Hi,I used a low phase noise signal source as my artiq frequency referenceļ¼ŒThe reference frequency is 100MHz

This is the phase noise and spectrum of my external 100MHz frequency.
But my artiq DDS output frequency produces some other unknown noise besides the sidebands. And his phase noise curve is not very good

He looks bad, the frequency of the output produces other noise besides the sidebands.
Below is the frequency output and phase noise curve using artiq internal clock reference

After artiq connects the external clean source as the frequency reference, the phase noise curve should be better, but it doesn't seem to be, I don't know how to solve this problem, thank you all.
I modified refclk=100000000, clk_sel=1, pll_n=40 because the frequency reference I entered is 100MHz.

8 months later

Hi, I would like to follow up on that since we see similar issues with our systems. We have one crate with Kasli v1.1 running artiq6 and clocked with 125 MHz and one crate with Kasli v2.0.2 running artiq7 clocked with a10 MHz signal.

We see additional phase noise (that is not present in the clock signal) on the DDS (AD9910) outputs on both crates, however with different characteristics (see the phase noise plots). On the crate with Kasli v2.0.2 we see the same phase noise on the phaser output, indicating that it is not just a problem of urukul.

Phase noise of the DDS output from the crate with Kasli v1.1, clocked with 125 MHz (the left peak is also present in the clock signal, however not the high frequency noise).

Phase noise of the DDS/phaser output from the crate with Kasli v2.0.2, clocked with 10 MHz.

For comparison, here is the phase noise of our current 10 MHz source connected to Kasli v2.0.2:

Has anyone seen similar behavior, and what it the best clocking setup to minimize phase noise?

The first plot doesn't look like it is locked at all. The second looks like a standard 50 Hz ground loop. Try a balun or a different grounding scheme.