Yeah it’s good. I even took it all apart and used another led.
So a disconnect from the LED could cause the FET to go? That’s interesting stuff. I dunno what happened- I’ll replace it though and hopefully doesn’t happen again 
It doesn’t light up to the normal through the FET, the gate is just stuck
Thanks Wight… I’ll double check all my connects to the LED next time. I didn’t know a disconnect could cause something like that
I’m pretty sure it’s possible, but I wouldn’t really expect it to happen under normal circumstances either. More of a hypothetical problem. Long heavy gauge wires would probably make it more likely to happen, maybe especially if they were looped around a certain way (like an inductor).
If the FET is running cool then current flow isn’t bothering it. So while the FET is running cool the only potential problem is voltage spikes, and voltage spikes come from the inductive stuff going on in your wiring as the FET switches on and off.
I feel like you’ve probably hit the nail on the head when you say that you might have gotten a bum FET.
In bench testing I've done all kinds of unapproved stuff like that - holding the emitter leads to the driver pads instead of soldering, just to get a quick reading on something, fingers slip, wires come loose, etc. If there's a wrong way to do it, I've done it, but no dead FETs.
Fair enough. My own behavior along those lines is what I was thinking of when I typed #11. I haven’t done a fraction of the fooling around with FET drivers that you have, but I’m sure that if suddenly disconnecting the LED normally killed FETs we’d know by now.
So, that said… what’s your theory on what did kill the FET? Fellfromtree has done the requested checks as far as I can see and the FET sounds like it is in fact dead…
When there's no rational explanation, blame it on ESD. :D
Even though ESD happens with robust parts like this about as often as a rogue cosmic ray strike, you get taken seriously if you blame ESD.
I have a 12*7135 (I stacked 4 chips) on a Convoy S2+ build I did, it ran fine for a while but now it won’t come on but in moonlight (lowest) mode for some reason
I wonder if possibly the diode feeding voltage and for reverse polarity is bad…I will have to check
I had that happen and it turned out to be a flux corrosion issue on one of the inputs to the XM-L emitter. Same deal, just a low moonlight output. Driver was fine, and the emitter was fine once removed from the MCPCB.
It’s a 3x XP-G…ran great for about 5 min testing it out, shut it down…won’t come back on, I used a prototype ramping from turbo down to high (1 tick ever .25s) vs the large step down…I think it’s the firmware…but will re-flash just to make sure
But I will definitely check the inputs just to make sure…thanks!
Well, worth remembering that ESD adds up slowly over time, it’s not like being struck by lightning.
They used to call it “wounding” in the old days. A semiconductor has a narrow range within which it changes from being an insulator to a conductor — and ESD can change either side of that range by some small amount. Once enough components have wandered slightly off spec in various ways, the sum total performance of the circuit is, I believe the technical term used to be ‘cattywampus’
But maybe things like that don’t happen any more. I recall they never happened to young people, only to older people ….
Just fixed it reflowed with a heat gun and tested- looks good
Good to know hank, I havent wore a ground strap in years, maybe that would be a good idea :bigsmile:
You fixed it by reflowing the existing FET or by reflowing to swap on a new FET?
Doh you know I gotta start checking subs… It was actually quite a crazy ride on this one. I took off the old FET with a heatgun, put on the new FET. But then I was tired and hooked up the positive and negative backwards on the bench xd The diode was bypassed because I was testing what mods I could make to the board- so I had a wire going B+ to the cap, bypassing the diode.
I wanted to see if I needed the cap or not, as I don’t monitor that or write to memory on the Tiny13a- trying to optimize it for future builds xd If you will… Anyways The whole top end of the board started smoking. So I replaced the Tiny13a (checked out dead) and put in another FET as it wouldn’t fire up anymore. The stuff on the bottom all checked out. Then this is funny. I’m tired and solder fumes are giving me a headache. I clipped the led on, causing a short and shorting out the path. Turned it on to test- fried the diode (which I had just bypassed a minute ago).
Ok now I’m running out of parts. I put in a new diode and replaced the FET again just in case- it took a lot of juice through it with that short. Now my wife comes in and is like- what chu doing willis… Well Fing up
So I learned a lot that night, I really didn’t lose out. I learned what blows in reverse polarity. I learned what blows on a short. So now I have a deep understanding on what not to do and hopefully wearing a mask will let me undo the rosin flux damage to my brain
Hmm.
It’s all working now. Only fried the diode after I took out the bypass… Which is funny cause I had it bypassed a minute before clipping the leads to the led without a isolator put on the bottom of the noctigon xd
Yeah reverse fried the mcu with the bypass… Glad parts are cheap, and I learned a lot in the mess. Basically stupid stuff- I’m usually careful but get careless when I’m tired…
Thanks for the link on that cap wight :)… Was wondering if I could get rid of it. The code I used STAR Firmware by JonnyC - Source Code and Explanation
EDIT I see it the OTC would be under Star 3 OSH Park ~
Yeah I’m stubborn too when it comes to getting tired. Think my best ideas always come to me in my sleep. Like I used to wake up and write code that actually worked back in the day- same code I would be working on for two hours strait the night before with no results.
The diode, when I shorted out the led leads, forgot to put a piece of paper under the noctigon. Those MNKE batteries are supposed to do up to 32a- wish I didn’t throw away the bad FET from the reverse, it was melted on top lol
This is with an XM-L2 or something else? The XM-L2 thermal pad is isolated from both + and . In other words AFAIK you could attach either GND, BAT+, PWM signal, or whatever to the bottom of the Noctigon and not cause a problem. I do not understand what problem was caused by not using your paper isolator what do you have the thing resting on?
I clip them on with clips… It’s probably not a good idea now that I think about it. Put a piece of paper under the mcpcb so the bottom of the clips don’t connect. *EDIT oh I see what your saying- no it just makes a strait short between the two
I looked at the diagram and I couldn’t see how the diode went either… Think it would all go through the FET? All I know from experience is the diode goes if you short out LED+ and LED - on that driver.
Ahhhh, I understand now! I have my test MCPCB screwed to a very small heatsink, just solder wires onto it. Thanks for sticking with me until I understood.
Fair enough, it’s a mystery.
I guess… About half a volt makes it through the bad diode though. All I know is it’s stupid stuff… You shouldn’t be bypassing the diode, then on top of that apply power backwards… But somehow that happened. Then shorting the LED + and - when your tired shouldn’t happen either :bigsmile:
Good news is it’s done and the finished light worked great… Sent it out yesterday morning That’s the new U3 1A dedomed
