From: David Nice PALFA colleagues, This is to update all of you on PALFA activities at the Observatory over the past week. The crew down here has been very busy, and there hasn't been time to communicate all of our news -- most of it good -- to the full collaboration. People on site for PALFA work include Paulo Freire, Jim Cordes, Wouter Vlemmings, Julia Deneva, Dunc Lorimer, David Champion, Fernando Camilo, Jason Hessels, Ramachandran, Beth Reid, and David Nice. (Julia, David C, Jason, and Beth are grad students from Cornell, Jodrell, McGill, and Princeton, respectively). 1. Search pointings. 2. Quick-look processing. 3. New pulsars! 4. Old pulsars! 5. Feed rotation. 6. The database, or how I learned to stop worrying and love MySQL. 7. Disks disks disks. 8. Upcoming observations. 9. Next generation spectrometers. 10. Upcoming proposal deadline. 11. No PALFA meeting planned for Aug/Sep. 12. More Information, PALFA logo, etc. 1. Search pointings. We have had six sessions of Galactic plane time and four sessions of anticenter time. There are two more anticenter sessions this week. Pointings have been chosen from a "sparse grid" developed by Paulo -- start with his "type 1 tiling," which covers nearly every point within the 3db contour of one of the feeds, but only observe every third grid point. The other points can be filled in later. The Galactic latitude range covered is somewhat larger than that in the proposal, and the longitude range in the anticenter region is also larger. Due to a 1950/2000 coordinate error (don't ask) some of our observations were made a bit further off of the plane than originally desired, but the pointing grid has been arranged so that the erroneous points mesh with the full grid. The first pulsar was discovered using the erroneous grid. Integration time has been 135 seconds along the plane and 67 seconds in the anticenter. Ramach generated some maps of the search pointings. plane: http://www.naic.edu/~palfa/notes_04Aug08/gal_plane.ps anticenter: http://www.naic.edu/~palfa/notes_04Aug08/gal_anticenter.ps closeup to show beam pattern: http://www.naic.edu/~palfa/notes_04Aug08/beam_blowup.ps These plots were generated *entirely* from information in the PALFA database. Black = pointing grid Red = pointings with beam 0 Green = pointings with beams 1 through 6 Yellow = known pulsars from ATNF catalog 2. Quick-look processing. Dunc and David Champion have set up a data processing pipeline. This has been the highlight of all the efforts this week, and has worked far better than any of us could have anticipated. During an observation, data are written to WAPP disks. Immediately after the end of a scan, data are FTP'd to the ASP cluster, where they are split (each WAPP file has intermingled data from two beams) and processed using sigproc routines. Data are decimated by a factor of 16 in time (from 64 us to 1 ms) and a factor of 8 in frequency (from 256 to 32 channels). Processing of a typical 2-3 hr run is completed around an hour after the run ends. Candidates are posted on an internal webpage, and we've all spent many hours staring at candidate plots. Some of the data have been put through "long processing", with 128 us time resolution and 64 channels. This takes about 3x real time on the ASP cluster. (Note that the primary purpose of the ASP cluster is realtime data acquisition, so it isn't always available for off-line computing tasks such as palfa "long processing." We expect the ASP will normally be available for quasi-realtime quick-look processing.) There is not presently a definitive plan for "real" processing of the data at full resolution. This must be discussed (and acted on) soon! Data are recorded in a 100 MHz passband centered at 1420 MHz. The passband has been remarkably clean, and little RFI zapping is done in the quick-look code. About 15% of beams show a large number of long-period (> 2 second) candidates, probably indicative of strong RFI, but the rest are clean. 3. New pulsars! There are one confirmed pulsar and two strong candidates. PSR J1928+1746 P=68.7 ms DM=174 Confirmed PSR J2009+3325 P=1428 ms DM=255 Unconfirmed after two tries :-( PSR J0540+3208 P=524 ms DM=120 No confirmation yet tried The 68 ms pulsar has been timed over several days, and might be a young pulsar with a p-dot of around 1e-14 for an age of 80 kyr. PSR J2009+3325 looks excellent in the original search data but, surprisingly, has not been redetected. PSR J0540+3208 was discovered this morning and there has been no chance to confirm it. 4. Old pulsars! Several old pulsars have been blindly re-detected. All are from the Hulse-Taylor survey except J1913+1011, which is from the Parkes multibeam survey. PSR J1913+1011 P= 35ms DM=170 SNR=15.5 2.8 arcmin away PSR B1914+13 P=281ms DM=219 SNR=214 1.8 arcmin away PSR B1914+13 P=281ms DM=219 SNR=10 7 arcmin away! PSR B1915+13 P=194ms DM=102 SNR=133 2.3 arcmin away (16 sec search) PSR B1925+188 P=298ms DM=166 SNR=15.7 4.5 arcmin away PSR B1919+14 P=618ms DM=74 SNR=24.8 0.4 arcmin away (22 sec search) (It isn't clear as I write why two of those searches were shorter than the standard 67 or 135 second integrations.) One old pulsar was not re-detected at a distance of 2 arcmin. This is a bit of a mystery. A pointed observation the following day easily detected it. PSR B1911+11 P=600ms DM=80 2 arcmin away 5. Feed rotation. There are some issues with the ALFA feed geometry which impact the search area and scheduling. First, the feed is supposed to be able to rotate, so that the hexagonal pattern of the outer beams can assume whatever orientation is desired. Due to a failed drive motor, feed rotation has not been possible this week. All observations have been made with a zero rotation angle, which means that the feed is aligned with the azimuth arm (i.e., the pattern rotates with the telescope.) This plays hell with pointing grid coverage. At first, attempts were made to observe particular grid points at specific azimuth angles, but this proved impractical, so we decided simply to tolerate mis-pointings due to lack of rotation. A second issue to keep in mind with ALFA is that its beam pattern is elliptically shaped, with major axis in the zenith direction. Thus, even when feed rotation becomes available (in a day or so), the beam pattern changes slightly depending on azimuth angle. The pointing offsets associated with the seven beams have complicated the question of "where was the telescope pointing?". We think we understand it, based on detections of new and old pulsars, but there is a nagging concern that the failure to re-detect candidate J2009+3325 might be due to pointing issues. 6. The database, or how I learned to stop worrying and love MySQL. Jason and Beth have developed a MySQL database and some tools to fill it with PALFA information. So far it includes: (1) Paulo's pointing grids, (2) Pointing records, harvested from the CIMA log, (3) Search results, from the "quick-look" processing, and (4) A catalog of known pulsars. Tasks like cross-checking the pointing records with the pulsar catalog become relatively easy, and MySQL/python/pgplot has proven powerful for making plots like those given above. 7. Disks disks disks. Data are generated at a rate of 64 Mb/sec or 230 Gb/hr. It is recorded onto the WAPP disks. It is copied to the ASP for quick-look processing, but then deleted from the ASP disks. Arun (who deserves our eternal gratitude) has been moving data off the WAPPs onto RAID disks. The on-site RAIDs are filling up, so things have been a bit touch-and-go in the short term. However, long-term prospects are good: -- there is a 12 Tb array to be delivered to the observatory any day now -- there is a plan for a 50 Tb array so that the whole of the precursor survey will be on-line here -- there is a $1.8M grant to the Cornell theory center to permanently house the data archive. Archiving will likely not be able to begin there until October because they are just now arranging the fund transfers and hardware purchases. It may be possible to process some data at Cornell before the archive is ready. -- there are several ~0.5 Tb "Lacie"-type disks for transporting data to processing sites or to the Cornell theory center in a manner to be determined. We anticipate needing about 12 to 15 of these disks, preferably with both USB2 and firewire ports. The current list of contributions and expected contributions will probably handle the immediate need. A guess at who can supply disks is: Cornell (5 x 0.5 Tb), Columbia (1 x 0.5 Tb), UCB (a few expected), Jodrell (two expected), McGill (has offered a few), UBC (likely can supply a few?). [Contact Jim C to discuss this issue further.] The disks are labelled digitally and physically (on the disk enclosure) as, e.g., palfa_conrnell_X where X=1,2,3,etc. Eventually disk location will be tracked in the MySQL database. Jim reports the following on the long-term aspects of archiving: A question that came up here concerns the longevity of the CTC archive and whether there will be a duplicate copy of the raw data anywhere. The short answers are (a) a long time (30 years for the tape media; > 20 years for the CTC as an institution) and (b) no. The robotic tape system + hard drives + associated computers are expensive enough (including personnel to maintain the hardware and database). So we will be counting on the longevity. In my estimation, this is far better than any of our sneakernet archives of old-style Exabytes, DLTs, SDLTs (9-track tapes?) on which we have data. 8. Upcoming observations. Observations start up again in late August, with five Galactic plane sessions and seven anticenter sessions between August 27 and September 6. To round out the 100 hours of the precursor proposal, there should be eleven "3 hour days" in late September or early October. There is nothing that precludes remote observing and remote quick-look reduction in the future. 9. Next generation spectrometers. No news is bad news. It is likely that the spectrometers won't be available for another year. In the meantime we will be limited to 100 MHz WAPPs and we will be unable to make parallel observations with EALFA, who also use WAPPs. It seems wise not to wait for the new spectrometers but rather to go full steam ahead with the search work. (This blurs the line between the "precursor" survey and the "real" survey, and suggests that we should modify our ongoing precursor observations to mesh with whatever we decide the "final" program should be. 10. October 1 proposal deadline. Technically, the current precursor search program terminates on October 1. We need to submit a short follow-up request to the Observatory director to continue this program. A full proposal is not needed for extending the precursor search. Details will be forthcoming from the Observatory. Fernando volunteers to take care of this. The Observatory expects the ALFA collaborations to submit major proposals for the October 1 deadline to cover one year of observations. After discussions among some of us here, a straw-man plan is to submit two proposals: (1) a major search of the Galactic plane (and possibly other areas; this has not been thoroughly discussed), including time for confirmation of candidates and quick characterization of binaries and other interesting sources, and (2) a long-term timing program, something like monthly observations of confirmed pulsars, possibly with denser runs to ensure phase connections. Jim and Fernando have volunteered to take the lead on the first proposal, and Ramach and I will draft the second. I cannot emphasize strongly enough that no decisions have been made on sky coverage, pointings, number of passes, etc., for the search proposal. These topics must be discussed among us all. And soon! 11. No PALFA meeting planned for Aug/Sep. There had been talk of a collaboration meeting in August or September, with a view to working on the October 1 proposal. The meeting has not panned out, so telecons will be arranged to discuss the proposal. 12. More Information, PALFA logo, etc. Bits and pieces of the project information are moving to the PALFA website http://www.naic.edu/~palfa. It will become fancier in due course. We now have a PALFA logo, courtesy of Jason; see http://www.naic.edu/~palfa/graphics/palfa_logo_1.gif. Millisecond pulsars spin faster than a kitchen blender. Or could there be some other meaning? -- David, for the PALFA blender subcommittee