So, I am looking at a project that will require a buck driver for 3S driving 3 XM-L2’s and will use an e-switch. I would like decent output, will be using good protected 18650’s and will be piggy backing it… maybe on a custom contact board to increase heat dissipation for the driver. I would really like it to be tiny13a based as it will run Star Momentary.
I have never used a buck driver or built a buck driver before, so I am really looking for opinions here!
Thanks Matt
What max size diameter will fit your project?
Personally, I would go 3S cells driving 3S emitters. Bucks are generally around 80% efficient. With 3S/3S, you will have 3 times the voltage overhead over 1S/1S. So you can more easily achieve higher currents in DD.
Basically it is an SRK contact plate. So diameter is not a real problem. I would need to use 12 volt LED’s in a 3Sx3S config, right? The light uses and SRK style reflector, so it will be all flood and a little throw.
What the f√¢k?
I think what ImA4Wheelr meant when he said “3S/3S” is, literally, 3 li-ion cellpacks in series (for a cellpack being any number of cells in parallel) driving 3 “led emitters”. Usually, 3 white led emitters in series drop around 10V at typical max driving current.
What did you meant when you said “12 volt leds”? What is SRK?
Cheers
I think your only option is FET-based DD if you want to use the 13a & STAR.
2-3 li-ion 3-mode no silly flashes 2’8A driver
Reported to ramp output up enough for 2S emitters, which means it can do XHPs and MT-G2’s. It may do 3S emitters when fed with 3S li-ions, yet no available info on this.
Very nicely priced.
Cheers 
Sorry, I wasn’t very clear. I meant I would direct drive 3SxP emitters with 3S cells. You need a very beefy buck to put out a lot of current and the efficiency will be less than DD. Now, With 4S cells and 3S emitters, I would go buck.
LD-33 Triple XM-L 2.6A 9-17V Circuit board
Seems efficient, doesn't it? Is that Gold or Platinum rating? 
Cheers
So, I have zero experience with 2 or 3S lights… But, it sounds like I could use a standard FET driver with the required voltage drop for the MCU and place all 3 LED’s in series and maybe not cook them? If so, that is the route I will go. I was concerned that there would not be enough voltage drop to do this, but around 3-3.2V per LED in series would likely get the job done.
FmC, would there be a better way to go?
Sorry I threw out the “SRK”, that is Sky Ray King. It is a 3P battery arrangement driving 3 LED’s in a 3P arrangement.
BLF Fet driver, 3s cells, 3s emitters is your best option for what you want.
I think RMM does them with LDO instead of Zener, which is best for eswitch.
I got the impression the OP wanted high current. 2.6amps split over multiple emitters is pretty low. And if the emitters are in series, there won’t be enough voltage overhead for the buck to regulate well.
If both the cells and emitters are in series then the current isn’t split, it’s 2.8A flowing through all three consecutively.
^
Yep. My first sentence address if the OP goes parallel. Second, if the OP goes series.
“3S XM-L2’s” and “decent output” is about the only stuff I can clearly glance above this.
That LD-33 is really amazing, man. The 3S cells curve on high is terrific, with 94+% efficiency. Less power lost in the driver means it runs cooler (less wear and less heat on the flashlight’s body) and also potentially a better beast for modding. Chapeau…
I bet it’d also be easily modificable for higher currents.
Cheers
But it’s not Attiny 13A but Pic 12F so no star program. Some are capable of doing the switch but that’s not an option here. There are a few nice drivers around using Pic mcu’s but very little in the way of shared knowledge and programming experience compared to Attiny. The Ax2002 is a nice 1.5A buck driver but it also uses a Pic.
Ok, I did warn ya all, I am a noob to multi-cell serial builds. All the reading I had done indicated that maybe I wanted a buck driver, 9-12V in and 3.xx output and the LED’s wired in parallel. I am beginning to rethink that process. Maybe the route I really want to go is a FET driver with LDO and the 3 LED’s wired in series and that would allow them to make the voltage drop across them, rather than in the driver. In that case, I guess what I am looking for is a driver that would let somewhere around 2.5 or 3 amps out to the LED’s…
Am I understanding correctly?
So, I am thinking this driver from Mountain, configured for 9V and driving 3 XM-L2’s would be the way I would want to go…
FET-2S±MOM-LDO
What would be the expectation for my batteries? Would 3 KeepPower protected and matched cells be good in this build? I think I want to stay with protected batteries in a series build…
Thanks All, still learning and listening. This is quite a bit different that the builds I would normally do.
Yep, that’s the one I was thinking of.
Any (matched) cells will do the trick, but as with any DD driver, they are going to be your ‘rev limiter’ in regards to maximum output.
Is it safe, with good care and feeding, to use unprotected cells? Assuming charged and discharged together and monitered performance.
If you are the one using the light(not loaned out to a ‘floob’(FlashLightNoob)), & are aware & monitoring the cells, then no problem.
The general rule for buck driver is a 2S battery to match 1S 3V LED because when the battery is discharged to 6V, the driving current is still in regulating. Vf (3.2V) + Driver Overhead (1-1.5V) = 4.7V is still lower than the cut off voltage of the battery pack (6V). If 2S 3V LED is driven by 2S battery pack, the driving current will not be regulated the whole discharging cycle of the battery pack. So it is better to drive 2S 3V LED by 3S battery and 3S 3V LED by 4S battery.