IPAC 2MASS Working Group Meeting #62 Minutes

IPAC 2MASS Working Group Meeting #62 Minutes, 5/23/95

Attendees: C. Beichman, T. Chester, R. Cutri, T. Evans, J. Fowler, G. Kopan, B. Light, C. Lonsdale, M. Moshir, J. White

AGENDA

  1. Plans for Bar-Code Tape Inventory Control
  2. Cross-Scan Offset Misalignment in April '95 Data
  3. IRTS Use of 2MASS Computers
  4. Survey Strategy Meeting
  5. Position-Dependent Responsivity Study
  6. Effect of Seeing on Hybrid Photometric Algorithm

DISCUSSION

  1. Plans for Bar-Code Tape Inventory Control --- Based on current technologies, it is expected that the number of tapes sent to IPAC from the observatories will be between 5,000 and 10,000. Some form of automated tape inventory management is desirable. The U. S. Naval Observatory uses bar-codes and readers for this purpose, so it has been expected for some time that 2MASS will use a similar method. In order to get some progress started on this aspect of operational readiness, splinter meetings have been held with R. Beck, J. Fowler, G. Kopan, B. Light, and D. Wittman. The tape stacker on lugosi has a built-in bar-code reader, and its control software includes capabilities for programming it. D. Wittman will look into how to use this so that the 2MAPPS software can obtain the right tapes from the stacker without requiring any special tape-stacking order. There is also expected to be other non-pipeline software for automating certain aspects of tape inventory control.

    The tape stacker came with sample preprinted bar-code labels, and these all contain eight-digit numbers, implying that at least eight-character codes are compatible with the hardware and software. Nine-character codes are probably possible, and this is the number currently envisioned for tape labels, six for the observation date, one for the band, one to indicate the last tape for that band, and one to indicate hemisphere.

    B. Light will contact D. Monet at USNO to request information on their bar-code system; B. Light gave D. Monet the software for generating bar-code labels, and has tested this on IPAC laser printers with success. The main tape labels would be printed (or at least applied) at the observatories, but since the plan is to make backups at IPAC, we would also need this capability. R. Cutri has suggested considering less degradable media for the backups, such as high-capacity CD ROMs if the technology improves sufficiently by the time we could use it.

  2. Cross-Scan Offset Misalignment in April '95 Data --- J. Carpenter has observed an error in the telescope's scan-to-scan cross- scan overlap control in the April '95 data. This was discovered while constructing a mosaic of the galactic center. The overlap is not constant during the scan, but instead an error develops so that when (for example) returning to the northernmost declination at the end of a descending/ascending scan pair, an error of about ten pixels from the desired cross-scan relationship is seen. R. Cutri reported that the declination balance of the telescope had been found to be incorrect and was adjusted around May 3, so we may find that this problem has disappeared after then, although we have not yet processed enough data to know. Conversations with H. McCallon indicate that the observed error may not be particularly significant compared to the older protocamera data. A ten-pixel error at the end of a six-degree scan pair, assuming that half the error came from each scan, amounts to a twist angle error of 0.03 degrees, whereas twist-angle variation is typically several tenths of a degree fullwidth. In any case, it appears that the error originates with the telescope and is therefore not something that 2MAPPS needs to worry about, so unless further evidence for concern is encountered, this phenomenon will not be tracked further at IPAC.

  3. IRTS Use of 2MASS Computers --- IRTS is entering a stage of its data processing in which it will need about 300 hours of CPU time, and nebula (H. McCallon's machine) is unable to provide all that is needed. It has been decided to mount the IRTS disks to karloff and run the IRTS software on that machine. This may result in occasional periods of CPU drain, but it should not produce significant problems for 2MASS, and if it turns out to be a significant impact on the 2MASS analysis work, a different machine configuration for IRTS will be worked out.

  4. Survey Strategy Meeting --- The upcoming survey strategy meeting at UMASS is currently scheduled for June 19-20. T. Chester, R. Cutri, and H. McCallon will attend this meeting. One of the goals is to define the general formal data structures that will be involved in database access linkage (e.g., from sky position to scans covering that position) that will be implemented in 2MAPPS. This is needed by midsummer if later software retrofits are to be avoided. J. Fowler will send a memo by the end of the week to T. Chester defining the 2MAPPS needs; this memo will be reviewed by H. McCallon and G. Kopan.

  5. Position-Dependent Responsivity Study --- R. Cutri showed the first results of the study of responsivity as a function of position in the focal plane. This is the expansion of the idea previously embodied in J. Fowler's ROUND program, which used ordinary scan data to derive in-scan responsivity variations but had no way to determine cross-scan variations. The new study uses the specially designed set of raster scans to probe variation in both directions. So far only one point source has been processed, but already some structure in the responsivity variations is evident, including some puzzling scan-to-scan variations that produce a washboard-like pattern. A more complete analysis using more point sources should be available by next week.

  6. Effect of Seeing on Hybrid Photometric Algorithm --- As discussed in last week's minutes, the hybrid photometric technique (involving a magnitude-dependent combination of PSF fitting and aperture photometry) has some problems when the seeing gets bad, because of the loss of flux falling outside the aperture. C. Beichman reported that the approach to be used to get around this in the upcoming data-analysis run will be to jettison the aperture photometry when the seeing goes beyond a threshold. The seeing is measured by the pfrac parameter from G. Kopan's bumps program; the threshold will be set before June 1.