More on frequency precision

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

More on frequency precision

Post by Sam » Fri Aug 13, 2010 12:11 am

When the MSA commands to a certain frequency, it can't always hit the frequency exactly. The DDS is used to fine tune the PLL clock frequency to get as close as possible. But if you use Show Variables and look at the Actual Final IF, which you might expect always to be 10.7 MHz (or whatever your RBW filter is), you will see that it actually varies slightly.

If you cabled the TG to the MSA input and did a VNA transmission scan in a small range around a low frequency like 1 MHz, you might expect that the transmission phase would be pretty flat. And it will be if you calibrate. But an uncalibrated scan shows an interesting effect. You will see what looks somewhat like noise. This occurs because when you move from one point to the next, the Actual Final IF may change a few Hz. This will cause a phase change, whose size depends on your RBW filter.

This illustrates one method that the MO frequency drift causes phase changes. The Actual Final IF changes as the MO changes, though the MSA is not aware of the amount of the MO change, so the displayed Actual Final IF does not change with temperature. As the IF2 changes, the phase delay through the final filter changes.

This also demonstrates a limitation to trying to achieve Ultimate frequency precision in the MSA. Even with a GPS disciplined MO, the actual MSA frequencies will vary slightly because of the resolution of the PLOs.

I don't say this as a criticism of the MSA. I'm just pointing out that there is a limitation to the benefits of a super-precision MO.

Sam W.

Post Reply