[evlatests] L-Band Default Frequency Test
Rick Perley
rperley at nrao.edu
Mon Jun 17 13:24:31 EDT 2013
I have suggested that the default tuning range in L-band be set such
that the upper edge of the AC and the lower edge of the BD tuning be set
to 1520 MHz (rather than 1500). Some concerns were expressed that there
may be aliased RFI from the 'INMARSAT' satellite (1525 MHz and up)
showing up in the ~1500 -- 1520 MHz zone. I ran a short test over the
weekend to check this.
I observed the northern calibrator J1800+7828 for 2.5 minutes, then
a 'blank' field at J1800+8000 for 2.5 minutes. A tuning arrangement,
with the center frequency of 1520 MHz was made, with 1 MHz channel
width. Thus, the frequency span went from 1008 through 2038 MHz.
The J1800 source was used to calibrate the bandpass and set the flux
density scale (I know the latter from the recent 'flux densities'
runs). I then inspected the calibrated spectra (with Hanning smooth on,
of course!), and made images of the 'blank' field in both Stokes 'I' and
Stokes 'RL' for each channel from 1485 MHz through 1519 MHz.
The 'blank' field isn't blank of course -- there is about 250 mJy of
real flux in it, seen equally in all images.
Inspection of the 'I' images shows that every frequency channel
gives an identical image -- none show any of the effects expected with
strong RFI within the field. The rms of each image was calculated in
both 'I' and 'RL' -- this shows a slow and steady rise from from 1485
MHz to 1515 MHz (of about 10%) -- this is surely the decreasing L-band
sensitivity, as shown (for example) in EVLA Memo 152. The noise rises
sharply above 1515 MHz - this is as expected from the roll-off of the
anti-aliasing filter at the upper end of the IF.
One channel only within this band shows higher noise than expected:
corresponding to a frequency of 1504 MHz. Perhaps this is low-level
RFI? Nothing is seen in the individual spectra.
I made vector and ampscalar averages over all baselines longer than
500 wavelengths (see below for explanation of this lower limit),
excluding ea19 (see below for the reason). These show that the subband
from 1456 through 1520 MHz to be as clean as any other (and arguable
cleaner!). I've included these four for your perusal. HiVec =>
Vector averaged from 1520 through 2032 MHz, HiAmp => Ampscalar average
from 1520 through 2032 MHz. Loxxx => Corresponding plots for 1008
through 1520 MHz.
It will be noted that the noise in the uppermost subband (1968 to
2032 MHz) notably rises -- this is due to the receiver rolloff. But the
rise is not great (factor ~2), and it is certainly better than in the 20
MHz 'lost' at the lower edge (988 through 1008 MHz).
I recommend adoption of this setup as the default in L-band.
----------------------------------------------------------------------------------------------------
Some other details:
1) ea19 was eliminated due to very low sensitivity. (Long-standing
issue, known to FE).
2) ea23C had a DTS problem, and was flagged out.
3) ea25 in LCP (C and D) gave no fringes and was flagged (known
problem).
Inspection of the visibilities as a function of baseline showed that
all baselines shorter than 500 wavelengths showed much increased noise
-- for all channels. That is, it's a continuum issue. Both RR and LL
showed this, although 'RR' was more strongly affected in the subband
that I was looking at (#8). These short baselines were not utilized in
the images or the spectra discussed above. As this test was run at
about 7:30 AM, local time, I'm blaming the sun for this.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: HiVec.png
Type: image/png
Size: 87332 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130617/b63dfe87/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: HiAmp.png
Type: image/png
Size: 77905 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130617/b63dfe87/attachment-0001.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: LoAmp.png
Type: image/png
Size: 80698 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130617/b63dfe87/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: LoVec.png
Type: image/png
Size: 89963 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20130617/b63dfe87/attachment-0003.png>
More information about the evlatests
mailing list