Well, I started to type something up and then went and looked at the owners manual for what you have online, and realized that most of what I was typing about was a moot point. Specifically, the FG I bought uses a quartz crystal timebase - it could have a better one, and it can take an external to allow use of as spiffy a timebase as you might like. Leaving it on for a few hours greatly improves stability as the crystal warms & stabilizes.
The FG you have, by contrast, uses a capacitor charged by a constant current as its time base, and both "presumably" and also "from a quick look at the manual" has a fair number of potentiometers (trimmer as well as front panel) in the circuit (presumably, these affect what the constant current is.) Those tend not to age well, IME. You can certainly try a can of contact cleaner on the pots. I don't know what they'd use for a timing capacitor, but I would assume it would be something (a type) more stable than an electrolytic that would be prone to drying out, so the pots is what I'd mostly suspect. They tend to corrode and get weird with old age, so the moving contact does not make reliable contact, and may be making some sort of contact but with a film of corrosion that affects the contact value.
Then again, a dried out electrolytic elsewhere in the circuit could be the culprit (the other standard old age issue) or something completely else could be wrong.
Since there does not seem to be a method to feed the thing a different timebase, the "simple in my case" diagnostic of putting a different timebase into it does not appear to apply.
I had used some of the "similar in concept" Beckman FGs in the past, which is one reason I went looking for a quartz-crystal based FG when buying my own (HP3325A, a few years past yours but still quite retro from these days.) I hated fiddling with the dial trying to get a precise frequency. Contrariwise, some folks hate the 3325A precisely because it does not have a dial....
Dial or no dial, I think stability and quartz crystal run together in these things. An ovenized crystal is better, and with an external reference input, you can get as precise as you can afford with GPS-clocking, rubidium references or the like (or just one really good ovenized crystal that you use for all your local reference inputs. At least your counter will agree with your generator in that case.)
------------------------------------ I tried to lubricate all contacts and potentiometer with no success. A capacitor? It could well be, but it is very difficult to locate the components on the board. Also, the schematic diagram drawing is of poor quality.
Frequency measurement results:
- Just turned on. Frequency = 5.000 MHz Delta F = 0
- 15 minutes after. Delta F = -98 kHz
- 30 minutes after. Delta F = - 116 kHz
- 45 minutes after. Delta F = -132 kHz
- 60 minutes after. Delta F = -141 kHz
- 75 minutes after. Delta F = -147 kHz
Again my question is: is my generator faulty or all instruments based on the same (analog) frequency generation principle (charging a capacitor at constant current) behave more or less the same?
73 Tony I0JX