Qlite driver problem

I seem to be having problems with the Qlite (supposedly Rev.A) driver. With one XML2, low mode flickers as if the battery is making poor contact and firefly won’t light at all. Once, when the battery died on high, it did go to firefly mode but never again. My other XML2 (from vestureofblood) works fine. On my XML, low and firefly won’t light at all. To narrow down the problem, testing was done with 18 gauge wire soldered to both ends of the battery, negative to board and the loose positive used as a switch. None were left on for more than second so heat can be ruled out also. High and medium modes work fine. Anyone have similar results or care to shed some light on the matter? Is the pulse width (on time duration) too short? LED out of tolerance?

Without the lead connected from the battery positive to the driver positive, have you measured the battery output? Another user had a problem with tailcap current recently that turned out that the battery was drained. Just a sanity check…

I have experienced the same thing with my Qlite + dedomed XM-L2 also, sometimes in moonlight mode the LED lights up but sometimes it doesn’t. I think it is something to do with some variance in LED.

I have to assume one working LED is enough proof that the driver is fine. I’ll try bypassing the star. If that works, I’ll know where the problem is. I’m just glad that the XML2 on copper is the one that works :slight_smile: .

UPDATE: Soldered directly to emitter didn’t help.

I've had a few where the LED refused to light up on moonlight mode, it was something wrong with the emitter and swapping it fixed the issue. My only guess was that they got too hot during the reflow, no way to confirm that though.

The Qlite driver moonlight mode is too low to work with many LEDs. Also, they may work for a while, but after the ’7135 chips have been run on high for a while, their response shifts and will no longer work with the very narrow PWM pulses used to get those moonlight modes.

I ran into the issue with my custom SRK driver. I had it set up to do a low mode with 1 lumen per LED, but after running the thing for a few hours in the higher modes, the lowest output that I could reliably get was around 15 lumens per LED. I thought that the LEDs were the cause, but it turned out to be the ’7135 chips. The response shift in the ’7135s seems to be permanent… they don’t recover to their original response profile.

aye carumba, was this a problem for all ’7135s or just a driver here and there?

I think it is probably all of them… it looks like it is just what AMC7135s do. They may not all do it to the same degree, but there are 8 chips in parallel on each LED and each LED channel was acting the same. Swapping the processor or LEDs did not change anything… that just left the AMC7135 chips.

After running the light on high (70+ degrees C) I notice that I no longer had my low low mode. Tweaked the code to boost the low level, ran it on high again, same thing. It finally stabilized with a reliable low PWM level of 8/256. It started out at 3/256.

i was thinking of ordering some of those qlite drivers from intl outdoor with 20mA moonlight, but this is quite the kick in the pants, i have no reprogramming ability
it seems R&D testing consists of trying the levels for a one second each on the prototype then mass production

None of my XML2/T6/3C leds work with any of my Qlite 1A’s on moonlight mode…

They screwed up with the 1A. On the rev 1 Qlites the moonlight and low modes were too close together. Instead of increasing the low mode in the 1A, they decreased the moonlight mode to the point that it is not stable.

if i use 2, 25 and 100%, are those modes stable? (star 1)

I got several of them and all work fine. The first release (which I have) actually had some non-380mA reaching batch, I actually swapped chips with others to test that.

Yes, that’s what I run mine at. Also, star 1 does not need to be connected… it is already ground.

Further testing using direct drive shows that the LED that won’t work on low and firefly gives up at about 3v and a working LED stays lit somewhere below about 2.5v. I wish I could at least get the 2% mode.

I read something about ripple currents causing problems. I don’t understand it fully but it sounds like it might have something to with what’s happening. It’s on the last few pages:


I have installed a Qlite rev.A driver with 13*7135 into my SS ZY-T08, and when the light is turned on on High for says 30 seconds the moonlight mode will not work after that. However after turning off the light for more than 3 minutes the moonlight mode works again, and it is always like this. I wonder is that something to do with heat.

I wish I would of read this thread before I built my last batch of P60s. Bought 3 Qlite Rev A from Hank along with one of each of the new xml2 hes offering on the noctigons. After I put them all together the 1A tint Xml2 wouldn’t light up on moonlight mode So I decided to tear it apart and inspect everything. Still wouldn’t light up on moonlight So i decided to just go back to the default mode and used solder braid to remove the bridge to the last star. Now when I turn it on after 5sec or so it starts switching between low and medium like warning mode. IT does this no matter what mode its on and is very frustrating. Battery measures 4.01 volts and doesnt do this with the other drop ins.

Any ideas?

Sounds like a different problem. Did you accidentally compromise the ground bridge to driver with the braid? Solder joints aren’t always as good as they look. I test mine by pulling on them. A tight fitting driver can’t be tested that way so I make sure they drop in and fall out easily before soldering.

The ones I had the moonlight issues with were boards from several different sources (IlluminationSupply standard 105C, Fasttech 2-group 3/5 mode, & pre-rev.A IOS Qlites) but all had been flashed with Nlite code. Anyone know what PWM setting DrJones used in that one for the moonlight mode?

And in my case, changing the drivers didn't fix it, I had to swap LEDs. The drivers would run other LEDs at the moonlight mode, and new drivers wouldn't run the problem LEDs at moonlight.

Random thought, has anyone tried adding some more capacitance? Say an extra 10-22uf ceramic, you could even try it on the spring side. I caught myself out once, with some nasty blinkies when I only had a 100nf cap, instead of the 10uf+ that I think is required. Moonlight mode may even require more.

The cap on the Nanjg drivers is a 10uF ceramic cap.