[WIP] *15mm* DD+single-7135 driver - single sided Dual-PWM compatible using 3.3x3.3mm FET

This is a single sided driver which makes use of an LFPAK33 or other similar MOSFET in tandem with a single 7135. It is intended for use with the Dual-PWM feature in JonnyC’s line of STAR firmwares. See the bottom of the OP for the driver’s life story. :wink:


  • Unattractive appearance.
  • ~1.0mm ‘tall’ component keepout around edge. (7135 tab extends to no closer than 0.5mm of the edge.)
  • 0.95mm electrical keepout around the edge. I just couldn’t quite swing 1.0 this time. It’s a terrible letdown I know.
  • FET + 1*7135: alt PWM is hooked up to the single 7135
  • Full solder dam for the 7135’s tab - makes assembling this small driver easier.
  • 3.3x3.3 sized MOSFET, recommend LFPAK33 because it has exposed leads.
  • Offtime ready.
  • Large 1.5mm GND ring on top & bottom (where possible). Exposed copper is 0.7mm on top and bottom.
  • Like some other recent drivers, this driver does have a reverse-polarity protection diode but bypasses the diode for the voltage divider. Due to this we are able to use any diode desired. Ideally we use a protection diode with as low a Vf as is reasonable.
  • The exposed spring pad is 10mm. Total pour diameter is 12mm.
  • LED +/- pads are 2.0mm in diameter. They are precisely positioned across from each other.


  1. This driver may require very minor shaving of the SOIC clip near the FET, depending on what type of FET is installed. I do not expect shaving to be required for LFPAK33.
  2. This driver should sand down to 14mm w/out issue. 13mm is too much.
  3. Please do not ask questions about potential shorts seen on the OSH Park preview without checking the Gerbers in gerbv.



Hand assembly recommendations:

  1. Solder FET
  2. Solder 7135
  3. Solder MCU
  4. Solder R2, C1, OTC
  5. Solder D1
  6. Solder R1
  7. Solder LED wires

RMM, NeilP, ToyKeeper, and others posted about 15mm FET drivers over in RMM’s sales thread. This isn’t a case where I’d even considered making something along these lines. As with another recent project, I didn’t think the stuff would fit! Until RMM said that he was working on one that is. I still wasn’t really interested though, I figured he was using a baby FET… Then RMM stopped by the [WIP] 17mm DD+single-7135 driver thread and asked if I was playing with LFPAK33 at all. Looking back I’d say that it was right around that point where the one-upmanship kicked in… a few hours later (after laying out a functional 15mm FET+7315 driver) I dropped RMM a line to make sure he was OK with what I’d done (he was). Since RMM has been influencing my drivers recently and this one was 100% motivated by RMM I figured it would be a good time to mention that on the driver itself.


Here’s what I showed RMM (v006) next to v010:

I wasn’t really thrilled with the LED+ pad in v006, but nothing really seemed to fit better until I decided to move C1. Once C1 was moved the LED+ pad fell into place nicely and easily aligned with LED-. SOIC clip tolerances around the MCU are very tight, so while it may appear that there is some room to fudge on this driver that’s mostly not true.

This is great…

Things just keep getting smaller…

Now how about a 12mm with the same capabilities? (double sided of course) :wink:

Like this? [WIP] °12mm° DD+single-7135 driver: double sided 10440 fun for Dual-PWM

I’m just going to assume that you had already started on this before I said something… otherwise you are ridiculous

Are all of your designs going to switch to the 33 fet now? Should I bother buying any of the bigger ones?

Very nice, it is going to be interesting to see what future builds this will enable.

I started a few minutes after you asked about it. I didn’t have anything else I wanted to work on at the time. I almost stopped about 40min ago when I zoomed out and realized that I’d left out R2 entirely. That was a pretty rude shock! :frowning:

No, I do not plan on switching larger designs to LFPAK33. There’s significantly less variety and quantity available for the smaller LFPAK33 size. If LFPAK56 fits I will continue to use it.

oh wow…nice

Very nice driver, I really like it, thanks for doing this, wight! So with the right firmware this is a 15mm driver with modes ranging from sub-moonlight to true direct drive. That is the almost perfect EDCflashlight-driver (the perfect one does all this and also boosts from an single eneloop or alkaline ;-) ).

There are several 15mm lights out there that could benefit from this driver (+leds on copper), not just the sk68. My favorite 15mm driver light at the moment is this one.

But this also gets to a point where I get pretty hesitated by all the DIY-stuff that is required to make these new drivers work. I hope, if the driver tests as spec'ed, that Richard is going to sell these assembled, and for a reasonable price :-) .

djozz, how is the COFLY KX-H10? solid pill? is it easy to dissemble?

Thanks! I’m pretty happy with the functionality, just not the aesthetics. I hoped to at least have enough space to line up the 7135/MCU/MOSFET down the center, but no such luck. Beggers can’t be choosers though, I’m thrilled it all fit w/ my normal goals and on one side.

Thanks djozz. I’d like to do a boost driver, but low-voltage boost is off my list until I get a buck driver working. :~

I don’t have any 14500s or 10440s, all my stuff is 18650 and larger or runs on a conventional AA.

Is there something holding you back about this driver in particular? It should be a very simple reflow. (The only thing to watch out for of course is putting a little too much solder paste on the gate/source pins. No big deal though…)

It is very easy to disassemble, and it has a hollow pill. There's a great dislike for hollow pills on BLF, but I use my leds on copper boards, and for small lights (and most 18650 lights for that matter) the (thick and copper, if Noctigon or Sinkpad) ledboard itself is way sufficient for the heat transfer to the pill. And I just use the plastic press-fit disc to fix the ledboard, alternatively you can use an adhesive like Arctic Alumina Adhesive.

No, the reflowing itself is not what holds me back, I think i can get it done quite nicely, it is the yet again new (mostly magic to me) components that I need to find in small quantities with affordable shipping to the EU, and my unability to flash my own firmware (I am not much into electronics and in computers, even do not have a proper Windows one).

I think there's many folks out there that would love to use all those great drivers, but despite of not being afraid of some DIY soldering, lack one or two skills to make a functional driver starting with a bare Oshpark board.

Oh hell yes. This is going directly into my titanium D25A. And hopefully also into the brass AA lights from CNQG, if it’ll fit. I think I have hosts for about half a dozen of them.

The D25A will probably get special firmware, since it has a sensor for fully-tightened vs slightly loose. Not sure what to do with that yet though.

Thanks djozz, i have seen your measurements all over this place :slight_smile: so i have to be honest, when you say it is possible enough with the hollow pill it gets me curious, because you are right about the great dislike about this here, and me as new at this have almost adopted this as a truth, without any real personal experience backing it up.

So i wonder have you used any thing in that size with a fet driver and didn’t feel it lost a lot of output fast by bad heatsinking? i
thought a solid pill/shelf where almost a requirement to not have it sink like a stone in lumens output.

A DTP MCPCB is a requirement for that, and so is a decent path from there to wherever all this heat is going. Beyond that, more contact the better of course… ideally the whole chain has very few thermal interfaces and the interfaces are all good.

Interesting this means i have disqualified some possible nice host wrongly then, i will keep this in mind for future builds, thanks wight and djozz for setting me straight.

Way to step up, wight. When I get a chance I need to update the Oshpark projects thread with all this new stuff. Won’t be anytime soon though.

I think it might be a good idea to mention that one reason for bypassing the diode to the voltage ladder is to allow Zener mods that still have lvp with just a change of R1.