One of my favorite EDC lights is a Convoy S2 I have modified to include triple Nichia 219 LEDs on a Noctogon MCPCB, including copper heat sink, Carlco optics and the QLITE driver with 12*7135s (can provide about 4 to 4.5 Amps). The driver software is flashed with custom modes (thanks to RMM’s store!)
Seems that the higher current has caused the QLITE’s copper pad at the base of the spring to weaken over time. This is the copper pad to which the spring for the positive battery terminal is soldered. This copper pad has now detached from the QLITE board surface (and of course has taken the spring along with it).
After examining the remaining surface on the QLITE board, I do not think this is readily fixable. It seems the simplest approach is to buy a replacement QLITE driver. There just does not seem to be enough space and contact vias to attach a brass button or replacement spring and then make a reliable electrical connection from the battery’s positive terminal into the rest of the QLITE electronics.
Unless, that is, some of our talented community has come up with a creative to save such a driver when this issue has surfaced?
As noted, I think the simplest thing is to replace the QLITE driver, but thought I would first ask.
In addition, has anyone found a way to bypass or strengthen the battery spring / contact pad on a QLITE?
I'll have to say that I've never had one rip off yet--even when I was running them at 6.5A+ as an FET driver. Definitely not a common occurrence.
Since the pad sits directly underneath the MCU, it is a bit tricky to make a "clean" fix. If I were to do it, I would find a clean spot on the PCB (with no traces), drill through it, then run a positive wire up to the top. The spring would then be epoxied in place with the wire soldered directly to it. I did this to a few of the qlite drivers with FETs on them as a further current bypass over a year ago, so you should be able to make it work.
I had the same thing happen on a MT-G2 Convoy L5 build. It wasn’t amps or heat, the battery tube was short for 2S 26650 batteries, causing both springs to be coil bound and bottom out, twisting the stout driver spring, till it cracked the pcb. So I replaced it with a single sided driver and a copper spacer in between the battery tube and brass pill. The Qlite was cleared of it’s components, and the chips were added to another driver for a total of 20 chips in a Y3 and still performing today. It sucks but it also gave me the opportunity to learn how to stack chips.
Richard - I think I understand what you’re describing as a fix. I may try to epoxy a brass post instead of spring to affect the repair.
KawiBoy - in looking closer at the pad/spring piece that has ripped off, I think you may be right in that the source of the problem was something mechanical such as you experienced rather than too much current.
The spring appears to be compressed at a weird angle. My QLITE driver may have suffered from the heat sink and pill not being screwed far enough forward into the S2’s head.
During S2 assembly, I put the S2 ’s glass Lens in front of the carclo optics and also used the O-ring in front the lens. That combo may have shortened the effective battery tube length by a couple of millimeters, putting undue stress on the QLITE driver spring.
I hope I can fit an epoxied brass post in place of the spring with a soldered wire going through a new hole that I will drill into the QLITE driver board as Richard described.
You may be able to just reflow a thin copper disk in the pad's place. There are like 3 via's that connect the pad to the MCU side of the board. They should fill with solder and hold the disk. You could then clean up the board and use some kapton tape or something to reinforce it.
Hmmm, thanks for this idea as well. The original spring is still soldered to the pad. Perhaps some solder paste to get solder to flow between the pad and through the vias, supplemented with epoxy on the remainder of the pad for mechanical bond to the driver board. I wonder if cured epoxy will stand up to the reflow heat?
At this point, it is more a challenge to see whether I can “out-skill” the events that caused the failure.
