[evlatests] P-band switched power observed at S-band

[Barry Clark]

A nice piece of work.  We should do something about the problem.

I still worry that as well as the cal, some receiver noise from
the P band might be leaking in.  It can't be much, as Fraser says
it didn't show up on stress test gains, but it might be more
sensitive to use this approach to see if S band psum changes
when the P band receiver is turned off.

The simplest fix for most purposes is to run the S-Band Pcal
at half the frequency as the other cals (or twice the frequency,
or displaced in phase by 90 degrees).  Paul will hate this
suggestion - it does nothing to help the pulsar problem.

On 06/17/2016 02:42 PM, Paul Demorest wrote:
> hi everyone,
>
> While looking into various switched power issues recently, I noticed
> that the state of the low-band (4/P) cal switching causes a significant
> change in switched power (aka Pdif) measurements done at S-band.  This
> is not a subtle effect; for several antennas the S-band Pdif changes by
> a factor of ~1.5 to 2 (!) when the low-band cals are switching.
>
> Note this is _not_ the same effect as the gain modulations that lead to
> apparent "Pdif compression" as we have also been discussing recently. As
> far as I can tell, the low-band cal switching has no detectable effect
> on the amplifier gains at S-band.  Rather, the effect observed here is
> that when the low-band cals are switching, there is an extra amount of
> switched power added to the S-band Pdif.
>
> This is easy to test by separately controlling the state of the P- and
> S-band cal switching and plotting the resulting Pdif values vs time, as
> produced by the correlator and recorded in the SDM switched power table.
>   The attached three plots show the results of this for three different
> antennas.  In these plots, the red labels and dashed lines show which
> cals were enabled at different times during the test.  All cals other
> than P and S were disabled the entire time.  Note that at each scan
> boundary (black dashed lines) the system temporarily reverts to the
> default state (both S+P cals on) until I send a command to change it.
>
> The different antennas show a wide range of behavior:  ea01 looks great
> - a consistent S-band Pdif is seen independent of P-band, and the Pdif
> level goes to zero when the cals are turned off.  In contrast, for ea03
> the amount of switched power coming from P-band seems comparable to that
> coming from the S-band cal itself! (ea03 was the worst case in this
> test.)  The third one shown here, ea05, is somewhere in between with a
> ~10% effect.
>
> This is summarized for all antennas in the table below, which shows the
> ratio of the P-band-only to S-band-only Pdif values, sorted by the IF-A
> value.  The starred antennas are those with 4-band MJPs installed (at
> least according to the war room white board); there does not seem to be
> much correlation between this and the Pdif values.
>
> I have not yet looked at this effect on receivers besides S, or checked
> carefully for frequency dependence within S-band (this test used a
> single 128 MHz subband centered at 3.0 GHz); I may look into this some
> more in the near future.  Other questions, comments, suggestions are
> welcome.
>
> Cheers,
> Paul
>
> ----
>
> P/S Pdif ratios
>
> Ant    IF-A   IF-B   IF-C   IF-D
> ea03*  0.758  0.763  1.355  1.350
> ea07   0.503  0.492  0.512  0.506
> ea09*  0.416  0.422  0.292  0.286
> ea15   0.400  0.400  0.282  0.295
> ea25   0.247  0.299  0.015 -0.014
> ea21   0.244  0.248  0.457  0.498
> ea26   0.151  0.158  0.125  0.116
> ea12*  0.136  0.133  0.494  0.474
> ea05*  0.119  0.122  0.077  0.074
> ea14*  0.112  0.116  0.147  0.146
> ea28   0.097  0.102  0.065  0.065
> ea23*  0.086  0.087  0.167  0.179
> ea04   0.074  0.077  0.061  0.057
> ea16   0.072  0.072  0.073  0.070
> ea13*  0.065  0.057  0.029  0.030
> ea24   0.061  0.060  0.060  0.060
> ea27*  0.054  0.054  0.057  0.059
> ea17   0.050  0.048  0.069  0.050
> ea06*  0.044  0.044  0.049  0.049
> ea22   0.039  0.036  0.054  0.063
> ea20   0.038  0.037  0.048  0.051
> ea19*  0.023  0.024  0.008  0.008
> ea18*  0.002  0.001  0.006  0.006
> ea01* -0.000 -0.000 -0.000 -0.000
> ea10* -0.001 -0.000  0.004  0.004
> ea11* -0.008 -0.007 -0.021 -0.021
>
>
> _______________________________________________
> evlatests mailing list
> evlatests at listmgr.nrao.edu
> https://listmgr.nrao.edu/mailman/listinfo/evlatests
>