I first repaired the minimum stuff required for testing: fixed broken traces,replaced the broken/missing resistors. Did not replace any caps, just removed the ones that were cracked and could cause a short. Also added the missing 14.318MHz crystal.
After that I was able to get the master VSA working, so I went on.
Second VSA sort of worked after reflowing it, but in the end I had to replace it. Most likely some of the traces in the substrate were damaged, as some excessive force was applied to it at some point (could see fracture lines under the microscope). When desoldering it, 4 pads also lifted from the PCB. These pads were located in corners, and probably already lifted separated from the PCB while the heatsink was forcefully removed at some point.
Anyway, after replacing all the missing SMDs, fixing traces and pads, new inductors, new crystals, new electrolytic caps, new VSA and some upgrades (128MB RAM, DVI output), here it is:
The attachment 20231215_211736.jpg is no longer available
The attachment 20231215_212321.jpg is no longer available
The attachment U1.jpg is no longer available
(sorry for the shadows, I'll upload some better pictures later).