Seem like I damaged my NanoVNA V2 Plus 4. I measured wire with capacitance at the far end. Probably it was still charged to 24V when connected.

All measurements look the same, calibrated and uncalibrated. Also connecting load or short have no effect to results. Uncalibrater pictures attached. Other functions seem to work normally, Led601 and Led401 are on. I have also checked that Port1 outputs frequencies normally over the full frequency range in CW mode.

I thought this was easy fix and I have changed some of the MXD8641 switches with no luck.

Any hints what could be broken?

When you do the calibration, are you 100% sure that the 3 probes make contact?
Check this again. When you mount the probes, the dot on the screen must move in certain positions, if it does not move, it means that it does not make contact.
--
Eugen - YO9WHI

Hello,
On a dirrerent topic is there a transfer swtch available  for nanaovna ? (diy or built)
regards
john davis

Yes, I am sure I did it correctly. I cheated a little and did not insert the open standard as seen in pictures.

It used to work logically before. At the moment any kind of load at port 1 (CH0) has no effect to results. Connecting port 1 to port 2 moves the CH1 logmag ~+60dB.

--

Anssi

I finally got some time to play a round with my broken unit. I took x-rays of the board at work. All the bond wires on chips were intact. There was bad solder joint on one of the undocumented pars under the white coating. I took off the coating and corrected the soldering but that had no effect.

Those QFN-14L packages are a quite tricky to measure in reliable manner. Every time I think I found the faulty part, re-measurement gives an opposite conclusion.

Some Good hints are still wanted. Feels like I need to force myself to measure all the passives or to make a simulation model to verify my measurements.

--
Anssi

After forgetting this for while I managed to get it work now...mostly.
It was quite a ride. I probably changed all MXD8641's three times before realizing I had one broken via at the output of U3. soldering a wire trough the via fixed most of the stuff. Also in a closer inspection, U601 did loose the lock every now and then. This was caused by the too slow slope at the LE line. Removing C622 fixed it.
I also tried to fix the reset issue that is happening when I disconnect the power cable and switch to battery mode. Anyhow this issue seems to be feature of the IP5305 (U701). I could not find any easy way to get around it.

There is still one issue that I am facing. My results are really noisy under 20Mhz. 20Mhz to 4.4Ghz results are repeatable. Any Ideas what could cause this?

Moi

Have you checked those parts which are nearest the ports. I have seen posts here where changing those helped.

It looks like soldering quality is not 100%, I have seen other boards slowly dying at first power up.

T Leif