Came home last night to find it had turned itself on while sitting on a shelf. There’s no way it was on when I left as all other lights in the house were out making it really obvious. Just sitting on a shelf all by itself and turned on with no interaction from anything. The voltage was still 3.7 so it can’t have been on for too long and it was war but not uncomfortably hot so at least thermal management was doing its job.
Has this happened to anyone else? Anybody familiar with the firmware/hardware have an idea how it would be possible?
like to buy a second one, but i think the threads are low quality ,the tube must be a litlle bit thicker ( also better for heat dissipation) to make the threads deeper. Also the ring to put the lanyard i think is very thin (metal on metal) better is a nylon lanyard. And final i like to see a better holster.
Than for shure i buy a other one mmm maybe two :person_facepalming: , for the rest i like so much this light.
Someone can inform me witch throw farther the sst 20 or the XPL hi?
i have the one with aux leds, with NarsilM ( I think bought it in september) mine english is not so good to understand the manual.
I want to chance the moonlight (higher) someone can explane me in simple words to high up the moonlight?
If your in ramping, you go into the menu (this might require 8 seconds or 16 seconds), wait for 2 fast and 2 slow blinks, then click maybe 5 or 6 times. Make sure they are fast clicks and you see the confirmation blink with each button press. Then let it exit the menu on its own.
It’s weird that it mentions me, since Tom wrote about 99% of it.
Anyway, I am not aware of any bugs which would cause NarsilM to turn itself on, but I don’t really know enough about it to say a bug doesn’t exist. The closest thing I saw was a thermal management bug in a fork of Narsil, where if it was hot and then was turned off, with just the wrong timing, it could try to “step down” from off to whatever level would have been next in the step-down sequence. But usually it wouldn’t happen while off; it would happen next time the light was turned on, which could be an unpleasant surprise if it was at moon level the next time.
But I don’t know if that’s even possible in NarsilM 1.2. It could just be that something made it think the button was pressed.
If something made the switch close for a moment, then yes, it would turn on… it’s just odd for that to happen while it’s sitting unused on a shelf.
The time window for weird firmware behavior is only a few seconds. Certainly less than a minute. If it acted up more than a minute after being touched, it’s most likely a hardware issue, not firmware.
I had a similar experience a while ago with my first firmware for momentary switches. The light sometimes turned on when my hand came close. I suspected spikes caused by static electricity which triggered the interrupt on the port where the switch is connected to. So I added debouncing code after wakeup which actually cured this problem:
NarsilM has some de-bouncing. Instead of using the pin change interrupt, it polls the switch on each WDT tick. This avoids most of the noise. Then it requires a steady state for at least N ticks before considering the button state changed. However, it looks like the duration is only 1 tick for button press (activates on first cycle), and 4 ticks for button release, using only one measurement per tick… so it could possibly trigger if the timing lined up exactly wrong.
Anduril has a different style of de-bouncing, because two of its hardware targets have very noisy switches — FW3A and lightsabers. The lightsaber switches in particular bounce for a really long time. At first it used the pin change interrupt, but that proved too noisy even with pretty strong debouncing, so now it polls via WDT. But when it polls, it waits for 32 consecutive identical values measured about 0.063 ms apart (stable for 2 ms). This seems to work pretty well even on the noisiest switches I’ve tried.