With the basic functions of the 3580A restored (at least, it is sweeping again), some attention needs to be paid to the mechanics. Fortunately, all is good with the intricate tuning assembly and digital display, but the knobs have some cracks, probably, a combination of age and stress.
Everything taped up, and the cracks filled with rapid-set epoxy resin.
To apply the resin to the small cracks, you can use a piece of stiff plastic foil, cut to a tool of appropriate size –
Here, a few close-ups of the tuning mechanism. It has fast tuning, and fine tuning, a clutch, several gears – all good old analog technology.
Finally, some test of the sweep circuit – but how to test a 200 second per division (i.e. 2000 second per screen) deflection for accuracy and linearity? Well, I connected it to a 34401a multimeter, and recorded the values for several hours by GPIB interface.
As you can see, the sweep is very linear, only some minor deviation at low voltages (maybe connected to some offset voltages or similar effects of the operational amplifier), at least, we can’t see any leakage current of the capacitor, which would show up as increasing sweep time with higher voltage/later divisions.
Also interesting, see the accuracy of the sweep speed, with warm-up of the instrument (each measurement is 2000 second). Still, after all these years, well within the 5% specification of the sweep time! Amazing!