Custom 20mm high current Nanjg style driver

After doing my custom SkyRay King driver board ( ) I wanted a 20mm driver that could do some high current for the Defiant Super Duper Pooper Thrower. Well, 20mm drivers are few and far between and very few of them can handle decent current levels.

So what’s a boy to do? Why build their own of course…

The regulator board. 12 AMC7135 chips on a 20mm board can handle 4.5 amps. The regulator chips are mounted vertically. The legs solder into through-holes. :

The CPU board. 17mm diameter. Connects to the regulator board via a 4 pin header. The CPU can be a ATTINY13 or ATTINY85 chip. The pinouts of the CPUs and the components are compatible with the Nanjg 105C regulator and any of the DIY firmwares available on BLF. I’ll probably write my own firmware… maybe.

I might also lay out a single board version that can use 8 regulator chips.

so this would be a 20mm single cell high current board sandwich? given how many budget lights have a 20mm board, please say you’ll make this available, maybe fasttech, cnqg or intloutdoor would buy the design or pay royalties on it? I’d certainly be ib for a few for stock.

Yes, that is why I built it. Soooo many lights, soooo few drivers… I don’t know how many lights the board height might be an issue with. That’s why I’m thinking of also doing a single board version with fewer regulators (hopefully lying flat so that they can be stacked).

Even at OSHPARK prices, the bare board are cheap… about $1 for the regulator board and $0.70 for the CPU. The components area 0805 size SMD parts for easier soldering.


so aside the 7135’s, which obviously we’d need more of, can we just farm a nanjg driver for the other parts? I can see my m10 getting a driver/emitter update, the other board I’d love to shove into my uf-v3. 4a on a copper mounted xp - g2 should be very very interesting.

Oh my god. If you make these and the srks buyable you are the New King of flashlights

Must be nice to just design what ya need when it’s not commercially available!

I had issues with a custom build that required a 10mm driver. Ended up going lower power than was planned. Works well, it just doesn’t have the punch I was looking for. The Mini FluPIC works good and supplies 700mA, which is not the same as the 1.4A target.

You think you might design some drivers for the MT-G2 emitter? I just got the S2200 and it’s great, but a bit underpowered. I got a tail cap reading of 2.2A off of protected AW 2600mAh cells. Comfychair upgraded the S1100 to an MT-G2 and found a dielectric layer between the emitter and the copper star, might be interesting to put a true copper star with more power under that MT-G2 and see what it’s really capable of.

This! Count me in for both types if made available. Thanks for all the effort, TP! 8)

You could reuse the CPU, but the stock ones have crappy firmware/PWM. Dr Jones’ NLITE firmware is MUCH nicer. He could also burn his programmable Lupodrv firmware into one.

The passive components on the 105C are tiny little buggers and are very difficult to work with, hence my choice of larger sized ones.

I have the single board one laid out. Just need to clean it up some…

ok. so any chance of a stock list of parts with recommended supplier, obviously mither doc jones for his program/programmed chips and an assembly drawing?

Depending on cost I would be in for anywhere from 1 to a huge pile.

19-20mm 3-3.5A 3 cell buck driver next, please. :party:

@ Pyromaniac - you keep this up and I will have to drive over to your place.Tongue Out

These ideas sound good, looking forward to seeing how it all turns out.

Are those *s for modes? Nice looking boards. DBCstm, I’m not sure if a nanjg type board can power a mtg2 from 2 cell liion, too much overhead for the 7135’s. Maybe lower V Lipo or 5 cell nimh but any of them would need regulation for the mcu.

Yes, but I do need to re-label them to match the standard NANJG naming… the stars are connected to the same processor pins the NANJG drivers use, so any NANJG compatible firmware should work.

Thanks RBD, of course I was thinking 3 cell Li-ion but I probably don’t have the knowledge to make such a mod anyway. Better I leave it alone! :slight_smile:

Tell me about it. It would be nice to have an IC that incorporated all of the passives except the 7135’s since those parts seem to be set.

It would probably make life more difficult if they were all in one package… you would have many fewer layout and routing options available. Drivers are very small and densly packed. Plus they are essentially single sided circuit boards. Components on the back side have issues…

Going off track here, but I've brought it up a few times and never got a concrete answer unless I overlooked it...

I know the current regulation side of the 7135 only cares about the relative difference between Vf and supply voltage as far as what it can handle, but what about the Vdd pin? Like in my MCU-less MTG2 driver with 2x16340s (lots of sag, which keeps the voltage difference in a tolerable range), are they OK with full input voltage jumpered onto the Vdd?

No! You might be able to get by with a voltage divider driving the Vdd pin.

BTW, the way AMC7135’s are used in flashlights operates the chip in ways totally not specified in the data sheet. There are no specs on things like what happens when you pulse the Vdd pin to PWM the light or how the chip will operate when the Vdd pin is driven low. I have seen AMC7135 chips that allow a couple of milliamps to flow through the LED when the Vdd pin is used to turn off the LED. That behavior would be a killer for my SRK driver since its parasitic drain depends upon the chip drawing essentially no current when the LEDs are turned off.