Strange issue with new builds, looking for ideas

I put together 2 S2s a few weeks ago and ended up with an issue that has me puzzled.
All parts were purchased from Mountain Electronics:
Convoy S2 hosts
Triple Samsung LH351Ds on Copper MCPCB (one in 5000K, the other 4000K)
MTN-17DDm drivers with standard firmware w/memory (6 modes total), memory and 60 second Turbo Timer with wires installed including spring bypass
2 Sony/Panasonic Protected NCR18650GA batteries
And of course spacers and optics
I also had switches with bypassed springs on hand.

I assembled the lights with artic silver compound, the only soldering I had to do was wires from the drivers onto the MCPCB.
After fully charging the batteries I tested the lights, initially everything looked fine, all modes correct from moonlight to 100%.
However I found one of the lights, when on 100% would turn off after 2 or 3 minutes (not step down, fully off). If switched off/on it would come back on at 100% but turn off again, more quickly the second time. If allowed to fully cool the cycle would repeat.
All other modes would remain on well past 5 minutes, issue only occurred on 100%.

Swapping batteries between lights then tail-caps made no difference, the issue remained with the light.

Thinking I had a bad solder at the MCPCB I disassembled and re-soldered both, no change.

My next thought was that the incident light may be drawing a bit more current tripping the protection circuit on the battery (rated 10A).

I put a fully charged Samsung 25R in and ran through all the modes leaving on high for about 5 minutes with no problem. It did appear to be noticeably brighter with the 25R but still more than bright enough with the protected cell.

I figured I had found the issue and was happy to live with it since I seldom use this type of light on 100%.

Out of curiosity I installed the 25R in the other light and ran through the modes, again admiring how bright it was with that battery.

This is where the real problem starts.
I reinstalled the protected cell in the light and right away noticed I lost moon mode, in fact there is little if any noticeable difference in the first 3 modes. Modes 4, 5 and 6 get progressively brighter.
Putting the other protected cell in the other light I get the same results.

Back to the 25R again and same results, 3 visibly identical modes followed by 3 progressively brighter modes.

I set the lights aside for a couple of weeks and took a look at them today to confirm what I’m seeing.
Measuring current draw through the modes with fully charged protected cells (Fluke meter with short, heavy leads) I get roughly the same results with both lights.
Mode 1 (moonlight) 1.2A
Mode 2 1.2A
Mode 3 1.2A
Mode 4 2.4A
Mode 5 3.5A
Mode 6 6.0A

All 3 LEDs light and appear to have equal output.

I’m at a loss as to what happened and it occurred on two lights.
I have at least 5 other similar S2 using the same driver but with Cree emitters. In fact the 25R I used came out of one I use often.

It seems like it must be a driver issue but I don’t understand what’s going on.
I have a few Q-Lights with moon mode, I may just install those and be done with it.

Anyone have any ideas? Sorry for the long post.

2 Thanks

Maybe thefreeman or TK or Dale or someone will happen by with better specific input, but sounds like bum chips on the drivers. If they were new, it’s possible that he’s gotten some bad ones with all the recent supply turmoil. I haven’t bought any drivers from him since before the pandemic I think, but I’ve read that there have been more issues than usual…and there are still big universal issues with the 7135 chips as well but I don’t know if a bad one (the only one, here) would cause this kind of thing. Shoot him a message and let him know what your meter readings said…could be helpful to him. Nothing special or different about the emitters that would cause anything. Glad you put a vanilla cell in there because that was the first thought of course. :slight_smile:

As an aside, if it’s been awhile for you then you may not know that Simon changed the threading on the S2/+ models some time back (traditional trapezoidal to a squared off type), so parts may not swap between hosts depending on what you have in hand.

You might have a look at the FET beside the 7135, I found the main pad under the FET had tiny nibs that stick out slightly and the leg of the 7135 is SO close, if touching it messes up modes. Give that a close look and see that it’s clear.

This is a different, larger driver but check this area.

1 Thank

I don’t have a picture from the smaller driver where I ran across this… perhaps if you could take a pic of the driver so we could look it over?

Did you, by chance, choose the thinner board? With bypassed springs the protected cell being longer may well be flexing the board and causing a glitch.

If all else fails, remove the driver, unsolder wires and springs, put it on a frying pan and reflow it do not put flame on high, keep it on low and wait, once solder melts give it few seconds and turn off the heat. sounds like you maybe losing a contact due to bad solder on one of the components, if you have a soldering paste put tiny dots of it on all solder joints. If that does not work, trash the driver. broken traces on a mcpcb are not easily fixed cuz they are next to impossible to find.
DB has a point, if the board is thin and flexing, broken traces/contacts are very likely.

Thanks for all the suggestions.

There was no option on the board when I ordered the drivers last month would think by now they would all be the 1.6mm, I say that only because the note has been posted on the site for a while now.

Also don’t think the protected cells caused the issue since it started after using the unprotected 25R on both lights and is present regardless of which cell I use now.

At any rate I’ll try to find the time this weekend to take one of the lights apart for driver inspection. While out I’ll confirm which PCB it is and take some photos to post. I’ll also measure the batteries for the record.

If all else fails I’ll email Richard to see if he’s had any similar reports.

1 Thank

Hopefully it’s sorted soon. If you do find out a definitive answer about the failure mode, please do share that. I’m always curious about that as I try to learn more about driver circuits but it might be handy someday for some other person troubleshooting, too.

I pulled the driver out of one of the lights tonight and inspected it with a jeweler’s loupe. Everything looks good as far as I can tell, no sign of a cracked board, solders or unwanted solder bridges and traces all look intact.
It is the 1.6 mm PCB.

Using a single XML2 mounted on a CPU heatsink I did a quick test using 3 NiMH AAs in series. With that set up all driver modes are visually the same, none distinctly dimmer than the others. I used to use this set up when stacking 7135s on regulated drivers and could always see the steps in the lower modes. I’ll try again when I have time with an 18650 laptop pull, not looking for high output just to see if there’s a change between modes. When I do I’ll document the draw on each mode.

I didn’t measure but side by side the protected battery is longer than the 25R as expected.

I’m going to try to post some photo, haven’t done so on the new site yet.



That was easy!

Very nice pictures! I don’t see anything right off the top, as expected because Richard uses a very nice computer controlled re-flow oven. But having had the same kind of issues with one of his drivers and about to use 4 very much like yours I am invested in finding out what is going on here!

On it…

Three NiMH cells (1.2V) should technically be 3.6V which is pretty low to run one of these drivers, sag under initial start may well have it below operational threshold.

I soldered an Omten to the neg end of an old Sanyo FJ years ago with 5” 18ga leads off the switch and positive end of cell for testing purposes…switch is taped to the body of the cell. Works great, the switch makes testing go smoothly.

@DB_Custom Yes, I realize the limitations on using three NiMH cells, and on top of that they’re installed in a cheap holder with very thin wires. What I do expect to see is a change in current/output on the very lowest modes.

With the incident driver I get approximately 0.65A (inductive meter on the + wire between driver and LED) on all modes and no noticeable difference in light output. No doubt the maximum the cells can put out.

After charging the cells I did the same test using a standard Q-Lite REV.A 3.04A driver, no stars soldered so 3 mode. Results were 0.029A, 0.225A and 0.793A with noticeable change in output.

It appears the low modes were lost after using the 25R with the incident drivers.

I don’t know exactly how the 7135 functions in this driver. I though it was just for the moonlight mode but Richard’s site says “The 7135 gives an ultra-efficient moonlight and low mode…”. Looking at the readings I have from when the light was assembled I think it’s just the first two modes that are lost.

Maybe something shorted in the 7135?

Quite likely. They’re a bit fragile and tend to fail closed when pushed too hard. (closed as in a closed circuit, letting current through) And from what I’ve heard, they’ve gotten more fragile over time, easier to burn out, as compared to older production batches.

The usual symptom of a dead 7135 is that the low modes are all too high, and identical… and then the higher modes (from a DD FET or similar) work as intended.

However, usually the low modes on a broken 1x7135 light are all 350mA or 380mA… not 1.2A. So something doesn’t add up. Maybe it just failed harder than the others I’ve seen, or maybe there’s something else going on.

Hardware isn’t really my thing though, so take all this with a grain of salt. There’s a pretty good chance I’m wrong.

2 Thanks

I just put one of the four 22mm Bistro drivers in an X6. Pretty much identical driver but with a larger border.

And again, oddities. Everything tested fine on the bench, modes looked good, no problem. But once assembled… one mode. Took it apart, double checked everything, tested again on bench, one mode. Looked over all the contacts, nothing out of place. Tested again this time to a loose XPL-HI on a 20mm star, no issues.

Reassembled, checked with the head still bare… one mode. Took it BACK apart, ran a pick around edges of contacts looking for a ground short. Zip. Tested it on the HI… worked fine. Reassembled it into the X6, to a 4750K SST40, old stock but otherwise brand new fresh out of the tape just mounted on a new previously unused 20mm SinkPAD. One mode, bumped it through like 5 times with only about 2 seconds on at each stop, suddenly bumped to moon and is now functioning through 7 modes as expected.

Reversing is about a 3 second press and batt ck blinks
8 straight blinks on my Panasonic BD cell, but otherwise it happily puts out .9 to 1410 lumens.

Beats me.

Seems as though it had a weak short to ground bypassing the driver but it never showed up anywhere and apparently burned out under load, eradicating the issue. That’s the best guess I can venture.

1 Thank

Ironically, as it went together so smoothly I was thinking about a thousand builds, the times it took a few minutes and was flawless vs the times it was a PITA. And it occurred to me that for a surgeon doing hip replacements it is likely to be similar… easy peasy then arse busting ridiculous.

Then my light didn’t work. Lol. Took a little longer, had to jump through extra hoops, but in the end it’s a nice light!

Some days you get the bear, some days the bear gets you.

I hope there is no arse busting in your future. Just easy peasy all the way. :mending_heart:

Thank you. I guess I’m getting all my bad luck out of the way on flashlight builds. Lol

Strange, so the Bistro driver appears to be working fine now?

I did a little testing with the second light (same issue but the light is still together) using a protected 18650. It’s a little tough to tell what modes I’m in on the low modes since I’m testing with the tail cap off holding one meter lead on the battery and the other on the light body. Unintended double taps occur easily.

Modes 1 and 2 varied from 0.27 to about 1.0 amp and were inconsistent. I always knew when I reached mode 3, 1.2A, mode 4 2.7A, mode 5 3.5A and mode 6 6.2A. Readings are all approximate with slight variation but the top 4 modes were consistent.

An interesting new issue came up. I was holding the light on for longer periods, maybe 10 seconds or so in each mode to get better readings and ran through the modes quite a few times, long enough for the light to start getting hot.
After testing I put the tail cap back on and found I had modes where the light would flash on then immediately back off. Turn the switch off then back on and the issue would repeat. Double tap a few times to get to the higher modes and the light remained on.
I let the light cool off and it’s back to having the not so low modes.

I have about a half dozen of the old 380mA 7135s still in the strip in my stash of parts but I’m not sure I could successfully swap them out on the drivers. I could probably put a new one on a clean board but I’d probably damage something trying to remove the old one.

I’ll probably email Richard tonight to see what he thinks. If all else fails I’ll try swapping the 7135, and if that doesn’t work I’ll be installing Q-Lites.

Practically speaking, 3A is more that I need out of a light this size but that’s no fun.