[evlatests] Phase change with antenna elevation.

[Vivek Dhawan]

Antenna tipping tests show that attempts to correct the phase change
with various tests (listed in K. Morris memo) have made no dent -
still ~30 deg change at X band for elevations 10 to 80 deg.

Wednesday's meeting produced some suggestions:

1. Enclose other LO signals in the round-trip measurement. e.g. the 4096
    could be divided and sent back as 512. Also check other comb lines,
    L301, L302, etc.    Hardware is being gathered for this.

2. In the tipping tests, use all 4 IFs, to see differences between the
    2 L302's on an antenna. Use widely separated IFs to see how the error
    scales with frequency. (A related test report at the end.)

3. Tip more than one antenna to see if they track.

       We already have 19 and 21 outfitted for these tests, 2 IFs each.
       If we have enough hardware, I suggest we add 11, 17 or 26 which
       show clear phase change. (1, 2, 4, 5, 8 and
       25 also show little change, but are taken for WIDAR)

4. Redo the earlier tests where CW was injected directly into the de-
    formatter. Use weak signal + noise. Earlier tests had only CW,
    the VLA digitizers were probably unhappy with this even though
    the levels were adjusted.

5. Cover K-band feed with absorber, disable ALC's to avoid amplitude
    change with elevation, from ground pickup - Avoids phase change with
    attenuator setting.

--------------

A 2-hour 'baselines' test was done at L-band instead of the usual X.
I fixed the antenna positions at their values found in the last run
at X band (i.e. phases flattened at X, June 28).

The result is:

a. VA 15, 27 show <10 deg phase change with elevation. All other VLA
    antennas show no change (~5deg rms, bad weather)

b. EA 1, 2, 4, 5, 13, 14, 18, 25 look good - little change with elev.

    11, 17, 21, 26 show a very clear, ~50-deg phase change over 10 to
    120 deg elevation change. The AC and BD IF's are the same.

c. Comparing the tip tests on 21 with this L band data, the magnitude
    and sign of the change is the same at X and L bands (as seen in
    AIPS).

d. The facts are clear - the interpretaion is not. Following the signal
    through frequency conversion and correlation to AIPS, the chain is

L band: L301  L302  T304   Samp.  Deformat Correlator IDCAF FILLM.
f GHz   12.9  14.4  4.096  2.048

X band:       L302  4.096  2.048  Deformat Correlator IDCAF FILLM.
               11.3

I believe the first 4 operations at L band, the first 3 at X band,
each change the sign of the phase. The last 4 operations make no
sign changes for these 2 bands. Is there is one module implicated
or exonerated by this test? I'm not sure.

Vivek.