Got a new function gen! (and with new I mean old**)
** But not that old
(Attachment Link) " alt="" class="bbc_img" />
It's a Minipa Mfg-4220. Definitely not a precision device, and seems to suffer from a bit of frequency drift, but judging from the specs it's apparently to be expected for an instrument in this range.
I'm currently testing its accuracy against a 10Mhz GPSDO, and seems to be off by a bit (as pictured). The manual mentions a "temperature controlled oscillator" so I was hoping that with some warm-up time it would improve a bit, but I've left it ON for like an hour now, and the last two digits are still going back and forth between 22 and 21, sometimes settling on 21 for a while.
While I wasn't expecting this to be SUPER accurate I may open it later to see if the oscillator can be calibrated or replaced by a higher-accuracy part.
It's interesting to notice that it's supposed to be able to measure frequencies of up to 3Ghz (the likes of which I'll probably never need to measure) so I this is a pretty solid addition to my collection of devices that go *WELL* beyond the kind of signals I'll ever work with.
Well the counter IS stable if it if flipping between 21 and 22, it's just not accurate.
Just about every digital counter (and other instruments) has an uncertanty of plus or minus one count of the last digit. So the counter is stable within the limits of your measurement. If the TCXO in the generator / counter has an adjustment (it should) then you can remove the 21 Hz offset.
EDIT,
Found a specification for the counter - 5 ppm per year so assuming it's not been adjusted for at least a year it is within specification. 5 ppm at 10 MHz is +_50 Hz. That said you should be able toadjust it to a better accuracy.
Agreed. Somewhere in the vicinity of the oscillator crystal there should be a trimmer cap. Using a plastic driver you should be able to set it within +/-1Hz.
Warm up times: 1 hour is an absolute minimum. For critical adjustments as this example I consider 2 plus hours a minimum. Sometimes I'll even leave an instrument idling for as much as 12 hours before a final adjustment.
That's kinda what I thought I was going to find inside. Will try to do this today after work, thanks! It's great to have some confirmation _and_ guidelines for the procedure.
And yes, the counter IS definitely stable. The frequency drift I mentioned is for the signal generation circuitry. That turned out to go away with some warmup time, but still; you can leave it in a set frequency, and it will start moving up and down within a range (it was originally going just in one direction and just kept going and going... but after warmup it stayed within a range, and does no longer try to escape to infinity). It's not dead stable. I think the datasheet mentioned something like 2% FS per range, which is
definitely meeting, though. (Although it was listed under "Precision" in the datasheet? I would have expected something like "stability" but that is nowhere to be found in the specs, so whatever)
Anyway, even if the output is not SUPER stable, it still gives a rather clean signal, and the counter seems to be quite decent, so if I can adjust the offset in the frequency measurements I think I'll be more than fine.
Thanks you both for your input!