<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Paul, thank you for re-running this test and plotting the results
against the last run. I agree it's pretty obvious the filters didn't
have a significant effect on the out-of-band emission in either
antenna modified (ea09 and ea11), so this is clearly not a solution.
<br>
<br>
Mert's idea about radiation off the board as a possible cause may be
correct. There are some long exposed traces carrying the cal signal
to different areas of the PCB, coming out of an 8-way power splitter
driven by the noise diode. The LBR enclosure is weathertight, but
would provide little (if any) shielding at S-band. The DC connectors
are also non-metallic and unfiltered, so RF could also be coupling
to the harness inside, and radiating off the LBR power cable leads
down from the apex. Unfortunately, there's little that can be done
here without a significant redesign of the box and board.<br>
<br>
One thing that might help is to try a noise diode with lower output.
The current device has an excess noise ratio (ENR) of ~51 dB, which
is pretty high. On the P-band side, the cal path already has a fixed
20 dB attenuator on the board, as well as a provision for an
external coax attenuator to set the Tcal level. Normally, the coax
attenuator ends up between ~10-15 dB, or 30-35 dB total attenuation,
ahead of the cal coupler. The noise diode manufacturer has another
part that has a lower output (31 dB ENR), but is the same otherwise.
We could retrofit the LBR in ea09 with this part, resetting the
attenuators to get the proper Tcals, and reinstall and retest in the
same antenna to see if we get any improvement.<br>
<br>
There was actually one difference between the two antennas we
retrofitted for the 2nd test. On ea11, we added filters on only the
P-band noise cal signal paths, while on ea09 an additional set of
filters were put in external to the receiver on the 4-band inputs
from the MJP dipoles. Both antennas have MJPs installed. <br>
<br>
In the ~4 weeks between the two tests, we did replace the LBRs on
ea01, ea06 and ea16, which would explain the changes seen on these,
versus the others which were quite stable, as Paul mentioned. The
flat line on ea04 may be due to a problem internal to the receiver,
like the noise diode or switch, or maybe something dead in the
control path to the switch. I'll make a note to check this.<br>
<br>
-Wes<br>
<br>
<br>
<div class="moz-cite-prefix">On 4/25/2017 10:18 AM, Paul Demorest
wrote:<br>
</div>
<blockquote cite="mid:e9b0a1a30e8eb73732aba977d7e88028@nrao.edu"
type="cite">hi all,
<br>
<br>
This is an update on the effect of the low-band noise cal (aka
switched power) system on L- and S-bands (see various old emails
from me to evlatests about this). Note that currently the
low-band cals are turned off during all non-low-band observations,
so this is not an ongoing problem.
<br>
<br>
Last week, Wes and co. installed some extra filters into the
low-band systems on ea09 and ea11 to see if this effect could be
reduced via a HW fix. To test this, I took some new data both
before (2017/03/30) and after (2017/04/24) the filters were
installed.
<br>
<br>
The short answer is that the filters do not really seem to have
helped. The pattern of power vs freq does not look exactly the
same as it did before, but the effect is clearly still there.
ea09 may even look a bit worse after this work. I don't know what
(if anything) this tells us about the nature of the problem. Any
ideas?
<br>
<br>
For those who want to see the data, the attached PDF shows the
results from the two S-band tests for all antennas where data were
taken. The quantity shown is the ratio of "on-off" power at 10 Hz
when P-band cals are switching to when they are off, as a function
of frequency. S-band cals were switching the entire time, so
ideally we want to see a flat line with a value of 1.0 (ie like
ea04 which for some reason is not showing this effect currently).
This was measured using binning-mode widar autocorrelations.
<br>
<br>
Most antennas show nearly identical results in the two tests, so
this effect is apparently stable over few-week timescales. The
power vs freq pattern clearly changed for ea09 and ea11,
presumably as a result of the filter work. It also changed for
ea01 and ea16. Maybe there was some other work on the low-band
systems on these antennas during this time range?
<br>
<br>
I also took L-band data but it turned out not to be very useful
due to a bug in the data recording software. ea09 and ea11 still
do show a strong effect at L-bang though.
<br>
<br>
Cheers,
<br>
Paul
<br>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
evlatests mailing list
<a class="moz-txt-link-abbreviated" href="mailto:evlatests@listmgr.nrao.edu">evlatests@listmgr.nrao.edu</a>
<a class="moz-txt-link-freetext" href="https://listmgr.nrao.edu/mailman/listinfo/evlatests">https://listmgr.nrao.edu/mailman/listinfo/evlatests</a>
</pre>
</blockquote>
<br>
</body>
</html>