect vs the HD (in fact less then that in some cases).
Remember that in the XP-L, Cree automatically deducts in the binning for the lumens losses… Hence an XP-L V6 becomes an XP-L HI V2. 
Edit: It’s in that range anyway. I can remember building an FET X6 with XM-L2 U4 1C that made 1811 lumens, de-domed it and saw some 1500 and ugly tint. So I pulled the de-domed and went full dome again for max power. So they’ll all lose some lumens and take a tint hit when losing the dome.
They do this for the XHP35 as well, officially in the data sheet the HI version only exists up to D2 or D4 bin vs the HD E4.
Are there some F4 bin HD XHP35’s floating around by chance?
It is rather strange that the ratings would be this far off. I might just have to risk an LED, maybe just test it to the official rating of 1.05 amps to see what it does.
Well I am not sure what to make of it. I just tested a 3c E4 XHP35 HI and it made ~1900 lumens @ 2 amps and was not gaining a lot with more current. I am guessing that it would have maxed out around 2300-2400 @ 2.75 amps or so.
The odd part is it took twice the current to reach the rated lumens. That is not like Cree, they are usually very close to spec.
On the plus side it is still almost twice the lumens of an XP-L HI and should net similar throw figures.
I need to order an HD version to test side by side.
I put an HI variant in my MaxToch Shooter 2X and it didn’t quite match throw numbers but of course it’s making a lot more lumens so it’s all good. Had to machine a 2x battery extension tube so as a 4x 18650 it looks somewhat funky, but boy does it work well! lol
Will be trying one soon in an L6, probably shave the HD dome off…
At the very least the XHP35 is the brightest “single die” LED we have that doesn’t have issue with donut holes and can net some pretty good throw figures too boot.
If only it was a 3V setup.
I’d be happy if it had a 6v setup… I’m not sure why they couldn’t do the same as the others; maybe the XP footprint didn’t have room for separate pads?
If we get a smooth reflector for the L6, I just might have to buy 4 26350’s like Dale and do it anyways
Dale, did you get yours from Banggood? The generic purple ones? If so, how are they?
Yeah PD, they’re kind of a light lavender, marked as 2000mAh.
I haven’t tried a direct FET driver on one yet to see what it delivers in comparison to other cells but need to do that. I got 10 of em, figured I’d find uses for em all pretty quick and wish I’d gotten more. lol
I CAN say that the L6 with an L2 single cell tube and 2 of these is a sweet combo! 
An Ultimate test…
I have a TangsFire TS2010 with SBT-70 and an FET driver. The 4000mAh gold King Kong does 6.71A (rested, probably not a full charge) One of these little 26350’s will do 6.98A! Top dog cells allow considerably more of course, but they’re not too shabby.
Edit: The King Kong cell is at 4.12V while the 26350 is at 4.19V. Still…
Edit II: 4.40A to an XP-L in a F-13 with Bistro X6 driver, same cell.
That is quite impressive for the size. Glad I grabbed a couple. I tried the L6+single cell tube the other day and agree it is an interesting size light and amazing performance for the size.
Hmm now I need to figure out what little 26650 light I want to put an MTG in…. An L2 with MTG would be sweet…. but I already have one and I never buy two of the same light.
TA they fit in your L2 tube lengthwise? I know Dale already bored his out