TPS92200 Led driver failed

[Muffins]

Hi guys,

I'm busy making a circuit with some high power leds powered by a 20v drill battery. I'm using a TPS92200, I got a little PCB made up, attached power to it and it blew up with a decent sized jet of flame.

I did apply power to it without connecting the leds so I thought that might have been a problem but in the datasheet it says it has 'no load' protection. Maybe I just need a second set of eyes on this, any ideas on why this failed? I've attached some images of the schematic and PCB layout.

Also, it seems to have failed around the ground pin. Could this be an inrush current problem? Power was applied via a slide switch.


Looking at the datasheet, it seems I didn't put a the feedback resistor in the circuit as I was following a diagram without one in it. But since there was no load connected it would have been pulled to ground.

Update:
I tried a second PCB since I had two assembled. This time I had the LEDs attached and it didn't explode but the leds didn't turn on either. Then I connected it to some leds and resistors on a breadboard and they came on but the dimmer did not do anything when the pot was turned. I'm guessing this is has something to do with the driver not being able to achieve it's set current value due to the high resistance. Then after turning it off and on a couple times the driver popped again. Considering using a completely different device now.

[jmelson]

The oddity is this IC has a low-side transistor.  I don't see the need for that in a buck LED driver.  If the low-side transistor and high-side one ever turn on at the same time, then an explosion is guaranteed.  I have used the LM3405 with good results.
Jon

[Muffins]

Think I might want a bit more current then that. Looking at one of these right now https://www.diodes.com/part/view/AL8843#tab-overview

Am I right in thinking that if the supply voltage is 20v, and that is within the output range of the IC, if 3 LEDs with forward voltage drops of 6v are connected on the output it should be fine? The total voltage drop of the LEDs just needs to be less than the maximum output voltage of the chip?