Drift with 30 kHz RBW filter

This contains the scottyspectrumanalyzer yahoo group backup
Post Reply
Sam
Posts: 1151
Joined: Sat Apr 28, 2018 6:49 pm

Drift with 30 kHz RBW filter

Post by Sam » Thu Jul 29, 2010 4:08 pm

I have posted my scan of phase and mag drift of my MSA using the 30 kHz RBW filter:







The scan covers about 2.5 hours. Magnitude drift is only 0.1 dB, all of which occurs in the first 10 minutes. Total phase drift is slightly over 1 degree (not counting the first minute of warmup). After 20 minutes, phase drift never exceeds about 0.3 degrees in 15 minutes, and after about 100 minutes phase becomes very stable. The ambient temperature was stable during this test, so the results do not indicate the effects of temperature cycling caused by heating or air conditioning.



Clearly, a wider bandwidth is useful when phase measurements are most critical. The only downsides of the wider filter are (1) higher noise floor and (2) extremely sharp and deep magnitude changes are better measured with a narrow filter. For most uses, a wider bandwidth is fine. As I mentioned before, drift is primarily relevant to determining how often to repeat Through or OSL calibration. With this wide filter it might not be necessary to repeat the calibration at all in many applications.



These results suggest that most of the drift with the 2 kHz filter was the result of master oscillator drift, which causes small changes in the actual IF2 frequency, which alters the phase change as IF2 passes through the filter. Some of it could also be the result of the filter frequency changing slightly with temperature.



Sam W.

Post Reply