I'm trying to use these to drive 10W LED panels. Initial test went swimmingly, But now, something has changed. And I don't know what.
The LED panels are the same. Polarity has been checked over 5 times on both LED panels and the controllers. from the PWM and Gnd pins, I have two wires running out. One is spliced into the main ground (it's tied together internally anyway, and the way that dimmer works is that it has to be high to enable the light. So I simply apply 5v (sharing common ground, etc - after smoking the last driver i thought I'd tested everything and was sure there wouldn't be a repeat) to the PWM pin for 100% duty cycle. Seems reasonable enough.
Apparently the silicon disagrees. Sparks and a cloud of smoke ensue as the LED driver destroys itself.
Hm, and now leaving that pwm header disconnected entirely also instatrashes the drivers, LED panels are unharmed. Off to aliexpress to order another 10 drivers
That;s what I thought at first too, but I can't reconcile that with the measured 0v difference between the two grounds, and the fact that no smoke is emitted until the driver board is connected.
And it happens with just the dead simplest configuration. So that's 100% ruled out. Single supply, just the LED panel which last week I was adjusting the brightness of up and down no problem.
I bought a couple of those last year, haven't got around to trying them out.
I don't know what DOC aliexpress provides, but you can take a look at what this amazon seller made available --
Nice schematic but it should be top down left right. Plus on top, ground on the bottom. Input on left output on the right. It is a bit confusing when first looking at it. Post a link to the driver you are using that gives technical details.
On your schematic show all power sources, grounds, etc.
It says that in a few listings, but the majority of listings say it's okay for different voltage panels, listing 10v, 12v, 14v, etc. (I think the input = output is the limit, the highest voltage it can do because of the buck topology). Anyway, I had one of these damned things working beautifully last week.
And I did just 5 minutes ago, then I disconnected it to put the scope on the input, so I can see if there's a spike or something from an SMPS that might be causing an issue (I've had that blow out devices before)
Bad news ... unfortunately, its much worse than it looks!
Some non-isolated step down converters come with diagrams that use ground symbols like its candy. Lets take a look at U1 PT4115 shown in the above circuit ...
Here's another similar component found on DigiKey. Note the GND pin label on the IC, then look at their application circuit which is labelled correctly. The signal is actually high voltage DC - (non isolated)
This buck converter is not supposed to be isolated. The 20v DC power supply is! as is the 5v power supply. They're SMPS supplies and I can tie ground together and not get smoke or flame.
Its this connection that has my curiosity. I would check this with a multimeter where the multimeter common lead is connected to earth ground. What is the AC and DC readings?
Could try a suitable series resistor from this connection to earth ground to see if the above readings were static or solid.
Is the scope probe ground connected to earth ground? (not a problem if you're using a battery powered digital scope).
Just for testing (and to see if it resolves the issue), it might be a good idea to have a 120:120 isolation transformer powering everything.
