[evlatests] New vs Old ACU Tests Report

Wed Jun 25 12:17:43 EDT 2014

     All:

     Attached is a report of the test done Monday, with plots.  (The 
report is at the bottom of the attachment list)

     Rick
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-U-Az-E0.eps
Type: application/postscript
Size: 54802 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-U-El-E0.eps
Type: application/postscript
Size: 49622 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0001.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 21-U-Az-A0.eps
Type: application/postscript
Size: 49440 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0002.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-U-Az-A0.eps
Type: application/postscript
Size: 49514 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0003.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 21-U-El-A0.eps
Type: application/postscript
Size: 48985 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0004.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-U-El-A0.eps
Type: application/postscript
Size: 48885 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0005.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 21-C-Az-E0.eps
Type: application/postscript
Size: 44939 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0006.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-C-Az-E0.eps
Type: application/postscript
Size: 44939 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0007.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 21-C-El-E0.eps
Type: application/postscript
Size: 44510 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0008.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-C-El-E0.eps
Type: application/postscript
Size: 44710 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0009.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 21-C-Az-A0.eps
Type: application/postscript
Size: 48217 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0010.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-C-Az-A0.eps
Type: application/postscript
Size: 48217 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0011.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 21-C-El-A0.eps
Type: application/postscript
Size: 54818 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0012.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-C-El-A0.eps
Type: application/postscript
Size: 54818 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0013.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-L-Az-E0.eps
Type: application/postscript
Size: 46732 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0014.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-L-El-E0.eps
Type: application/postscript
Size: 39638 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0015.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 21-L-Az-A0.eps
Type: application/postscript
Size: 51481 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0016.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-L-Az-A0.eps
Type: application/postscript
Size: 51481 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0017.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 21-L-El-A0.eps
Type: application/postscript
Size: 50925 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0018.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 14-L-El-A0.eps
Type: application/postscript
Size: 50925 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140625/4b77dd3e/attachment-0019.eps>
-------------- next part --------------
A test was executed noon Monday (23 June) to compare the ACU responses
in antenna 14 and 21 (with the new design) to that of the remaining
antennas.  

We used the 'antenna pointing mode' test.  In this, the antennas are
pointed successively to:

On Source
+ El at half power
- El at half power
+ Azimuth at half power
- Azimuth at half power
On Source

Each of these steps has 20 second duration.  Two cycles through the
full pointing regiment were performed for each band, for each source.  

The test was performed at L, C, and Ku bands.  For these, the
approximate offsets are:  15, 4.4 and 1.8 arcminutes.  

Visibility data were dumped at 10 Hz, permitting easy resolution of
antenna oscillations (the fundamental resonance is about 2.2 Hz).  

Because the new ACU response is expected to differ somewhat depending
on the rate of change of azimuth and elevation, we observed two
sources:
1)  A source chosen such that dAZ/dt ~ 0
2)  A source chosen such that dEl/dt ~ 0.  

ea21 twice failed to execute the pointing regimen: The first was at
L-band for the 2nd source (dEl/dt = 0), and second at Ku-band, for
that same source.  In both instances, the data were flagged for being
off source (and examination of the visibilities certainly showed the
antenna was nowhere close to the target).  ea14 successfully completed
all pointings.

The antenna gain solutions for each 0.1 second integration were
determined.  The plots referred to below are in decibel power units,
normalized such that OnSourcePower = 0 dB.  

Each plot shows either ea14 or ea21 aligned with a non-retrofitted
antenna, with the same power scale for easy comparison.  For all plots
the time axis tick marks are in seconds (not hours, as printed).  Note
than in each plot, the curve rises from near -4 dB to 0 dB, then
declines again.  This is because the antenna main beam has passed
through the source en route to the other side.  

RESULTS:

In every test except one, the new ACUs outperform the old by a clear
margin.  The single exception is discussed below in item 4-c.  

1) +El -> - El, when dAz/dt ~ 0

1-a) L-band:  Motion = 30 arcminutes.  Plot 14-L-El-A0.eps shows the
response for ea14 (new) and ea17 (old).  Plot 21-L-El-A0.eps shows the
response for ea21 (new) and ea24 (old).  The results are the same --
the new ACUs move smoothly through the peak (the maximum at 0 dB), and
settle to the new position with no overshoot.  The time taken is about
3 seconds.  The old ACUs overshoot badly (about twice the offset),
then recover.  Time taken is about 6 seconds.  

1-b) C-band:  Motion = 9 arcminutes.  Plots 14-C-El-A0.eps and
21-C-El-A0.eps.  Same result as for L-band, but the new ACU duration
to settle is 2 seconds, while the old is at 5 seconds.  Note the 2.5
Hz antenna resonance showing up in the 'old' ACU plots, but not in the
new. 

1-c) Ku-band:  Motion = 3.6 arcminutes.  Plots 14-U-El-A0.eps and
21-U-El-A0.eps.  Similar to C-band, except the antenna resonance is
now much stronger in the 'old' ACUs.  It is also (barely) visible in
the new ACU response.  A slight overshoot also now visible in the new
new ACUs, which extends the settling time to about 4 seconds.  

2)  +Az -> -Az, when dAz/dt ~ 0.  

2-a)  L-band, Motion = 30 arcminutes.  Plots 14-L-Az-A0.eps and
21-L-Az-A0.eps.  For the new ACUs, the response is nearly identical to
the elevation motion.  For the old ACUs, there is again a strong
overshoot, followed by a second overshoot backwards.  

2-b)  C-band.  Motion = 9 arcminutes.  Plots 14-C-Az-A0.eps and
21-C-Az-A0.eps.  Similar to L-band, with the old ACU overshoot very
large and prominent.  This is completely eliminated in the new ACU.  

2-c)  Ku-band.  Motion = 3.6 arminutes.  Plots 14-U-Az-A0.eps and
21-U-Az-A0.eps.  The same as C-band, except that the 2.2 Hz antenna
resonance is now visible, more strongly in the new ACU response.  It
is in the old as well, but is partially masked by the very large
overshoot.  

3) +El -> -El, when dEl/dt ~ 0.

3-a) L-band.  Motion = 30 arcminutes.  Plot 14-L-El-E0.eps shows a
very similar response to dAz/dt = 0.  This source (3C147) has much
higher SNR, which permits the small antenna resonance to show clearly
in the old ACU response (for ea24 here).  ea21 failed to point for
this trial.  

3-b) C-band.  Motion = 9 arcminutes.  Plots 14-C-El-E0.eps and
21-C-El-E0.eps.  Again very similar to the dAz/dt = 0 case, but with
the antenna resonance clearly visible in the old ACUs, but not in the
new.  

3-c) Ku band.  Motion = 3.6 arcminutes.  Plot 14-U-El-E0.eps shows the
responses for ea14 and ea17.   Again similar to dAz/dt, with strong
antenna resonance in the old ACU nearly completely suppressed in the
new.  A small overshoot is visible with the new ACU.  ea21 failed to
point for this trial.  

4) Az -> -Az, when dEl/dt ~ 0.  

4-a) L-band.  Motion = 30 arcminutes.  Plot 14-L-Az-E0.eps shows the
response for ea14 and ea24.  The response is notably different from
the dAz/dt ~ 0 case -- slower for both the old and the new.  There is
a small, and irregular-looking overshoot with the new ACU.  ea21
failed to point for this test.

4-b) C-band.  Motion = 9 arcminutes.  Plots 14-C-Az-E0.eps and
21-C-Az-E0.eps.  Again, the new ACU response is slower than for the
dAz/dt ~ 0 case at this band.  There is no overshoot for either ea14
or ea21.  The 2.2 Hz antenna resonance is clearly visible in the old
ACU response.  

4-c) Ku-band.  Motion = 3.6 arcminutes.  Plot 14-U-Az-E0.eps shows that
the new ACU has gone into large-scale, slow-motion, damped
oscillation, with the 2.2 Hz antenna resonance superposed.  This is
not a one-of occurrence -- the pointing cycle was executed twice, and
the same response was seen on the 2nd trial.  (And the same behavior
was seen in the motion from +El to -Az).  ea21 failed to point for
this test.  

SUMMARY

Of the 12 tests performed here, the new ACU clearly outperformed the
old in 11.  This means that the antennas stabilized to the new
position more quickly, and with much reduced antenna resonance.  

In one test, the new ACU was clearly worse.  This is for the short
move (3 arminutes) in azimuth, when the rate of change in elevation
was close to zero (i.e., near meridian transit).  

We have not attempted a test for super-sidereal motion.  This will
have to be done by others.  