What did you mod today?

I don’t know anything at all about this kind of driver, but it seems like the MCU is getting hot or one of the Capacitors is. The emitters are working, the chips are working, the problem only starts when it heats up, so there’s something in the voltage to the MCU or the heat, which might be coming from the chips getting hot reducing 4.2V to 350mA. Sink the chips, see if that does it…

I don’t know exactly Ouchy, I have done several mods with this driver, even experimented succesfully with different resistor values and FET’s, but they kept working fine. Two observations that do give a clue perhaps what is happening, but may not get you to a solution:
*the driver is very sensitive to a lighted tailcap I found: I even tried without bleeder and with several bleeder resistor values, it all messed up the UI. So what is happening in the tail easily influences the driver timing. Did you do a spring by-pass, then try the flashlight without it, it may just improve. Try a lower drain battery, or higher.
*when the light is off, it resets to ‘start with white’ . In this driver this reset is fast, like 0.5 seconds. It sounds like perhaps that has become shorter than 0.5 seconds in your case. Try very short taps to change modes to try this idea out.

I though about the tailcap, and swapped out the back end from different S2s. I don’t think it’s the driver. There’s something else going on here.

I reheated a wiped the board so many times trying to reflow, that perhaps I ruined the integrity of one or more pads and the solder adhesion is iffy. Also. I ground down this board and drilled holes through the traces. Something could be moving once it gets hot causing some kind of short. Although the board isn’t DTP, it’s pretty nice. The entire surface is copper that is divided into eighths from the center out.
Wires? Who knows. If gone around and resoldered everything until all the cheap casings are melting into blobs. Still can’t be sure about the integrity of the pads.
Who knows, maybe the the AA isn’t cured completely and is causing a short. I removed the centering ring because it was sliding around and not curing to the mask.

I tried it a few minutes ago and it went through 8-10 cycles before it started glitching. Maybe it’s getting better. What if I wake up tomorrow to find it was all a dream and it works fine.

If it doesn’t get fixed I won’t have much choice but to scrape the whole board off and replace it with a standard RGBW emitter and board.

Did I mention that the optic is really good for these emitters?

Here’s what it looks like when it’s working.

Nice mod Ouchy. I like how you used the red, green and blue body pieces to go with your RGB emitter. :smiley:

Where’d you get the optic?

I had a triple with individually addressable leds (same RGBW driver as Ouchyfoot used above but modified, and not using RGB leds, link) with one of the leds a Pro-Light PC-amber one at ~5.5 lumen. On holidays when camping, or otherwise away from home with the family, I have used it as a night light for my son because he is afraid in total dark, it runs for 11 hours per night. But I like using it as my EDC as well, so it bothered me that I could not use it when he is asleep. So tonight I made a new flashlight with this led especially as his night light. The requirements: a ~5 lumen setting, safe, reliable, good runtime. I used:

*a Convoy S3 host I had around for ages and never felt much love for,

*a ProLight PK2N PC-amber led,

*a BLF-A6 driver (the led has such a high voltage that direct drive on a 18650 it still draws well under 2A),

*and did 2 spring bypasses to at least get some light out of it on high setting (for my own pleasure, the night light mode is only 30mA).

Because of the high voltage of the led, mode 4,5,6 and seven have close to the same output, but mode 2 is a nice 3 lumen setting for night light use (mode 3 is 30 lumen, a bit too much). I run it on a protected NCR18650B for good runtime, the 30mA/3lumen night light mode runs for 110 hours, with the light being on all night for eleven hours, this means 10 nights on a battery. The highest setting is 1.13A/190lumen.

edit: I pimped the light a bit with a red tail cap and a red o-ring, and I must admit that despite it being a Convoy S3 it looks alright now:

Thanks, I think. It would be nicer if it worked.

I got the optics from kiriba-ru, but only because I was ordering heatsinks from him anyway. You can get them in most of the usual places, mouser, Luxeon star, etc. they’re pretty common. I think they might also be nice with hi CRI nichias.

This morning I ripped apart my RGBW that I built yesterday ( Post #1150) to find out why the light keeps losing modes after it’s been running.

Today I removed the driver and unsolder end all the wires. Went over the edges of the mcpcb with an emory board where I had filed it down and drilled holes through the traces to make sure there weren’t any minute bits that were shorting as the board heated up and expanded. I cleaned up and tested all the pads and rewired everything with 30AWG Teflon wire.

At first it seems fine. I can run through all the modes a few dozen times until it warms up and starts skipping modes. This driver is extremely sensitive, barely touching the switch to change modes. I’m talking barely making contact.

At first it starts skipping the last modes, police strobes etc. then it loses blue, then green until all that’s left is red and white. Finally it loses red. If I let it sit for a minute, all modes are functioning again normally.

It’s not the board.
It’s not the wiring. This is proven by my DMM and the fact that everything does work.
It’s not poorly connected emitter pads that loose connection as the board warms up, as I can turn it onto blue, one of the first modes that goes, and leave it on for twenty minutes. After leaving on this mode for some time and try to cycle, I only get white again.

Therefor, it must be the driver.
This driver was designed for multi die RGBW MC-E emitters. I have one exactly like it in another light driving an MC-E with no problems, so now I don’t know if the driver is defective, or if there is something about the voltage requirements of these Luxeon Z emitters that is confusing the hell out of it, causing it to warm up and start losing modes one at a time.
That’s about as far as my knowledge will take me on this issue…

How’s the switch?

(just beat me to it)

Switch?

Switch is good. It’s an S2+, so I’ve even swapped the rear ends of three others onto it to test. Same result. This driver is ultra sensitive to a barely touched press. Forget about soft press, this driver likes to change modes by the least amount of touch the switch allows.

I had something similar happen that drove me nuts, eventually i did pin it down to the switch, the switch had worn just enough that the pressure of the rubber boot was ever so just enough to press in the button to cause most of what your describing.

i added another plastic spacer/washer giving the button more breathing room and everything worked perfectly since then.

I tried four switches, all brand new. One S2 metal boot, two S2 rubber boot, and one non convoy fwd. clicky. I’m not worried about the sensitivity, I’m more concerned about why it keeps losing modes one at a time until there is just white left.
I’m really starting to think that there is something about those Luxeon Z emitters that this driver reacts to.

Can you reflow the Luxeon Z off and swap in a color MCE? If you have one and if they are the same footprint? If it works right with MCE then the problem would be confirmed to be the Luxeon emitter.

If the footprint is different, you could try an MCE on an MCE star. If it worked that would confirm that the problem is with the Luxeon emitter or star.

No. Can’t just swap emitters. The luxeons are four small separate emitters, each with their own footprint. I do have an extra MCE mounted on its own board, but I have to await a renewed driving force before I can tear this one apart again.

A dirt cheap SkyWolfEye (don’t bother)
1-mode 1xAA-2xAA driver and an old amber XPE emitter
for use with a “4/3AF” NiMH cell, to give to a friend who already has several such for nighttime winter use, way off-grid.

Old school, an SP-98 Sipik clone with a swirly multi tone blue/lightblue/white pattern either painted on it or wrapped. My friend wanted more out of it, so I stuffed a 1/2” thick copper plug in the hollow pill, glued it in with Arctic Alumina Thermal Adhesive, used a 20 to 17 brass adapter ring to mount a 17mm FET+1 driver with A6 firmware. Swapped in an XP-L HI emitter on a 20mm Noctigon.

Pretty bright! The square image shows the oddly patterned XP-L HI die quite well, not sure if I should have done something different there, might pull it and put a domed XM-L2 U4 1A in it’s place.

Ouchyfoot, it may be that the rubber cylinder in the button cover is preloading the switch giving it a hair trigger. Swapping out switch modules won’t change anything but shaving the cylinder .5-1mm can easily fix it. Or a thicker aluminum disc that stacks on the switch.

A Nitefighter F41C :

It came in a three tone brown/blue anodizing. I like the 26650 format with a side switch, but this little “ugly duckling” was hidden away.

After almost a year in exile I took it out and started the transformation :


Bead blasted.

The internals needed some work too :


XHP-50 5B

The driver is a LD-2 and the power comes from two 26350 batteries.

Nitecore EC4S XHP-50 3B (left) Nitefighter F41 XHP-50 5B (right)


Nitefighter F41 XHP-50 5B

Eagle Eye X6 with the same driver/LED.
I am sorry I didn’t take any photos of the innards, we all need a good laugh every now and again :wink: , it ain’t pretty but it works. I will have to open it up again in the near future to fix the e-switch the only glue I had was hotglue, not the best for a 6A XHP-50.

Sweet. That bead blasting does it for me. :slight_smile: