New Allen Telescope Array Development

• Feed Measurements
• SEFD Measurements

• These sources have been observed with the Beamformer

• Pulsars

• http://adsbit.harvard.edu//full/1994A%26A...284..331O/0000335.000.html
• Pulsars

• GUPPI Raw format
• Filterbank header
• Filterbank Ruby code from Dave
• Filterbank header spec (see section 3)
• Filterbank Voyager1 Tutorial/Jupyter

Email from August 07, 2017 from Griffin:

> On Aug 7, 2017, at 7:48 AM, Griffin Foster <griffin.foster@gmail.com> wrote:
>
> Hi Jon,
>
> I really appreciate you taking your time to help Andrew and I check out the ATA. Your knowledge was invaluable, 
and we learned a lot. On the way back yesterday we had a chance to think about next steps in trying to use the ATA 
for astronomy, and hopefully, you are interested in being involved further.
>
> In my mind, the next step is to simplify the digital chain and look at individual antenna outputs. We can setup a 
fast-dump spectrometer using a SNAP board. Sampling at say 450 Msps, we could sample the entire analogue band. With 
a SNAP board, we could get 2 or 3 inputs, then output spectra over 10GbE to a capture machine.
>
> We should be able to see astronomical sources very easily with a single dish. The 21 cm line should pop-out very 
quickly when pointed at or near the galactic plane. The gain should change significantly between a bright A-team 
source (Cygnus A, Cassiopeia A, Taurus A) and basically any other sky position. We should be able to fold and 
detect a bright pulsar like B0329+54 very quickly.
>
> I would suggest the following plan:
>
> 1. You come out to Berkeley, and we setup a SNAP board design to output spectra, which you take back to the ATA.
> 2. Run the spectrometer on each antenna (both polarizations) to get diagnostics:
>     * Observe the galactic plane for a few minutes, tuned to around 1420 MHz. This is a good check to see that 
there is nothing wrong with the dish, you should always be able to see the 21 cm line.
>     * Observe an A-team source, on for 2 minutes, pointed 10 degrees away for 2 minutes, across the band, say 1 
GHz, 2 GHz, 4 GHz, 8 GHz. We can use this to measure the sensitivity (SEFD) of each dish.
>     * Observe pulsar B0329+54, B0531+21 or B0950+08 for 10 minutes at say 3 GHz, this is useful to show that 
there is sensitivity and that the time reference is working.
>     * Observe an A-team source by moving the dish across the source to see the flux increase, then drop off as a 
function of the antenna beam gain, at around 3 GHz. This is useful to show that the pointing makes sense and the 
beam looks correct.
> 3. You can either transfer the output data, or we can show you how to look at it to make sense of things. What 
  ever we find, we can then figure out how to fix things.
> 4. Once we figure out how to report the diagnostics, you can keep the SNAP board at the ATA to regularly do these 
checks.
>
> It would be exciting to get the ATA in a condition to do astronomy again. Let me know what you think of this, or if you have any questions/ideas.
>
> -Griffin

Email from Andrew on April 10, 2018:

I just wanted to restart this conversation.

I think Griffin’s suggestions are great, here are a few other ideas.

Perhaps you could reduce the on/off time on the flux calibrator sources to the minimum permitted by the beamformer 
with max ~35% overhead.  e.g. if you can switch positions in 10s, observe for 30s ON, 30s OFF, and repeat a few 
times.

It would also be good to observe a handful of such sources - here is a good list: 
http://adsbit.harvard.edu//full/1994A%26A...284..331O/0000335.000.html

Positions for these sources can be found on SIMBAD, e.g.: http://simbad.u-strasbg.fr/simbad/sim-id?Ident=3C48&NbIdent=1&Radius=2&Radius.unit=arcmin&submit=submit+id

I would recommend cross scans at the nominal position in addition to the ON/OFF, as was done in the reference 
above.

For the purposes of analyzing these observations, we will want to be able to produce coarsely channelized and 
integrated spectral data, e.g. ~1s time integration and 1 MHz frequency resolution.  This should keep the data rate 
down.  0.1s and 100 kHz might be better for RFI flagging, but not strictly necessary.  As Griffin mentioned, 
multiple frequencies will allow us to assess performance across the band.
  
For pulsar observations, you can get coordinates and parameters from http://www.atnf.csiro.au/people/pulsar/psrcat/
  
The three pulsars suggested by Griffin below are good places to start.  For these, we will want to generate faster 
data products (100us integrations) with modest frequency resolution (100 kHz). 
  
For the moment, doing these tests with the beamformer configured with a minimum set of operable antennas is a good 
place to start.  Over the course of the next several months, we will want to repeat these tests with individual 
antennas.  It will be necessary to create some kind of ’standard’ data product for these tests.  If we start with 
simple total power (Stokes I), the ‘filterbank’ format makes good sense.

Cheers, 

Andrew

PSR B0329+54

RA: 03h 32m 59.368s
Dec: +54° 34′ 43.57″
In hours,degrees: 3.5498244,54.5787694

PSR B0531+21 (Crab Pulsar)

RA: 05h 34m 31.97s
Dec: +22° 00' 52.1"
In hours,degrees: 5.5755472,22.0144722

PSR B0950+08

RA: 09h 53m 09.310s
Dec: +07° 55' 35.75"
In hours,degrees: 9.8859194,7.9265972