DRAFT REPORT ON PULSAR OBSERVING AT GBT
2001 August 9
Zaven Arzoumanian, Don Backer, Bryan Jacoby, David Nice, and Ingrid Stairs
worked in Green Bank during 2001 July 22 -- August 1 toward pulling together
pulsar observing capability of the Berkeley-Caltech Pulsar Machine, BCPM1,
for use with the GBT. Two nights of observing were provided to help shake
down initial observing modes. A second rack, BCPM2, was installed.
Ed Childers, Frank Ghigo, Toney Minter,
Bill Shank, and Amy Shelton (and others!) provided significant
support in areas of equipment installation, guiding us through CLEO setup
of GBT receivers and subsystems and integrating Monitor & Control of
GBT and BCPM. Analysis of recorded data was done during the week+ by
those listed and also Scott Ransom. Development of BCPM/PSPM code is continuing
remotely by Stuart Anderson, Don Backer & Bryan Jacoby. Development of
analysis -- first look -- is continuing by Zaven Arzoumanian and Ingrid Stairs
and others when they return from holiday. Vicky Kaspi and Jason Hessels
are looking into developing system temperature extraction. So that's the
main players involved with some attribution as to activity.
1. BCPM HARDWARE
BCPM1 & BCPM2 in place with sparc/20 M&C cpus; 100 Mbs links are about
to be installed to allow few MBs continuous data transfer from each
to the scratch disk on vortex.
BCPM2 has limited operating range associated, we believe, with slow
xilinx part on Master Board. Combiner Board 0 on BCPM1 may be a bit
flaky and will be swapped asap. The MF's on BCPM1 will be improved
by shipping them back to Berkeley during planned observing lull
in early September.
The IF interface between GBT and 85ft IF processors and the BCPM
IF distribution network is complete. GBT IFs come from racks A/B
into BCPM IF switch positions 1/2, respectively. The 85ft 610-MHz
IF comes into BCPM1 rack, is attenuated and then split to feed
IF switch position 3 of both BCPMs. Software control has been
established.
2. PSPM/BCPM M&C SOFTWARE
Menus for principal commands -- MONITOR, TAKEDATA, TIMING -- have been
established and are stabilizing. These provide paths for bringing
primary telescope information into BCPMs. The manager under development
by Amy Shelton and Toney Minter will allow single interface to both
GBT and BCPM(s). Current effort is toward BCPM1 operation.
Basic timing capability via TAKEDATA and via TIMING was verified
using GBT data from our test observations. {include David's para here}.
TAKEDATA with output to disk requires several in progress developments:
fast ethernet installation; mounting vortex scratch disk; development
of continuous writing to disk in modest length files.
3. GBT/GO/BCPM OPERATION
4. PROMPT DATA ANALYSIS
Zaven A has constructed a Tc/Tkl gui called
``surfdat'' to inspect timing data.
See Figure 7 below. This currently lives in /home/bcpm1/src/analysis/bin/.
5. DOCUMENTATION
A Guide to iBPP continues to be developed
with audience for both general user and expert user/developer.
The TeXt is under RCS but a .ps file is posted here.
My BCPM web site
was developed in 2001 March, and needs updating.
Amy Shelton has started an
official BCPM web page buried down under "projects" below the main GBT page.
6. RESULTS!
We link below with brief captions a number of figures from the commissioning
run at end of 2001 July. There are single pulses from PSR B1933+16 & timing
of same pulsar using "search" data. A detection of M15C, PSR B2127+11C, using
a half hour of search data at 1.4 GHz is in Figure 3.
Importantly we have our first look at RFI. Figures 4-6 display various
aspects of a low-duty cycle, 12s periodic signal that is apparrently
an "old friend" from 140ft days -- the 1292-MHz airport radar at
Lynchburg, VA. We are unsure of which passbands we were working with
and there is some uncertainty owing to in-progress work on dealing
with inverted sidebands and data headers. But we think we had ~160 MHz
band around 1400 MHz which would have suppressed response early in
data path at 1292 MHz. The signal -- via addition or multiplication -- is
observed in *all* channels (middle of figure 6).
Note that in the BCPM the channels are
derived via independent analog conversions in six chunks (8 MHz in
this case) followed by digital mixing & filtering with 16 channels each
for total of 96 channels. The time profile of this signal is also
curious (upper left of figure 6): following the impulse from the radar
the level dips and then recovers with a ~1s time constant. Is there
an AGC in the fiber transmission with a 1s time constant? At present
we have no calibration. The peak is some ~50 times the noise as shown
and for a 20-K system temp and 2 K/Jy and 1300s integration and 50 MHz
and 2 polarizations, one could estimate 1.4 mJy with the undershoot
being a tenth of that.
Fig. 1: Single pulses (.eps) from PSR B1933+16
using BCPM1 at GBT on 2001 July 27 (D. Nice).
 |
Fig. 2: Timing residuals (.eps) for PSR B1933+16 using BCPM1 at GBT on 2001 July 27 & 31 (D. Nice).
 |
Fig. 3: M15C - PSR B2127+11C (.eps) from GBT observations
using BCPM1 on 2001 July 27 (S. Ransom).
 |
Fig. 4: 12-Hz Airport radar RFI (.eps) from GBT
observations using BCPM1 on 2001 July 27 (S. Ransom).
 |
Fig. 5: Blowup (.ps) of 12-Hz Airport radar RFI from GBT
observations using BCPM1 on 2001 July 27 (S. Ransom).
 |
Fig. 6: Analysis (.ps) of 12-Hz Airport radar RFI
from GBT observations using BCPM1 on 2001 July 27 (S. Ransom).
 |
Fig. 7: SURFDAT gui for inspection of timing data (Z. Arzoumanian).
 |