My QDX with Rasp 4 and WSJT-X is now running (see problem in #99825).
The lack of QDX response was due to unconfigured firmware. I had initially used firmware1.6 sent with the QDX kit and the built QDX was running OK (both RX and TX). I had later updated the QDX firmware to 1.9. The QDX then failed to respond (RX & TX). At the same time I had updated the Rasp 4 OS. Since the QDX operating manual states: "most QDX users need never use the terminal applications, it is not necessary for ordinary operation of QDX with WSJT-X etc".
I started then to look for errors in the Rasp 4 OS and WSJT-X configurations, because I was led to believe that the QDX firrmware was already configured for my QDX. The firmware was then put back to 1.6.  On connecting the QDX with my PC via Putty, I discovered that all 5 bands were set to zero frequency, zero AF gains and TXs were disabled. After configuring the band table etc using the figure in the manual for starting values, the QDX now works fine with WSJT-X .

The operating manual states that 30 dB is about right for just noise in an empty band. When I put a 50 ohm resistor- bnc plug on the input, I get 17 dB noise on the WSPR display. 
73

If you go back to 1.09 and do a factory reset, does it set up properly?

I have installed 1.09 into the QDX and then just opened the Configurations with Putty. The configuration was the same as with 1.06, that is, including my audio gain change for 20m to 66dB and +13 Hz. to Ref. Freq. 
I then did a factory reset. The resulting configuration was the same but less the above own changes.
The QDX with 1.09 continues to work OK. So I am still wondering what had happend when I first updated the firmware to 1.09.

Another problem: RF sweeps are not all OK: 30m Band, freq. centre is at -15 dB; 80m Band, freq. centre is at -7dB on left flank of peak; 60m Band, freq. centre is at -17dB, far left flank of peak; 40m Band, Freq. centre is OK within the peak of level of about -4dB; 20m Band sweep is OK, freq. centre is within large peak. So this will be my next task.

TX outputs into a dummy load and power meter with 9V supply on all bands, except 20m, are about 6 watts; 20m Band gives about 4 Watts.  

Thanks for tip.

73, Robin, HB9DEF
 

On 12/03/2023 14:50, HB9DEF Robin wrote:

Another problem: RF sweeps are not all OK:

Robin,

https://groups.io/g/QRPLabs/message/97251 and the rest of the thread. This is what Hans says about a bad sweep.

73 Alan G4ZFQ

Thanks Alan (G4ZFQ) for the link (97251). 
I have found that I was getting generally poor RF and image sweeps, although the TX output was satisfactory on most bands. 
Two Image rejections suggested a problem(s): 40m -26.3dB; 30m -5.7dB !; 20m -1.8 dB!; 80m -25.9 dB; 60m -32.6 dB. (Rev 3 PCB: image values are after removal of C71,72,73,74 and replaced with shorts).
I have gone through the Hans Trouble List for the QDX: T2 de- and resoldered with colour shrink tubing for each lead identification.
L12 was replaced and rewould with an extra turn at the end (42). L12 resoldered and checked for continuity.
L4 and L10 were rewound with number of turns given in manual and not using their corresponding inductance values. L4 inductance increased from 0.53 uH to 0.73 uH; L10 increased from 0.58 uH to 0.86 uH
I know there was some discussion here that inductances should be measured at the intended filter frequency, but my inductance meter only uses about 750 kHz.

The image sweeps are now: 40m -30.5 dB (y-axis -6dB); 30m -34.4 dB (y-axis -11dB); 20m -32.1 dB (y-axis -3dB); 80m -26.2 dB (y-axis -7dB) and 60m -27.6 dB (y-axis -19dB)

The 20m and 30m band image sweeps are now better. The TX-Test gives: 40m ~6W; 30m ~6W; 20m ~5.5W; 80m ~ 6W and 60m ~7W using 9 volts.
The 4 BS170s under the washer press on an aluminium strip (28 x 15 x 2mm), which gets quite hot after two WSPR TX periods. 

I think I will not do any further work on it but just use it. It gives an excellent performance (20m WSPR signals received in VK and ZL).
73 Robin