Picoscope with odd channel issue

[alank2]

I bought a Picoscope on eBay - a 3203D with 2 channels.  Channel A was having an odd issue, but I am thinking maybe I fixed it.  I am concerned that maybe it isn't fixed so I wanted to get opinions on it based on what the waveform looked like.

The issue presented itself when I turned on the AWG with a square wave, 1kHz, 1.6V amplitude.  Channel B is fine.  Connected with a BNC cable between CH.A and AWG.  BNC connector even locked would switch between the failed signal and the correct signal if you would twist the angle of the cable.  I also tested it by just putting a jumper wire in the signal pin from the AWG to CH.A and the issue would either be present or not present.  This is where it is odd, if I bent the BNC cable so that it was failing before removing it, then the failure was shown on the jumper, but if I bent the BNC cable the other way where it was fine, then the jumper would be fine.

The BNC port on the PCB is not loose at all, very solid with 4 heavily soldered grounds.  I did reheat the signal pin on the PCB, but I am not so sure this fixed it.  It might have improved it a small bit, but I think the problem was still present.  Then I decided to clean up all the left over flux that you sometimes see on PTH ports on devices with some IPA.  There was quite a bit of flux.  Once I cleaned that off, the problem seems to be gone for good, I can't make it come back.  Still, I wonder if it was the flux that was the issue.  It is humid today.  Part of me was really thinking there was some sort of issue inside the BNC connector itself.  Does any of this make sense with what the CH.A picture shows?

What I really find odd is that the amplitude of CH.A. is wrong when it was messed up.  I'm not sure how strong the output signal is on the AWG, so I wonder what relationship that has to whatever was going on with CH.A.

Measuring the rise/fall times with the Leo pulser I have shows CHA rise=3.731, fall=3.889, avg 350/avg=91.9 MHz, CHB rise=3.822, fall=3.972, avg 360/avg=89.8 MHz.  Not bad for a 50 MHz model.