I looked at the schematic and found the section that deals with vertical hold, there aren't to many components in this area so I started to probe around. After a short while I found what looked like a weird signal on the output of the 4049 at U88 pin 15. The 4049 is a critical link in several parts of the vertical circuit, and the cursor blink.

I didn't have a spare 4049, so I swapped one from what I believe is the cassette interface area on the SOL. The image is now rock solid, although I'd imagine putting the dud 4049 has screwed with my ability to load and save to cassette. I have ordered some new 4049's and that should be here soon and I will replace the funky IC in the cassette circuit.

I've been playing with my SOL-20 a lot recently. One of the side effects is that I have had a lot of components fail. In the last week alone I have had one chip fail causing a situation where no cursor could be seen, although the computer was still usable. Before that I had a 74LS163 fail that caused no picture at all. The no picture problem took me a while to fix, and I was a little nervous that I wouldn't be able to find the problem. Today I had a problem where the vertical sync went haywire. The image seems to be rolling at very fast pace producing multiple repeated images.

neoncluster.com

Screen rapidly rolling showing multiple cursors