I just bought a GPS module and it was working correctly at 38400 bps. After some power cycles I wasn't able to get more NMEA sentences from and instead I was getting some garbage.
I hooked a scope and realize the bit time which corresponds to 24400 bps. I switched baurate on mu PC to 24400 bps and I can get NMEA sentences however I can't change baudrate any longer. Even powering it off keeps the 24400 baud rate.
Tried to flash the firmware but flash tool does not recognize the module (I assume because of the wrong speed)
I found another user with the same problem: http://forum.trenz-electronic.de/index.php?topic=93.new;topicseen#new
Yeah, nothing stopping you from redesigning it (as I will on the next version). The diode trick is a quick mod. to the existing board. Mainly what I wanted to get across was an explanation for the weird behaviour.
In my case the LED isn't even on the same board as the GPS as I have 78 components on a 22x40mm PCB half of one side of which is taken up by the MT3329. Anything that is just a debugging aid is on a 2nd 'suport board' which I can plug the flight board into. So the LED is on the support board which has its own power etc. I will just buffer the signal on the next version.
Anyway, enough of my issues, hope you find the information about the behaviour useful and good luck to you all - UAVs sound a lot more interesting that what I am doing :-)