[evlatests] Settling Times for Referenced Pointing

[Rick Perley]

    Ken and I continue our testing with the goal of improving high 
frequency pointing and efficiency.   One of our tests involves two 
orthogonal cuts through the beam (to determine its location and 
characteristics) -- one in azimuth and the other in elevation.   There 
is a marked -- and important -- difference between these. 

    When stepping in elevation,  the 'settling time' is very fast -- 
amplitudes are stable with 3 or 4 seconds, although a small amplitude 
oscillation of period 4 to 6 seconds is discernible.  The amplitude of 
this small oscillation is 1 to 2 arcseconds. 

    When stepping in azimuth, the behavior is very different.  The 
response shows a slowly damped oscillation of remarkably large 
amplitude:  The amplitude corresponds to an overshoot of ~8 arcseconds, 
the fundamental period of the oscillation is 8 seconds.  Even after 20 
seconds, the oscillation is still visible in the data.   The damping 
time is about 10 seconds -- after this length of time, the oscillations 
are sufficiently small that useful data can be collected.

    I presume the origin of this azimuth behavior lies in the servo 
system.  Certainly the characteristics dictate that we cannot speed up 
'referenced pointing' by shortening the stepping time to less than 20 
seconds.  Reducing the amplitude and period of the azimuth oscillation 
would certainly be helpful!

    I attach two postscript plots to show the effect.  The data are 
taken at K-band. 



   
-------------- next part --------------
A non-text attachment was scrubbed...
Name: AZCUT
Type: application/postscript
Size: 38130 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20101119/3da3005b/attachment.ai>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ELCUT
Type: application/postscript
Size: 39625 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20101119/3da3005b/attachment-0001.ai>