comfychair-inspired quick+cheap+lazy single-sided 17DD FET-driver (poor man's nanjg92)

Yes, I did this for many times.

No, everything stays in place. Pull off the AMCs and add the Mosfet. That’s it…

Prisma, it isnt necessary to relocate the capacitor with the qlite and the Ak47?

For the zener mod, do everything you would do for wight's board, it is effectively the same driver. So I guees the cap should indeed be relocated, parallel to the zener diode.

This thread goes to my book :)

Does anyone else have this with this driver, which should not work different from wight's designs?:

Two cases of builds with this driver:

1) (small 16340 light with 219C) with the tail in place I get normal behaviour, the light works like a charm, but with the tail removed and while measuring current with my DMM the higher modes do not work, when tapped for the next mode it does that for a flash and then returns to lowest mode. Then I used a 0.01Ohm sense resistor for current measurement instead of the current going through the DMM and all the modes were back.

2) (Supfire F3 build with XP-E2) the stock tail worked fine (and was good enough for this 2A build btw), but when the tail was modded with a better switch and spring bypass, it would not stick in high mode. Low and medium worked but when tapped to high mode a quick flash and then return to medium.

I am not enough into electronics to understand what happens, perhaps if the minus wiring from battery to driver has too much capacity something happens??

In both cases when it doesn’t go to high, those are cases where you potentially may be drawing more current through the FET?

So maybe in those cases, because it’s drawing more current, it’s causing the voltage to sag momentarily which may be causing the mode switch?

Then, in the other cases, there’s more resistance, so less current draw, so no sag, so you can go to high mode.

A lot of maybes, but that’s a guess.

From memory some of the original HD2010's had a problem when testing not being able to measure current on some modes with the tailcap off.

I got one failure out of 10 of this driver. What could happen? I am really not sure cause it worked the night before I tested it. And next day did not. LED is OK.

Since I have nasty habit to hard solder driver to pill only way I could de solder my work is with heat gun, and of course it pops out after few minutes of heating, and so does other components on driver :)

I guess that I probably did bad wire soldering(22AWG)

I got one very important question for this driver(it is in picture):


Question in picture: No, the tab / back of the FET and the black wire is “out” to the led. Ground ring of the driver is “in”. Solder them together and you bypass the driver entirely. Same as if you did straight direct drive with no driver.

If I understand the question well, yes the FET will stick better, but you do not have a driver anymore: batt+ was already directly connected to the led+, now you propose to connect batt- directly to led- : direct drive bypassing the FET. The FET is the switch between batt- and led-, the switching is directed by the MCU.

Fixing the FET is a good idea because it hangs on a few thin leads on the side and could be ripped of by the led- wire when assembling the pill, best is to use some epoxy or ArcticAluminaAdhesive or whatever.

Halo beat me to it :-)

Thanks on explanation guys. I am real dummy for some things. But I can always ask you guys. :)

I use artic alumina(don't know if you guys tried that in this build). But it should be done in right order.

I suggest solder FET black wire before gluing with AA. Why? If you firstly glue with AA and then try to do soldering wire to upper side of FET AA will unstick itself cause it is heat sensitive.

One more question... I bend FET legs for better direct contact to soldering points on Djozz picture. Do you guys do the same?

Yes, I do bend the legs with small pliers. The last few I did I bended the third leg sideways and included it in the big solder blob to the ground ring (in the OP picture it just hangs loose). That makes it just a little bit extra solid.

I have a couple of the ’generic nanjg with 8x7135 from fasttech.
Does this mod work with those?

Also I have some AOD510 MOSFETs. Can I use those fr this nod?

If so I guess I strip off all 8 7135s, but then what?

Thanks,
Jim

Edit: I did the mod, but not working… No light. Which MCU pin should go to which pin on the AOD510 MOSFET?

Just took a look at the AOD510 datasheet for you. AOD510 appears to be opposite of the PSMN3R0-30YLD. So the bottom left pin of the AOD510 goes to pwm.

Instead of mounting it upside down like the PSMN3R0-30YLD you could place it right side up if you have some kapton tape to insulate the bottom of the AOD510 from exposed pads on the driver.

If you have no way to insulate the bottom you could place the AOD510 upside down but it would need to be oriented differently from the PSMN3R0-30YLD.

Thanks. I think I figured out what’s going on (in my brain).

Awhile ago, there was discussion about a new version of the EAST-092 driver that was showing up from FT, and there was lots of discussion about replacing the MOSFET on the “new” EAST-092. The AOD510 was one of the ones I think comfychair had found and was recommending, and THAT was why I have a couple of the AOD510s.

Later, comfy started the thread about the NANJG-92, which was making an EAST-092-like driver from a NANJG driver by removing the 7135s and then adding a MOSFET upside-down. The pics on that thread and also the Ebay link to the MOSFET on that thread are gone now, so I don’t know exactly which MOSFET was linked originally anymore, but bottom line is that I got the two things (the choosing a replacement MOSFET for the “new” EAST-092 driver vs. the NANJG-92 driver build) kind of glommed together.

Having said that, since I still have the AOD510s and a couple of the NANJGs, can I use the AOD510 to make a NANJG-92? I think I need to know which MCU pin should go to the AOD510 Gate pin?

Thanks for helping UNconfuse me :)!!

Jim

EDIT: Could someone ohm out and tell me which MCU pin # should connected to the Gate on the MOSFET? I know that it’s the PWM pin on the MCU but which pin # is that?

Bottom left is gate which goes to the pwm pin on the attiny, 2nd pin in from the capacitor.

Hi,

In the pic in the OP, the “dot” in the corner of the MCU should designate pin #1, then you count down that side then up the other side.

So, if I’m understanding what you said, that’d be pin # 7 on the MCU?

The Nanjg board I have in an 8x7135 one so it’s laid out differently than the OP, so I’d need a pin # since the cap may be in a different place.

Thanks,
Jim


Actually, looks like it should be pin 6 :zipper_mouth_face:
Yet, in the pictures above pwm seems to be connected to 7.But I’m not sure I even know what I’m talking about at the moment. :ghost: I need sleep. :weary: The schematic picture is correct at least, fwtw.

Hi,

I’ve checked and double-checked, and on the board I have, it surely looks like MCU pin #5 is going to the 7135 input (Vdd) pin (== the Gate on the MOSFET replacement)? I soldered the Gate pin on the AOD510 to the Vdd pin on that one 7135 location and ohmed it back to the MCU and it comes up as pin #5 still.

Well, at least now I have light, but I’m not getting modes, so I’m suspecting I’m only getting light either because of a short or because I’m in DD.