[evlatests] phase noise on outer antennas

[Ken Sowinski]

Since you first brought this up I have wondered whether the strict pad
dependence implicated a problem with fiber.  Your summary here reinforces 
that as an explanation.  Perhaps a check of LO quality at W32 (it's easier 
to get to) is in order?

Ken


On Mon, 28 Oct 2024, Paul Demorest via evlatests wrote:

> hi all,
>
> Now that we're in A-config, recent testing at high time resolution (10ms) has shown excess short-timescale phase noise affecting the outer three antennas (ea21 at E72, ea24 at W72 and ea18 at N72), as well as ea26 at W32.  A few plots are attached so you can see what I mean.  These were done via 10ms-dump-time observations of a bright source at K-band, and show two of the bad antennas (ea18, ea21) plus a good antenna (ea13) for comparison.  I also took data at X and Ka bands, a summary of these results is in the attached text file.  The worst ones have short-term phase RMS of ~20 deg at Ka, which is >10x larger than the good antennas.  This is enough to cause decorrelation (sensitivity loss) at the ~5-10% level and may be at least partially responsible for reduced high-freq performance seen on these antennas in recent stress tests.
>
> This noise has the following properties which make me suspect something like LO phase jitter is the cause:
>  - It looks totally random vs time.
>  - All 4 IFs for a given antenna show exactly the same noise (see zoom-in plot for example).
>  - The amplitude of the noise scales in proportion to observing frequency, higher freqs are more affected.
>
> It's also notable that all 3 outer antennas looked fine in B-config, and only started showing this noise once they moved to their A-config locations.  I found some 10ms X-band data from last A-config (Aug 2023) and it shows a similar pattern:  the outer pads and W32 are noisier than the rest, even though some different antennas were involved at the time.
>
> Since this seems to be pad-related rather than antenna-related, my hand-wavy guess is maybe LO optical power is getting a bit too weak over the very long fibers?  And as for W32 maybe it just has some problem with its fiber connection causing a similar effect.  Even if that's not the right explanation, I do think this is worth understanding and (if possible) fixing.  Let me know if you have any questions/suggestions.
>
> Cheers,
> Paul