I reverse engineered the sweep generation circuit.

It gets its clock from the line voltage. There is a tap on the transformer for that. Starting in the upper left corner. The signal gets limited with a diode, then squared off with those two transistors. A 14538 dual mono stable vibrator is used to set the duty cycles.
One of those two signals then gets buffered, fed into an integrating capacitor, into a Schmitt trigger and eventually into an integrator to produce the ramp voltage needed for the sweep.

There are five problems here.
1. I have no documentation for this unit. It has an external control connector, here labeled as J103, which would be cool to know the purpose of.
2. Knowing that would help me understand the purpose of the second signal generated by U101B and the MOSFET section.
3. The dual NAND gate 40107 is bad. It's got open drain outputs. I don't have a drop in replacement on hand. Still available on Reichelt though.
4. The value of C111 is quite large, making the charge curve incredibly flat and the adjustments very touchy.
5. The whole circuit is ass. You cannot control the sweep frequency, you have no external trigger. It's fixed at line frequency

I got it to sweep again but I was not happy because of the inflexibility of the whole unit

So I started to consider rolling my own circuit to get more out of it.
But is this really worth the effort? And how much effort do I want to put in there? Am I chasing some autistic obsession of the day down a rabbit hole with little to no gain?
As much as I was tempted to go full autismo on that and put a microcontroller in to control the sweep...
I want to eventually get a modern AWG, making this whole box obsolete

And then I got caught up in other things again life threw at me

Mhhhh... Sexy. Especially if the software hack still works to remove the artificial limitations and bring it up to the full 120Mhz the hardware is capable of