[evlatests] Beam Cuts at L, X and K bands -- A VERY curious result

Rick Perley rperley at nrao.edu
Wed Jun 19 16:53:13 EDT 2013 

    I reported yesterday that the settling time for X-band horizontal 
beam cuts was close to 10 seconds (rather more than the 6 seconds we 
normally think of), and that a further test, taken at both X and K 
bands, using 20 second dwell times at each point, would be attempted.  
This has been done, and the results are 'most interesting'.   The 
observations were taken of the northern calibrator J1926+7358, and were 
taken at HA = +0.5.  (This eliminates any possibility of the effect 
being related to the dAz/dt ~ 0 problem). 

    For all three bands, (L, X, and K), I took two orthogonal cuts, in 
azimuth and elevation, with oversampling 5, and eleven points -- 
essentially spanning the primary beam.  For the frequencies chosen, the 
stepsize turns out to be 265 arcseconds at L band, 44 arcseconds at 
X-band, and 20 arcseconds at K-band. 

    Attached are 12 plots, exhibiting the curious range of behavior. 

    Start with X-band:

    1) X-El.png shows the elevation behavior.  Each step is accompanied 
by a short duration overshoot. 

    2) X-El-Zoom.png is a blowup of a step -- the overshoot damps out 
within 3 seconds. 

    3) X-Az.png shows the corresponding azimuth behavior. 

    4) X-Az-Zoom.png is a blowup of the step.  There is a very clear 
periodic oscillation of period 10 seconds, which damps out in about 8 
seconds.  The pointing off offset is large -- ~13 arcseconds about the 
correct position.

    If this were a simple physical offset driven by the servo system, we 
might expect the oscillation amplitude to be much larger at K-band.   
But it isn't ...

    5) K-Az.png shows the elevation behavior for the entire cut.  The 
profile is very asymmetric -- go to the bottom for a comment on this. 

    6) K-Az-Zoom.png is a blowup of a pair of steps.  The same 
oscillatory behavior is seen, but the period appears to be a little less 
(about 8 seconds),  and the pointing offset is less -- about 9 
arcseconds.  The damping time remains about the same, 8 to 10 seconds. 

    7) K-El.png shows the overshoot is both small and short in elevation.

     8) K-El-Zoom.png shows that the damping time is about 2 seconds.

    So on the basis of this, we might propose the effect is more closely 
proportional to stepsize (or, wavelength -- we can't tell the difference 
with these data), so that at L-band the problems would be really easy to 
see. 

    9) L-El.png shows the full elevation cut.  Note that here, the step 
duration is 10 seconds. 

    10) L-El-Zoom.png shows that, as at both other bands, the damping 
time is short -- about 3 seconds. 

    11) L-Az.png shows the full cut at L-band (with 10 second dwells at 
each raster point).  Clearly, the overshoot effect is much less 
prominent than at either X or K bands. 

    12) L-Az-Zoom.png is a blowup.   Although the apparent amplitude of 
the overshoot is small (and the SNR not as high at this band), the 
physical scale is large -- 80 arcseconds!  (This is a rough number, due 
to poor SNR).  The effect looks minor since the beam is so large.  The 
damping time (to the level perceptible) is about 5 seconds.  The period 
is not visible, but is consistent with ~10 seconds. 

    We can conclude:

    A) In elevation, the overshoot is minor, and is damped out within 3 
seconds at all bands (to the SNR we have available).

    B) In azimuth, the overshoot is significant, with an amplitude which 
is approximately proportional to stepsize (or wavelength), a period of 8 
to 10 seconds, and a damping time of typically 10 seconds.   This is not 
due to the resonant frequency of the antenna (which is about 2 Hz), but 
is due to a resonance in the servo system. 

    Comparison amongst antennas shows that the phenomena described above 
(for ea21) is typical for all antennas. 


    There is evidence that the damping time is now considerably longer 
than it has been in the past.  Ken recovered some old chart records of 
antenna settlings, done in the early 1990s in preparation for the NVSS.  
He reports that:

    A) At L-band, the settling times were 7 seconds and 4 seconds for Az 
and El, respectively. 

    B) At C-band, the settling times were 4 and 3 seconds for Az and El.

    C) At K-band, the settling times were also 4 and 3 seconds for Az 
and El. 

    It seems the elevation settling times are unchanged, but the azimuth 
times are now much longer at the higher frequencies. 

    We might now wonder what the effect of the extended azimuth 
oscillation is for referenced pointing.  In this operation, each step is 
20 seconds in duration, so there is adequate time for a good solution, 
** providing that the data taken during the oscillations are flagged 
out**.    But this does not appear to be the case.  For my holographic 
data, there is no flagging at all between steps.  More to the point, for 
referenced pointing data, only the first few (<5) seconds of data at the 
move is flagged.  The system's flagging tolerance is too large to remove 
these wobbles  -- so that much of the extended oscillation seen at 
X-band is not flagged.    (Vivek has shown this, and may send around 
some plots for general viewing).  The error will not be large, since the 
10 seconds of good stable data (after the damping) will overwhelm the 
few seconds of unflagged wobbling data.  Ken and Martin are currently 
looking at this issue). 

    Finally -- A comment on the asymmetric K-band azimuth profile.  The 
peak of the profile is not on the center position, but is close to the 
step preceding this.  The data had x-band referenced pointing applied, 
so the normal interpretation would be a K-band azimuth collimation 
error.  But this is hard to accept here, since the stepsize is 20 
arcseconds -- the collimation error would have to be ~15 arcseconds!  
Either an error in the application of referenced pointing (within the 
holography loop), or a very large collimation error. 


 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: X-Az.png
Type: image/png
Size: 48542 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: X-El.png
Type: image/png
Size: 46938 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment-0001.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: X-El-Zoom.png
Type: image/png
Size: 36978 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: X-Az-Zoom.png
Type: image/png
Size: 39034 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment-0003.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: K-El.png
Type: image/png
Size: 51407 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment-0004.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: K-Az.png
Type: image/png
Size: 51787 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment-0005.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: L-Az.png
Type: image/png
Size: 53168 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment-0006.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: L-El.png
Type: image/png
Size: 52434 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment-0007.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: L-El-Zoom.png
Type: image/png
Size: 39019 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment-0008.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: K-El-Zoom.png
Type: image/png
Size: 41278 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment-0009.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: K-Az-Zoom.png
Type: image/png
Size: 42067 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment-0010.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: L-Az-Zoom.png
Type: image/png
Size: 37822 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130619/82bb2641/attachment-0011.png>

More information about the evlatests mailing list