Letting it warm up for two hours
as suggested in the service manual, I was pleased to find that the local oven-ized 10MHz oscillator running at a steady 10,000,032 Hz and drifting no more than +/- 1Hz. While that is well within spec, we will see if we can do better a little later. And to be honest, the oscillator was running at that frequency within ten minutes of turning it on and didn't move after that.
And how accurate did it need to be for contemporary analoge scopes? Their spec was usually 2%!
But at 500ns, things start to look not so good:
Same for mixed markers (500ns and 100ns):
There are lots of trimpots that interact with everything else. Maybe improving the PSU ripple will help.
It is not surprising that there's cross talk here, nothing in the box is shielded:
...
I will start by looking at the power supply more carefully tomorrow and see if I can clean it up. It has been a while since I worked on anything generating frequencies like this...
It is also worth understanding which parts are and aren't powered up on the various settings. Quite apart from overall power dissipation, the available signal power was insufficient to drive all subcircuits at the same time.
agree on all fronts. i am fairly certain that this box, without doing anything other than running through the cal procedure, will meet the manual specifications. the noise, crosstalk, drift, and so on won't make any difference if one is using this for its intended purpose, circa 1975.
but if i can make it do better, i probably will because, well,
i can.
the circuit design is really pushing the bounds of price/performance/size for that era, which also makes it an interesting thing to study, even if i can't improve the performance. building countdown timers using discrete capacitors as "ladles" and "buckets" and transistors as gates is pretty neat. the manual circuit description and schematics are worth looking at even if you don't have one.