DEL's OSH-Park driver boards

These were done around my preferences and criteria, but I thought others might find them useful as well. Most are based on the popular BLF FET ‘direct-drive’ + 1x7135 design. There is also a simple 3+1x7135 board. They all include pads for R5, the added resistor that solves the MCU instability issues in high-power lights. R3 is also there, the FET gate pull-down resistor that is needed when a very low gate voltage FET is used. 20 mm and up boards include extra stability components (C2 and R4) and the MCU pads are sized to fit anything from the tiny13A to the tiny85.

None of these boards, however, have pads for an OTC (off-time capacitor). So no conventional off-time memory or medium-press detection on clicky lights.

The board lay-outs are about as good as it gets (electrically) with what is basically a single-layer design. I have built several of the 17 mm boards and are using them extensively without any issues. I also have two M6’s running with rev00 of the M6 board, real happy with how those turned out.

DDm-17 rev04 - Built and tested.
17 mm DD+1 driver.

DDm-20 rev01 - Built and tested.
20 mm DD+1 driver.

DDm-22 rev00 - No plans to order.
Added 5 Feb 2017. 20 mm DD+1 driver.

DDm-Y3 rev00 - Built and tested.
23 mm DD+1 driver for the Yezl Y3 (1S). Ears need may need slight sanding to fit.

DDm-M6/SRK rev01 - Built and tested.
46 mm DD+1 driver.

DDm-M6/SRK rev02 - No plans to order.
46 mm DD+1 driver.
This one gives the option of one to eight 7135 chips for the regulated channel.

1400m-17 rev00 - Built and tested.
17 mm CC+1 driver.
This one is a favorite of mine for Convoy S-series lights! It is what I use in my EDC S3. Two regulated channels, based on Harlequin’s HQ4D board. Note: Only good for 1.4 A.

DDm-L4 rev01 - Built and tested.
DD+1 driver for the Convoy L4 (1S).

5.6m-SRK rev01c - Built and tested.
46 mm CC+1 driver for SRKs with stock non-DTP mcpcbs.
2-Channel, up to 5.6 A.

Pretty Purple PCBs :laughing:

Schematic for the DD+1 driver models listed above:

And these two oscilloscope screen grabs show why R5 was added during testing of the initial Q8 driver. With R5 = 0 ohm:

The yellow trace is the power feed of the micro controller at its pin 8.

With R5 = 2.2 ohm (I am using 4.7 ohm now):

Well done :+1:
These finished boards look great.
Oh, by the way thanks for your research of the R5. :slight_smile:

Thanks HQ. Hope you do not mind my appropriation of your HQ4D board :slight_smile:

I actually ordered and built a few of your boards, but OSH missed the vias on half of them. So I just added a few of my own design on my next order. Great balance of power and heat in a small light.

Hi Del,
I don’t mind at all. I didn’t invent the wheel with the HQ4D either and just adapted existing ideas for my needs.
Out of curiosity: Which vias were bad and how? I encountered all kinds of oshpark glitches, but til now the vias were nor affected.

They simply missed all of them on some boards. No vias at all. That was my only OSH Park glitch up to now.

I dont understand it at all DEL but do know its good work. :+1:

Wow DEL! That is a bunch of boards! I’d seen a couple of your boards at OSH Park, so I knew you were working on some. But I didn’t know you had this much going on! Impressive! Thanks for sharing these with us!

very clean layouts, good stuff.

It would be nice to update RBD’s Oshpark thread with a link to here.

Thank you, DEL, for all your contributions, research and help. :beer:

That thread is long overdue for a massive update. :person_facepalming: I’m not really sure of the best way to keep its op from getting unwieldy. It’s very difficult not only to edit a single already LARGE post but navigate it to find something. So many drivers only slightly different from each other, not to mention all the other project pcb’s. I’m inclined to just post author thread links under linear, buck, and boost and let the authors organize their own op descriptions and links and accept that there will be a fair amount of redundancy. Given that no one of us has editorial control to organize the separate threads into consistent layout under one umbrella it’s always going to be a patchwork but I need to do at least something to bring it up to date regardless of the result.

Thanks RBD. I do use your thread regularly and assume others do as well.

Nice work. Each new design might seem to add an unnecessary redundancy but to me that’s part of the evolution of the drivers to find out what works best. Having Oshpark available to try them in piecemeal quantities makes for a hand in glove relationship. It’s just difficult to keep track of.

Thanks for the kind words everyone. I hope to eventually add schematics and also scope traces to show the advantages of the few extra components.

Of course if you need an OTC, or would like to have the full complement of stability components on a 17 mm board, or a 3-channel board, TA is your man!

I will add a link on RBD’s thread as requested.

Those who design boards will always make new versions for their needs. Be it a small tweak or a big change. But if it’s shared, those who are not accustomed to board design will benefit from this diversity and the designers get new ideas. I’m always happy to see new designers joining the herd and new designs coming up.

And this, RBD, is where your thread is the center of it all. Be it patchwork or not, it’s invaluable.

Which reminds me I’m overdue with enlisting more than one board to your thread myself. :person_facepalming:

Nice! What do I need to add/change to get it going for a 2S setup?

Have to add a resistor and Zener diode in the same positions as on the ‘standard’ DD drivers. Here we have the luxury of C2. So swap out D1 for a 220 ohm resistor and replace (or stack) C2 with a 4.7 V Zener diode.

The same precautions hold as usual:

  • This is only good for 6 V emitters.
  • The 7135 will run hot at high PWM, but the 7135 has a good surface of copper for itself on this board and it should be OK.
  • Parasitic drain will be up to 20 mA while the tail switch is on.

EDIT: If the voltage reading is important, I would lower R5 to 1 or 2.2 ohm get a more accurate voltage to R1/R2. This is necessary because of that 20 mA parasitic drain.

Great, thank you! Ordered!