Keys up but no output

[Chuck Ritola]

I finally got the parts together to start testing my new RS-HFIQ (Rev D SN 601 Firmware 2.4a) but the TX output is extremely weak, measuring about -25dBm when everything is cranked to maximum. Changing bands doesn't mitigate. Analyzed the board, no anomalies other than the final FET looking like it was hand-soldered, with a smoky area around the solder points on the bottom of the board (not the side where the FET is).

The TX input seems deaf. No matter how much signal I cram into it I cannot get the clip light to come on when keyed (with one exception below). I'm using a Behringer UCA202, originally with the RCA outs to the TX input but switched to the headphone outs with volume at maximum for extra output. Verified with headphones that the signal going to the RS-HFIQ was hot but it still didn't work.

Early in testing the clip light was stuck on. I assumed something crashed in firmware or something so I powered everything down and brought it back up and the clip light was off and pretty much never came on again even if I tried to make it come on. This was before I measured the power - it was just routed into the dummy load while I figured things out.

The lights otherwise behave as expected: Yellow light comes on when keying up, both USB and power lights on, case warms up a little when keyed up, when keying down I can hear a pop in the RX signal from Quisk so there's at least some switching taking place.

Here are the 13.8V supply currents measured depending on condition:

Power switch, no usb: 100mA
Power switch, with usb: 180mA
No power switch, with usb: 0mA
Power switch on, with usb, quisk started, no keyup: 200mA
Power switch on, with usb, quisk started, no keyup: originally 680mA and slowly ramping up to 720mA

RX sort-of works: When I connect a random-wire dipole to it I can see equipment noise in Quisk which changes with tuning but can't pick up anything meaningful, including a strong local AM broadcast station. I even tried transmitting WSPR on 40 and got no spots, couldn't even hear it on a tabletop shortwave radio.

When setting it up I always had the RS-HFIQ plugged into a 300W dummy load and my keyups were no more than 30 seconds so I'd be very surprised if the final was blown.

Replaced 12V power supply as Bjorn suggested with 14-15V, no improvement.

Any ideas?

[Chuck Ritola]

Correction: The last current reading is WITH key-up. Also, the TX/RX jumpers were reseated with no improvement. Could the bias be set wrong? I see the adjustment resistor on the board and in the schematic; what readings should I be looking for with this board?

[jcve...@gmail.com]

The bias seems to be set correctly at ~500mA. Is the RX working OK?

[Chuck Ritola]

Thanks for the clue; I think I figured it out. There is a race condition between python's serial API used by quisk and linux where there is an indeterminate delay between requesting a serial port be opened and when it can accept communication. During this delay, traffic may be discarded. This caused the *OF2 command to be ignored, leaving it in mode zero, such that the RS-HFIQ was deaf and mute.

I made a fix involving a handshake loop in the open() method using the *W command instead of a time delay. If it holds up to testing and usage I'll file a bug report and propose fixes.