Pulsars with the GBT: Scott Ransom, NRAO

Transcription

1 Pulsars with the GBT: Scott Ransom, NRAO Basic Properties Current instruments Track Repairs Drift scan Large Projects Timing New instrumentation Backend Receivers

2 GBT Capabilities 100-m unblocked aperature (G = 2 K/Jy) Capable of seeing ~85% of the sky (EL > 5.5º) Excellent low-noise receivers: 350 MHz (~60 MHz clean BW) 820 MHz (~60 MHz clean BW) 1420 MHz (~200 MHz clean BW) S-band (~600 MHz clean BW centered at 1950 MHz) C-band (~1 GHz at 4.3 Ghz) Spigot is the main public backend GASP and CGSR2 are private coherent dedispersion machines

3 Current GBT Pulsar Instrumentation Pulsar Spigot (uses autocorrelation spectrometer) 100+ modes (only ~2 used) each produces ~25 MB/s 50 and 800 MHz BWs 1024x16-bit lags, μs sampling (main modes) 2048x8-bit lags coming very soon (by summer) GASP (Backer, Stairs, Nice, Demorest, Ferdman, etc) CGSR2 (Bailes, Jacoby, Kulkarni, etc.) 2- (CGSR2) or 8-bit (GASP) baseband sampling Up to 128 MHz BW capability (low DMs and all CPUs) Some modes of GASP will become public this year BCPM and Spectral Processor are still available

4 Track Repairs Summer 2007 GBT closed for track repairs from (at least): May 1 until Sept Track has had chronic (i.e. design) problems since 2001 Has cost ~$4.5M from NRAO ops over past several years Settlement from Lockheed- Martin may allow new initiatives...

5 350MHz Drift Scan Survey during Track Repair Team: Lorimer (PI), Ransom, McLaughlin, Hessels, Champion, Kondratiev, Lynch, Stairs, van Leeuwen, Kasian, Kaspi, Roberts, Cordes, Deneva (and others?) Spigot: 2048-lag, 8-bit, 82us 350MHz with 50MHz BW hrs of MB/s = 72+ TB! Pulsars@Home for West Virginia high schools... Expect: ~1 MSP/2-3 days, ~3-5 normal PSRs/day

6 Large Projects with the GBT Users Comm and others argued for more large projects for NRAO Up to 50% of obs time per telescope Sept 2007 starts first GBT large projects High-Precision MSP Timing (PI: Demorest) is first large pulsar project Monthly 800MHz and 1.4GHz obs of 12 MSPs Has become a high-priority for the observatory (i.e. $$$ and new instrumentation...) (Take these numbers with a grain of salt...)

7 High-precision MSP Timing cont: A Pulsar Dream Machine for the GBT Goal: Publicly available ~1GHz BW coherent dedispersion pulsar backend Team: NRAO, WVU, Berkeley, UBC, NRL... CASPER FPGA-based technology from Berkeley Modular HW and SW Useful blocks already (FFT, PFB, FIR, etc) Almost OTS... e.g. Parsons et al 2006; CASPER ibob with 2xiADC boards (2Gsps each) CASPER BEE2 compute board with 5 fast FPGAs and up to 12GB RAM

8 Dream Pulsar Machine Cont: 2 x iadc boards give 8-bit real sampling of 1GHz BW for 2 polns ibob board does initial coarse PFB BEE2 board for heavy lifting and output on 10gigE Fast switch sends to data server(s) Rough Design Stages /4096 chan DFB, total power By end of summer DFB + full Stokes Fall 2007? 3. DFB + full Stokes + folding Winter 2007/2008? 4. Coherent dedisp (2 x BEE2s?) Summer 2008? CASPER ibob with 2 x iadcs BEE2 board

9 L-high + S-low Receiver / Feed? L-band receiver has airport radar at 1230 MHz and satellites (GLONASS, GPS, Iriduim, etc) from MHz S-band receiver has Iridium and satellite radio ~2350 Mhz How about ~1GHz wide system? Mhz Cold filters ~15 K T recv? DM vars

10 820 MHz + S-low Simultaneous MHz of quite clean BW around 820 MHz ~600 MHz of semi-clean BW between MHz Two excellent MSP timing bands simultaneously Would almost certainly be some sensitivity losses, though... Other ideas?