How-To: Add 7135 chips to a Driver Board (Stacking)

It won’t help in the flashlight but will tell you when you have modded a board correctly so you’re not guessing. The last boards I modded I couldn’t get the current I was supposed to until I took the switch out of the circuit and then it was spot on 10 x 380mA. I have a new 10A switch to install but at least I knew there wasn’t a problem with my soldering.

I agree Rufus. :beer:

I measured 4.5A last night in a XinTD C8 v4 with 12 380's (4.56A) on 3 different cells: SONY 30A, Samsung INR 20R, and Samsung INR 15M. The 15M was pulled from a power pack. The XinTD has a XM-L2 U2 1A de-domed on a Noctigon, screwed down with AS5, lapped surfaces, 22 gauge wires cut to min size, copper braided springs, all contact surfaces treated, copper heat sink added to inside of pill. I'm now convinced I can probably achieve higher than 4.5A on a single cell XM-L2 Nanjg based driver, so plan is to attempt that. I'm amazed it held the 4.5A for quite some time, nothing formally measured, and lumens measurements showed not much loss in 30 secs.

Rufus - I've never seen a switch be a limiting factor on amps, always seems to be the spring, even on that cheap, small switch in a XinTD X3. It's easy for me to compare in a lightbox - put the light in assembled, test it. Then right in the lightbox, remove the tailcap, heavy wire to jump the cell to the housing, and compare. When I see a loss, I copper braid the spring, solder the spring up if it's lacking, and it recovers all or 90%+ of the loss every time. Not sure if this is a 100% legit test though, because even the heavy wire with a simple touch contact is not ideal, so there may be more lumens/amps to gain I can't measure so easily.

Pana PD's are good with their capacity, but nowhere near the ability of a SONY 30A, Samsung 20Q, or Samsung 20R in achieving top high amps. I'd recommend at least a pair of 20Q's or SONY's to try if you are doing a FT order.

Rufus - +1 on the copper wire in the ground tabs -- gotta try that one!

That's brilliant! Thanks for that tip!

-Garry

BTW Rufus, do you glue the 7135s down when you do that? Otherwise, how do you place the chips with the wire, and keep stuff in place while you’re soldering?

I missed the “remove the spring” and add the “inverse” spring part, so I’ve only been able to get to the 7135s on the top side (the side with the emitter leads), because I have one of the 20mm partition boards with the hole soldered onto the original spring.

I still say that both Fujik and AA seem to melt and the 7135 moves around when trying to solder the 7135 tabs though?

You can get them on Amazon or Ebay.

EDIT:

http://www.ebay.com/sch/i.html?\_trksid=p2047675.m570.l1313.TR1.TRC0.A0.Xsal+ammoniac&\_nkw=sal+ammoniac&\_sacat=0&\_from=R40

Well, since I always use bent pins, simple, easy pressure from a tweezers is enough to keep the 7135 in place - much easier with bent pins I guess. Absolutely the Hakko 888 improved my soldering "skills" as well. In fact when I accidentally cut the cord to the iron a few weeks ago, I tried with my old crap Weller, but within 5 minutes, ordered another Hakko 888! I use just the stock, small wedge tip to do the 7135's with.

I may just try it next time without bending the pins, as relic38 says, the bending may cause internal damage, but certainly the soldering would be more difficult. Have to pre-tin the legs though.

Sorry - deleting duplicate post.

I got some 14AWG wire from HD, and I tried a variant of Rufusbduck’s idea. Instead of laying the wire across the 7135 tab, I soldered the end of a piece of wire to the bottom 7135, then used a flush cutter to cut the wire to the height of the 7135 (careful! That wire will FLY when it’s cut):

I lined the 3 pins up (pins on 7135 were bent, fluxed, and tinned):

Soldered the tab from top 7135 to the end of the 14AWG wire:

Then soldered the 2 outer pins:

It’s a little more work, but at least for me, it took the “guesswork” out of stacking the 7135. I don’t know if this could be used to stack more than 1 on top though…

Currently with 14x7135. Still need to improve the wires from driver to emitter.

Hhmm - If you are doing the bent legs, I wouldn't pre-tin the legs because it interferes with seating the top one. Also for me, putting the driver in a bench top vise works so much better than those 3rd hand things, also much quicker for me, real solid position, no chance of anything moving. Also I always do the piggyback's with the wires off - de-solder the stock ones, do piggybacks - do close inspection on solder job w/20x mag, do continuity testing on all pins, then solder on heavier wires (usually 22 gauge, 24 gauge on EDC sized lower amp lights), rig it up on a simple LED test fixture (XML), connect up 1 cell and verify the amps is what I expect and the modes function (custom programmed of course!).

Just say'n....

Using a vise instead of a helping hand is a good tip! I’ll have to try that.

I don’t know if it’s the same with all of the 3.04A drivers, but with the ones from IS, it’s hard to tell where the emitter leads are soldered to, so I’ve been trying to avoid removing the leads. They kind of look like they’re solder to a pad that’s connected to one of the IC pins.

On a previous board, I had to solder on new leads (26AWG), because one of them got squashed when I put the driver into the head of a light (S5) I was modding. I just kind of “guessed” which soldered area the leads were soldered to, but it seemed to work.

Since the chips are all in parallel you can solder led- to any of the 7135 output pins. Led + has its own pad next to the mcu input polarity diode but could be soldered to the battery + pad if for some reason you want a really thick wire and can solder to the other side of the board.

BTW Tom, that driver in the pics I had is another IS 3.04A board. I had it hooked up to a bench supply, and the supply was reading 3.4 amps, and the emitter current using a clamp meter was actually higher that that, like 3.5 amps. I guess the difference between the bench supply readin and the clamp meter could be due to non-calibration.

Hi,

I added a 3rd 7135, but after that, it looks liike max current has gone down, to about 1.52 amps? I’ve checked the solder joints and all look ok. Is this a symptom of one or more of the 7135s going bad?

I have another same board, unmodified, and used the same setup, and shows 3 amps on high with the same XM-L2 emitter.

Maybe your battery can’t supply the needed current and it shift to a lower mode?
How about using a bench power supply? I got 5.04A just a while ago on my driver test. :beer:

Using my DIY bench power supply and DIY driver with 15x7135 I’m now getting 5.04A!
Placed the driver on a P60 pill and modified VBat- connection. I still need to improve my driver to emitter wirings similar to HKJ. :beer:

I am using a bench supply, rated at 5 amp/30V.

I had the same problem with my Nanjg drivers on my Convoy C8. The increase in tail cap current seem to be diminishing every time I add a new 7135. I wanted to ask the same question here in this thread but decided that I should seek the solution myself.

Yesterday, I made a new set of lead wires for my DMM, copying what many here in BLF did in getting the current readings of their flashlights. I used about a foot long (per color) of AWG 16 Automotive Wires, 2 Banana Plugs and 2 Alligator Clamps. When I used it to get the current readings, I was astonished with the results:

1) The LED current reading of my K40 (with 4 R100s added in parallel) became 5060 mA when the highest I could get using the stock lead wires of my DMM was only 3040 mA;
2) The LED current reading of my Small Sun T08 (direct-wired) became 5377 mA (yes, higher than my K40 which explains why it has a brighter beam) from the 2912 mA reading using stock lead wires;
3) The tail cap current reading of my Convoy C8 (with 12x 7135s) became 4169 mA from the 3012 mA reading using stock lead wires.

I guess changing the lead wires is worth a try. :slight_smile: