@David Hess
Thanks for the message. Please read on.
A quick update on what I did already.
I decided to mount the new chip on strips. I thought that mounting the chip on a conventional socket would somewhat restrict the air flow under the chip. Remember, the chip does not touch the mainboard even factory mounted, there is a space between the chip and mainboard just enough to fit that washer under the cooling sink.
Since the old Maxim chip is still usable until certain speeds, I decided to mount it in the strips and put the scope back together to see if it works. At least until the new chip arrives. I was eager to see if I messed it up during the replacement procedure.
Scope assembly worked ok with no problems. Powered it up and yes, still works.
Set the speed to 10ns on a 50Mhz wave, positioned completely to the left and as expected the wave started to behave exactly like in the videos I already attached.
HOWEVER...
I left it running and went to the kitchen to get a cup of coffee. When I returned in the room I checked again the behaviour by moving the position knob all the way to the left. Surprise surprise, no more drifting.
At first I thought I left it running at a slower speed but when I checked the readout it was at 10ns. Set it to 5ns, no drifting. However, magnification does the same mess to the wave.
This may or may not be good news since I don't have an explanation of the phenomenon.
As I mentioned before, after I put it back together the scope behaved as expected. I mean showing the expected fault. Now only magnification seems to do something bad there. If I magnify a wave displayed at 100ns, the magnification gets it to 10ns with the messed up wave. But without the magnification the scope behaves correctly at all speeds.
Thoughts?
Added three pics showing the chip before desoldering, the strips in place and the chip in place in the strips.