If you have a spectrum analyzer, depending on its resolution, you can
also measure the phase noise directly.
When I worked in the Satellite Communications industry, we needed up and
down converters faster than our vendor could make them. They use to go
thru a long QA process and they were very reliable. We waved the long
QA test and tested the local oscillators in the field. If they were
bad, we sent them back and they replaced them.
If you analyzer does not have the resolution, you can notch the carrier
and get more dynamic range or use the classic setup , which is the
"loose" phase lock loop.
Brian KD4FM
John Ackermann N8UR wrote:
Tom Van Baak said the following on 12/29/2009 04:24 PM:
But, depending on the equipment you use, a phase noise
measurement (e.g., script L of f) doesn't require the source
to be long-term stable; you get pretty much the same phase
noise results in the first minute as you would a day or week
later.
As a practical matter, you need to allow an OCXO time to stabilize or
else the warm-up frequency drift may cause difficulties in the phase
noise measurement. You need to keep the DUT and reference in the
proper phase relationship, and if the drift is too fast the PLL may
not be able to keep up. From what I've seen using the TSC test set,
you'll get glitches in the flicker-noise region (say, around 1 Hz
offset); the floor isn't much affected.
I fully agree with Tom on the main point -- I'm not talking days or
weeks, but an hour or so is probably a good idea.
John
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
