[WIP] 10mm DD+single-7135 driver: double sided 10440 torture for Dual-PWM

I like the strap idea since it takes battery pressure off the mcu but with just a wire over the top as B+ a dab of epoxy under the mcu before soldering should help support the mcu. The pill pictured above is solid JB weld between the pins and underneath the mcu and took some delicate work with a q-tip and solvent to clear the pins. I’m hoping it can still be reflashed set up this way but don’t have a clue really. That pill is for the smaller of the ladies night clickies(2-mode, 20/80% single 350mA 7135). The light functions perfectly but I had a major failure in the metal work and had to order some different brass.

While I haven’t actually built any 10mm drivers, I really don’t see a problem with the MCU taking the brunt of the physical battery-compression forces. I think that the legs should crush/bend slightly and then the MCU will be flat against the PCB: no harm, no foul.

That looks nice! I’ve got a couple of boards on the way from OSH Park where I’ve had this in mind, purely motivated out of being tired of ripping my lights apart to flash them with new firmware. I’m thinking a slightly bent copper plate to act a little as a spring for the cell, thin enough to allow me to flash it without removing it. Your solution looks good.

Am I seeing correctly? The middle of the 3 legs on the 7135 chip that’s ground doesn’t have ground under it, but there is a pad there… is this correct? That pin can be removed and with the middle of the board fairly clear a small post or pillar can be glued to the pcb. This can be the battery contact point, on the fet side of the board, with a connecting wire from the + via.

The MCU is too thick to be on the battery contact side of my TP.

Anybody know how this fet ranks for this package size?


PSMN1R5-25YL datasheet

I went ahead and ordered some of these, from what I could tell they looked to be about the best choice…not like I’d know or anything though. I’d appreciate it if someone would take a look and confirm or throw water on em. :slight_smile:

Not gonna work. You’ve got to be able to get LED- back to the LED side of the board. Have you tried sanding down MCUs? I have not. I assume approximately 0.5mm can be gained in that way…

All I’ve done so far for this package is make this list over here: [WIP] *15mm* DD+single-7135 driver - single sided Dual-PWM compatible using 3.3x3.3mm FET

Don’t know about this one, but the BLFTINY10 Ver 1.0 works fine that way.

You’re not listening to what I say. BLFTINY10DD V1.0 has a via for LED~~. This board does not. Therefore LED~~ is not available on both sides of the board… therefore if you install it upside down there is nowhere to hook up the LED to!

Remember Dale, smart prototyping will do small-ish runs of very thin boards( .8mm N/C and .6mm for $1 extra). Would a .6 mm board be thin enough to allow you to put the mcu on the battery side? If so, I’d be willing to go in on an order.

Not a matter of not listening, more an issue of having parameters to meet.

There will be an LED negative on the MCU side before I’m done. It might not be there NOW, but there will be…

Thanks Scott, the light was built around a single sided thin driver. Altogether there are, of course, quite a few things I would have done differently if I were commissioning this today. But it is what it is, and I accept the challenge.

Would you please start making threads instead of posting your drivers in other random threads?

You have yet to post any technical information, measurements, etc for the majority of your drivers. They all seem to get announced in places like group buy threads or my own threads and then they never get a discussion thread of their own. Your own discussion threads would be appropriate places to announce the drivers as well as share specs and measurements. And the same thing applies to links to those 4 firmwares, your discussion thread for the BLF10cc driver would be a great place to link to those - I had no idea that there was more than one demo firmware available. In fact I’d forgotten that there was any firmware available until you mentioned it. I was running around thinking that no ATtiny10 firmware existed at all. I still can’t put my finger on anything.

Has anyone measured the dropout on one of your FET+BJT drivers? I’ve been assuming that it’s exceedingly high.

As I see it, the long and the short of your BLF10cc driver vs this one is:

  • I assume that it has very high dropout.
  • I assume that it cannot provide 3A.
  • It has no reverse voltage protection.
  • It has no LVP.
  • It has no offtime capacitor.
  • … firmwares?
  • It’s single sided.
  • It looks like it has a healthy keepout on top, better than the A10DD-L. 0.7mm maybe?

It wouldn’t hurt to communicate that. (Which you have now. ;-))

Cutting off that center pin is fine.

I still don’t see anywhere I’d want to drill or notch this thing in order to bring LED- back through… and since you’d be using the 7135+FET side facing the battery, where you have little room… I think that you can’t really afford to move LED- around the board with an airwire, right? LED- is located right over top of MCU Pin8. I think the best that could be done here is probably to assemble the driver and then notch out the bottom corner where the 7135 is, carefully avoiding the MCU Vcc trace. If that’s what you have determined to do I can update the layout slightly to make that harder to mess up. (eg bring the trace further inboard) Hopefully the 7135’s bond wire for the output pin goes directly from the die over to the pin. If you have a 0.25mm driver ledge this gives you space for a slightly larger than 1.5mm wire to go through the notch.

And maybe more importantly - the 7135 is only 0.1mm shorter than the ATtiny13A. If you’re going to have to file down either one, why not file down the MCU?

I haven’t had the 7135 on there before, the small mosfets used were easily thinner and made the MCU the thickest thing on the driver. There are a couple of places on there that I can drill a hole and run a wire through if necessary, but as you say, if one or the other has to be modified it might as well be the MCU. I can press it down so it sits flat before I solder it on, there’s some extra space right there. And if I sand it down some before mounting it, well, maybe it’ll be enough.

I was looking at the area under the 7135, beside pin 5, for drilling a hole. I can nip the ground ring a little and even the 7135 a little, if necessary. I’ll use 24 or 26 ga wire and keep the hole small enough to make it a snug fit for the wire if that’s the route I take. So much is variable, will be done at the time of installation. I might just notch the edge of the board, slip the wire over the edge. Don’t know.

What size wires would be minimum for the 3A range? Minimizing the wires might help me seat the driver a little deeper in the small pill. There’s actually a very small threaded screw in the driver bay for ground, I removed the screw and solder the wire in. So it’s got good ground, but that’s an extra wire inside.

I will also sand down a piece of copper for the battery contact to get it as thin as is reasonable. Every little place to gain some space has to be utilized. A thinner board may well be an excellent bonus, Scott do you think we can get a small number of these done on that thin pcb?

Updated to v009. I attached the center pin of the 7135 to GND. I think I may have originally intended to do this but never got around to it. Now if you completely grind off the tab and outside corners of the 7135 it’s OK, no problem as long as all 3 pins are soldered.

As usual, this is untested, no version of this driver has ever been built, etc.

[There are no known show-stoppers with v005 or v008, v009 just adds creature-comforts. If you completely remove the tab on v005 or v008 there is enough copper there to scrape some and solder the underside of the 7135 to that.]


I think this will work, I can drill here at 1.2mm and should be able to easily make the connection and clear everything. I can mount the 7135 and MCU as much away from this area as the pads will allow and keep the hole out of the ground trace as much as possible.

It strikes me that if one were really dedicated, one could actually sand both sides of the MCU. It’s over 1.5mm thick IIRC, so maybe 0.5 off of each side? Destroying one is probably the best way to determine exactly how much may be removed.

The lines are small, but if you click on the picture I posted you can see where I was suggesting to notch the PCB. Anywhere else would require an airwire, thus wasting the space gained by flipping the PCB. You said Pin5, but I think you mean Pin8.

I’m not sure how thin you can get away with for 3A… at some point you start to drop too much voltage, this setup will be sensitive to that and current will go down. I wouldn’t go lower than 28 AWG, but I expect you’ll probably want more. I would. :wink:

Why not use the GND ring rather than a GND wire? That should save space/effort…

That’s fine, but just FYI it’s OK to totally break the GND ring there.

I don’t like to rely on the ground ring because it depends on soldering the driver into the pill. I haven’t had a lot of luck getting a good solid solder joint on the thin edge of this copper pill. Difficult to apply with .5mm threads there and the absolute being preservation of the pill. One off hand made parts, priceless.

Ah yes, pin 8, looking at it upside down and backwards. Dayum! I have trouble enough keeping all that straight when it’s right side up and facing me square! :smiley:

Of course: preserving that pill is the #1 priority when doing a build for your TP. But we often achieve GND with only a press-fit in other lights. Is the ledge simply too narrow for that?

I know what you mean. I had to recount several times myself after you said Pin5, I didn’t want to ‘correct’ you if I was the one who was confused.