My light is doing something odd, maybe someone has seen this before. I have a triple 219C with the above driver that drops out of high very quickly as if heat throttling. I have reset the driver, checked to make sure everything is tight, tried different 30q’s, Panasonic 18650b’s and Keeppower IMR18650 20amp batteries. With the light in thermal calibration mode it does not drop out of high, only in normal mode. This is a titanium host so I have the thermals set for about 20 seconds for a drop down.
I have had trouble a couple times with MTN fet drivers doing things like that—Richard explained that with the high current ,the ripple cause the driver to step down and sometimes change the mode group —usually you can use a higher value capacitor— don’t remember where he said he put it—I just get him to fix them—I’m older and these small parts ruin my light building experience—usually when the battery gets below aq certain voltage the problem stops— I have one that PF cells function perfectly in
HMM,
I have the same and same under 3* XP-L HI 1A running 30Q
Bistro is thermal sensor, not timed. TI sucks at heat absorption.
Have you tried to reset the UI? Think its 8 blinks or the last option.
Also the switch install is somewhat difficult to get working good in these. Just a thought.
Space is also minimal for battery. Another thought in case spring are binding, bending, grounding out, or POS wire is not pushing back through the hole.
Probably not relevant but some obstacles I had in the build.
Yesterday I made another triple 219C light with this fix, and when trying the ‘fixed’ (Banggood X6/X5) driver with a single XM-L2, the UI timings were nowhere, it was impossible to operate.
But when this same fixed driver was installed in the 219C triple, the UI was back to normal, with the added advantage that it allowed a fairly well lighted tailcap, which is something that I never got to work with low voltage led triples in combination with the X5/X6 driver before.
So the fix for me only works in these triples, but that was where the problems were in the first place
With the tail-leds really bright (did not write down the actual resistor values) the detection of ‘off’ went to infinity, so when the driver was switched on after having been off for a while, the time off was still interpreted as ‘go backwards’. Increasing the tail resistance somewhat (an extra 6.8kOhm) solved this.
djozz, thanks for the update. I am not using the exact same driver, but I bet the fix will work when I get time to test it. This is going in a Rey Triple Ti, so no lighted tailcap… but trits are installed! I have never had this happen on a single led or even in an XP-G2 4C triple. It took me by surprise.
This will give me a chance to removed the frosted narrow spot and install the clear narrow spot anyway.
djozz, that is not normal. The small value added resistor cannot affect otc or other timing in any significant way.
Two things to check:
value of the added resistor (we need only a few ohms, not kohms)
the effective circuit after the mod, can you maybe sketch it out?
FWIW I am in the process of slowly replacing all my FET+1 drivers with new oshpark boards that have pads for the resistor fix and improved pcb layout to help with mcu stability. These are both singles and triples used with both hi capacity and hi current cells. The board is shared, if anyone is interested. Just note there are no explicit pads for an otc (I have no use for it).
The other thing that will have an effect on this is the addition of the lighted tailcap without a good bleeder. I have found that the bleeder is fairly critical and sometimes the OTC needs to be adjusted with these boards. I have built 12 or 13 of them with 2 to 6 leds on the boards. Timing can sometimes be crazy.
So, looking at my driver… I am not sure where the resistor would go. I think it would land on the pad of the resistor that is end to end with the bigger diode. Gonna be hard to pull off on this driver if I am right. Or, I will move my H17f from my single to this light and put this driver in my single.
For power clicky lights I use the noinit_decay method to distinguish between long (reset to start mode) and short (next mode) clicks. E-switch lights do not need an OTC of course. I can link to something in pastebin if you want to try it. None of my firmwares have memory or user-adjustable settings, if that is important to you.
I did patch one of the red nuv drivers, not using the Zener clamp. I had the small R in series with D1, with each installed on one of the D1 pads and doing a pyramid with the junction up in the air.
