<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Well, I thought I would wade into this discussion...<br>
<br>
1. There is indeed one pulsar timer per baseband pair, meaning 4 timers
in total. Each of these can be set for a different pulsar ephemeris.
The restriction is that all of the correlations happening within one
sub-band correlator *must* use the same ephemeris. With 16 sub-band
correlators, this leaves plenty of flexibility to track 4 pulsars
simultaneously.<br>
<br>
2. Pulsar gating and pulsar phase binning are indeed two different
modes, and operate as Michael describes. Given the pulsar phasing
capability, I can't imagine why one would want to use gating...but it
is there with the ability to generate different gates on different
sub-bands to track dispersion. <br>
<br>
3. Officially, the correlator does not support "bunching" of phase bins
at some particular phase of the pulsar. However, the signaling to
generate phase bins is controlled by software, and so this is entirely
possible...but you should decide if this is indeed an important mode
reasonably soon so the real-time s/w people can program in the
capability.<br>
<br>
4. With my limited experience, having to get the pulse phase correct to
track the pulsar seems like an nightmare to me. I think a better long
term solution, if bunching of phase bins is required and/or pulsar
gating is required, is a feedback mechanism whereby for the first few
minutes (or however long it takes to get a detection), phase binning is
used with even distribution of bins across the period to find the
epoch, and then that epoch is immediately subsequently applied for
phase bin bunching or gating.<br>
<br>
5. The 65536 phase bins mode is possible, however due to CBE and output
bandwidth limitations the time resolution would be on the order of 100
msec (for all spectral channels...decreasing with fewer).<br>
<br>
Brent<br>
<br>
Michael Rupen wrote:
<blockquote cite="midPine.LNX.4.64.0608251009190.15610@scamper"
type="cite">
<pre wrap="">Hi Walter,
nice writeup! A few comments/questions...
</pre>
<blockquote type="cite">
<pre wrap="">This message describes my best guess as to the astronomers interface to
the EVLA pulsar modes with the WIDAR correlator. At the moment it is
mainly for the benefit of David Harland who is working on the obs prep
tool. Any comments are welcome.
The pulsar gating hardware: (EVLA project book 8.2.14)
~~~~~~~~~~~~~~~~~~~~~~~~~~~
WIDAR will support one pulsar timer per 2 GHz baseband (4 total). Each
timer is driven by a polynomial which returns pulse phase as a function of
the current time (UTC). The pulse phase is in the range [0,1). Each of
the 16 subbands per IF can operate with a different fixed delay relative
to the IF timer. Up to 1000 time bins can be accumulated. Each time bin
has a start time and a stop time. Start and stop times are specified by a
"pulse phase" ranging between 0 and 1. I believe that just as for
frequency channels the number of pulsar bins can be different for each
subband. Note: pulsars' spin periods range from 1.5 ms to 8.5 sec.
</pre>
</blockquote>
<pre wrap=""><!---->
I thought we had one timer per each of *eight* 2 GHz basebands (four BB pairs,
2 pol'ns each).
I would really like to stick with "baseband" rather than "IF", or at least not
mix the two. The VLA's IF modes confuse every observer I've ever met.
I'm confused as to gating vs. time bins -- I thought these were separate
beasties. Gating just means "correlate only at these times, and store one
output stream"; time binning means more output streams. I'm not clear on the
"beating" between the gating and the phase bins.
A couple additions:
* WIDAR can produce 2000 time bins if you don't mind some dead time while
downloading. The CBE purports to be able to provide up to 65.536
bins per baseline, through software accumulation.
* You can dump as many spectral channels as you want (apart from the
data output rate), at a cost in the narrowness of the time bins. If you
dump all channels, you get ~200microsec; if you dump only
64/sub-band/baseline, you get ~15 microsec.
I will check on the number of bins vs. sub-band question. It's all
in DUMPTRIG so I believe you can do this, but I'll check to be sure.
</pre>
<blockquote type="cite">
<pre wrap="">Common use cases:
~~~~~~~~~~~~~~~~~
</pre>
</blockquote>
<pre wrap=""><!---->
Am I right in thinking pulsar monitoring requires phasing up the array for
maximum sensitivity? WIDAR will be delivered with 8 digitally phased
sub-bands for a total BW of 1 GHz. WIDAR can phase up all 18 sub-bands of
every baseband, but we won't initially have that hardware. Antennas may be
assigned to multiple phased sub-arrays; phased sub-arrays may be defined
differently on different sub-bands. This implies for instance that one has to
have accurate positions which may be different for the different sub-arrays,
and for the different sub-bands.
</pre>
<blockquote type="cite">
<pre wrap="">1. Monitoring a single pulsar
The same polynomial will drive all pulsar timers. A different delay for
each subband will be added to account for pulsar dispersion delay (a
frequency dependence in pulse arrival time due to propagation through
ionized intersteller medium). This delay is parameterized by a single
parameter called the dispersion measure.
</pre>
</blockquote>
<pre wrap=""><!---->
</pre>
<blockquote type="cite">
<pre wrap="">The user will choose a number of phase bins <= 1000 and will specify start
and stop times for each. In the case of > 100 channels, it is likely that
the bins will be uniformly distributed across [0,1) as would be used in
"phase-ignorant" observations where the observer does not know the pulse
phase prior to observation.
If the pulse phase is known in advance the astronomer may place a far
smaller number of phase bins at pulse phases of interest to him/her.
2. Monitoring up to 4 pulsars
Same as above, but in cases where multiple pulsars are in the same primary
beam of the telescope it will be possible to observe up to 4 pulsars
simuntaneously, each with its own polynomial.
</pre>
</blockquote>
<pre wrap=""><!---->
Am I right in thinking this limit of 4 stems from the maximum of 4 phased
sub-arrays per sub-band? Does this increase if one is willing to spend
fewer than the full number of sub-bands on each pulsar?
</pre>
<blockquote type="cite">
<pre wrap="">"Monitoring" can mean either "imaging" or "timing", two distinct science
goals. From the perspective of the correlator these two actions are the
same. Typically timing will favor denser phase bins and imaging will
favor shorter integration times.
</pre>
</blockquote>
<pre wrap=""><!---->
Any thoughts on pulsar searches?
Anything special for higher frequency or astrometric observations? Though
the latter may be boring without EVLA-II.
</pre>
<blockquote type="cite">
<pre wrap="">Specifying pulsar information:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The exact mechanics of this are still to be discussed, but I imagine the
following is nearly accurate:
1. Specify "Pulsar Observing Mode" which will require the following
information:
A. For each 2 GHz IF:
1. pulsar ephemeris (polynomial table) specifier (see below)
2. pulse dispersion measure (DM) (note -- the purpose here is only
to warn the user if the dispersion smearing within one channels
is detrimental. The DM used in driving the phase bins will come
from the ephemeris.
3. pulse period (for same purpose as DM above)
</pre>
</blockquote>
<pre wrap=""><!---->
For David: there will be eight 2 GHz basebands (actually four BB pairs).
</pre>
<blockquote type="cite">
<pre wrap=""> B. For each sub-band :
1. gate enable (default enabled)
2. peculiar gate delay (default zero. This is in addition to
dispersion delay)
3. number of phase bins required (default: 64???)
4. table of start/stop phases OR one or two start/stop times for
the requested bins to be evenly divided into. (default: [0,1) )
2. Before the observation begins (but perhaps well after the scheduling
blocks have been submitted) the astronomer must provide the ephemeris
which contains a series of polynomials each valid for about 1 or 2 hours
and a dispersion measure used to calculate dispersive delays. Each
submitted ephemeris should be labeled with a specifier that the observe
file references. The pulsar ephemeris would be loaded into a database as
is done for the VLBA correlator. The reason that the astronomer might
want to wait until the last moment to submit a polynomial is that many
pulsars are "unstable rotators" which have phases that can drift by 10s of
degrees relative to an ephemeris on 10s of days timescales. Along similar
lines newly discovered pulsars often have very poorly known spin
parameters. The use of the VLA to determine a position can be very useful
in refining the these.
</pre>
</blockquote>
<pre wrap=""><!---->
Are these likely to be refined during the observation? or is that too painful
to contemplate? I'm wondering whether one might start with an even
distribution of bins, then zoom in on the phases of interest.
</pre>
<blockquote type="cite">
<pre wrap="">Note to the non-specialists: Dispersion measure (DM) is the column
density of electrons along the line of sight to the pulsar. Astronomers
use units of pc/cm^3 (parsecs per cubic centimeter) which is dimensionally
an inverse area. In these units typical values range between 2 and 2000.
This number is never negative. The delay incurred is frequency dependent
and is roughly equal to
dt ~= 4150 * DM / freq^2
with dt in seconds, DM in pc/cm^3 and freq in MHz.
</pre>
</blockquote>
<pre wrap=""><!---->
Is the ionospheric dispersion likely to be important at the lower frequencies?
</pre>
<blockquote type="cite">
<pre wrap="">At observe time if no ephemeris has been loaded into the database the
pulsar mode should be disabled and "standard interferometry" mode should
be used. An alert should be triggered.
Open questions:
~~~~~~~~~~~~~~~
1. Is fast-switching inconsistent with pulsar mode where a fairly
substantial correlator mode change is required?
</pre>
</blockquote>
<pre wrap=""><!---->
I don't believe so -- we'd asked about this earlier. Again I'll check to be
sure. Assuming one can reconfigure the correlator on these timescales,
we will want observe a calibrator with a wide-open bandwidth, then derive
and apply appropriate phases/delays for each of multiple sub-bands and
phasing locations.
Alternatively, given the initial hardware limitations, one could imagine
simultaneously observing the pulsars in phased-array mode using 1 GHz, and an
in-beam calibrator with a broad bandwidth. This calibrator might even be (a
gated version of?) the pulsar itself.
</pre>
<blockquote type="cite">
<pre wrap="">2. How does pulsar mode interact with sub-arrays?
</pre>
</blockquote>
<pre wrap=""><!---->
See above for some aspects of this. A related question:
* Can one trade phased antennas for phased bandwidth? I think not, but
this should be checked.
Cheers,
Michael
_______________________________________________
evla-sw-discuss mailing list
<a class="moz-txt-link-abbreviated" href="mailto:evla-sw-discuss@listmgr.cv.nrao.edu">evla-sw-discuss@listmgr.cv.nrao.edu</a>
<a class="moz-txt-link-freetext" href="http://listmgr.cv.nrao.edu/mailman/listinfo/evla-sw-discuss">http://listmgr.cv.nrao.edu/mailman/listinfo/evla-sw-discuss</a>
</pre>
</blockquote>
<br>
<pre class="moz-signature" cols="72">--
Brent R. Carlson
<a class="moz-txt-link-abbreviated" href="mailto:Brent.Carlson@nrc-cnrc.gc.ca">Brent.Carlson@nrc-cnrc.gc.ca</a>
Tel/Tél: (250) 493-2277 (ext. 346) | Fax: (250) 493-7767
Design Engineer | Ingenieur Concepteur
National Research Council Canada | Conseil national de recherches Canada
Dominion Radio Astrophysical Obs. | Observatoire federal de radioastrophysique
P.O. Box 248, 717 White Lake Rd | C.P. 248, 717 Rue White Lake
Penticton, BC, Canada V2A 6K3 | Penticton, (C.-B.), Canada V2A 6K3
Government of Canada | Gouvernement du Canada</pre>
</body>
</html>