My QCX-20 froze up this morning. I was running WSPR on my QCX-20. It had run through several sessions - probably at least an hour's worth at 10 min intervals. At one point I went over to look lovingly at it doing its thing - somewhere around noon, I think, and found that the display was frozen. The clock had stopped, and there were question marks in the last character position of both lines. The unit was completely unresponsive to control inputs. Horrors!
A power cycle brought back to life, and it hasn't done it since. None of the finals seem to have let their magic smoke out.
Regretfully I wasn't thinking enough to be absolutely sure if it was mid-send, though I'm almost sure it was. Shame on me for such sloppy diagnostics; I guess I was too distraught to think clearly,
Thinks: "a reboot on a three second hold-down of both buttons would have been real swell feature..., oh, but that would need more discrete hardware... hmmm, I know what: I'll pull out a reset line to a button when I get a round tuit and put it in a case (the QCX, not the r...)".
Thinks more: "Hmmm, I wonder if it stopped transmitting when it locked up... better ask Hans if the QCX can fail in a 'key down' state."
Possible factors: I have about 10-12 feet of butch, 14AWG speaker wire providing power from the 18.8V 3A regulated supply (Radio Shack 22-504) to my 40m unit, and a 14" daisy chain of decent shielded twin to the 20m. Power supply is on my desk, QCXs are on my window sill. My 40m unit ran all last night with the same power wiring, and was alive and well this morning.
DC voltage at the power terminals only droops by a couple of tens of mV between Rx and Tx modes, so I'm not suspicious of that, though I suppose I could put a couple of hundred uF across the terminals as a decoupler.
Interesting observation: I had "Frame" set to 10, and "Start" set to 12, because that just happened to be the next "x2" start time available when I set up WSPR. I was momentarily caught off balance when, after a send at xx:52, the indicator set the next time to 12, not 02. It makes sense, now I think about it: not all frame/start combinations are divisible into 60, and such combinations would cause the start time to creep around the clock every hour, so the code restarts each hour with the specified start time. No problem, of course, it just threw me for a moment. I can imagine Hans smiling and saying "Yes, Julian, that's right, now do keep up!"