[evlatests] Tardy Subreflectors

[Rick Perley]

    Analysis of the L-band data taken for the 'flux densities' run this 
past weekend has revealed a remarkable dispersion in the time taken for 
antennas to successfully complete a change of frequency from P to L bands. 

    For the run, the L-band observations always followed P-band.  
Careful test observations revealed that in order to guarantee at least 
20 seconds of good L-band data, a *duration* of 70 seconds was needed 
for these L-band observations. 
   
    Review of the subsequent data (filled without applying the flags) 
revealed that although the amplitudes were indeed stabilized at least 30 
seconds before the end of the 70 second duration, the *phases* for many 
antennas were rapidly changing up to about 20 seconds before the end.  
This is clearly a result of the subreflectors still in motion. 

    Review of the data passed by the on-line flagging reveals a 
remarkable disparity between antennas, which always repeats for every 
source, and every observation:

    1) Antennas 5 and 16 are always the first to complete the change 
from P to L.  This is unsurprising -- neither has a P-band receiver, so 
the on-line system never sets the subreflector the the P-band position.  
As the P-band subreflector position is 'all the way up', the time taken 
for these antennas to get their subreflectors into position is 
reduced.   There is always 50 seconds of good data on the 5-16 baseline, 
hence it takes 20 seconds to complete the frequency change.  Is that 
reasonable? 

    2) Antennas 14, 19, and 24 are *always* the last to complete the 
change (and 14 is always dead last).  What is surprising (to me, anyway) 
is that it takes an *additional* 30 seconds for antenna 14 to stabilize, 
over and above the 20 seconds for the frequency change.  Hence it takes 
this antenna a total of 50 seconds to safely change from P to L band.  
(Antennas 19 and 24 are always about 5 seconds ahead). 

    Ken tells me that the distance from the P-band subreflector position 
to the L-band position is typically 15 centimeters, and he figures that 
no more than 30 seconds should be required to make the move.  Some 
antennas (2, 3, 8, and 25, for example) more or less meet this 
expectation.  Some (14, 19, 24) aren't even close, taking up to 30 
seconds longer. 

    So the question is:  Why the large differences between antennas?  Do 
the tardy antennas have slow-moving subreflectors?  Or are they getting 
their commands late?   Or is the travel distance for these twice as far?