I’m trying to figure out how this could have caused the damage it did, but I think I plugged my USB to Serial adapter in backwards while it was connected to my PC:
Adapter-Target
Gnd-TX
TX-Nothing
Vbus-RX
RX-Gnd
I can only assume that’s what I did, because when I saw smoke I yanked it out again, and as it was still producing smoke I pulled the USB plug as well. After about a minute of looking for shorts and checking the components I decided there was nothing else I could do about it, so I plugged it back in and got to watch a trace burn out like a light bulb filament:
Yes, this was the board I reconfigured to request 500mA from the USB bus. Out of morbid curiosity I tried connecting the two through holes with tweezers, and to my surprise the red light came on! I soldered a small wire in place of the trace, plugged it back in, and got no LED, and no smoke, but a very warm USB plug.
So I’m confused. I could imagine how connecting Vbus and Gnd to my target board’s TX and RX (on an ATMega128) could have damaged the target, but it still works fine. I wouldn’t think grounding the target device’s TX line, or the adapter’s RX line would be a big deal. All the other components test fine, so I guess I must have created an internal short in the CP2102. I’m having trouble figuring out what that trace was connected to but that might shed some light on things. If it’s just the chip that’s destroyed I might throw one in with my next Digikey order and try my hand at surface-mount soldering again.
The burned-out trace is the Vcc trace. The part that goes left in your picture goes to the user 5 V connection, and the part to the right goes to the internal power for the board. However, given that via placement under the USB connector, I think there might be a short to the connector, which is probably grounded (through the cable, not through the PCB). Can you see if that’s the case?
There is no short between the wire I added and the case of the USB connector at the moment, but it seems like something has changed again, because now nothing gets hot when I leave it plugged in.
The cable I’m using does connect the socket casing to ground, but oddly I’m not reading continuity between the USB socket’s ground pin and the ground header pin. Is this right?
I think I had the adapter connected to my powered-up ATMega128 board, and then plugged in the USB cable, so if there was a preexisting short between the socket case and that VCC through hole the action would have started then. Do think this was unrelated to how I had it hooked up then?
Just got the new one (thanks again) and I’m really happy to see that you’ve added pin labels to the solder mask. When I’m using one of these in a breadboard for testing I’m forever repeating in my head “Ground…T X…V Bus…R X”. You might want to update the picture on the product page.
