Previous: Appendix 1: Station Reports Up: EVN TWG/Friends Meeting Next: Appendix 3: EVN CAL Observations
E-mail from Craig Walker (NRAO) to Paul Burgess -------------------------------------------------
Note that I intend to be in Jodrell for 6 months next year so that will be a good opportunity for direct talks and action.
The correlator is still being tested. "First science" processing is well advanced and the initial images are expected very soon. "Second science" processing is starting both to help find bugs and to gain some operational experience. At some point, the debugging should reach the point where we trust the results from experiments recorded in some modes (not clear which yet) enough to begin recycling tapes. Until then, we have a tape shortage. Given the experience so far, I would not like to predict when such operation will begin. It could be very soon, but I've thought that for a long time.
(2) MK3 vs MK4 and VLBA tape formats
MK3 format has been supported as well as possible with 8 BBCs at the sites for a long time. It was expected to be supported at the correlator soon, but it was recently found to not be supported by the software that finds the head offsets so there will be some additional delay. However, we expect to be able to process MK3 as the correlator becomes operational.
To our correlator, MK4 should look very similar to MK3 with some VLBA enhancements when the MK4 is recorded in VLBA compatible modes. Of course MK4 has modes that are not VLBA compatible (more channels and higher bit rates per channel, for example) which could cause trouble.
There is one problem with VLBA/Mark III experiments that we are experiencing that you might want to discuss. There is no good mode for polarization experiments using both Mark III and VLBA terminals. This is because either mode C or A must be used on Mark III (B and E cannot use both polarizations) and both of those require more BBC's than are on the VLBA antennas. What is needed is a Mark III dual polarization mode that uses 8 or fewer VC's. This would probably require recording upper and lower sidebands in the same head group.
(3) Thin tape useage
We are using only thin tape on pure VLBA experiments and thick tape in Network observations. We are not yet using the thin tape at the final high density - we are using the normal Mark III density so far. The problems with the durability of the thin tapes seem to have been solved, as long as the tape drives are properly modified and the tape path is kept clean.
Our intention is to vary the vacuum when switching between thick and thin tape on the correlator. However there have been hints from Haystack that this might not work when using the high density. If not, we are in trouble and will have to come up with some alternate solution.
(4) Scheduling: Status of 'Observe' generating software
The new Observe program is taking much longer to write than expected. I wouldn't count on much before the end of the year. For now, we use SCHED. For Mark III projects, the PI's use PC_SCHED and we transelate the $SKD output file to SCHED input. The geodesy group (NASA) uses DRUG to create VLBA control files directly, but we prefer that they remain the only ones who do this for now.
There has been discussion in the past over whether to send the actual telescope control files to the antennas or have those generated locally from a more global file. In Mark III terminology, this would be equivalent to asking where DRUG is run. Many of us favor having this step done at the antennas so that the latest proceedures and hardware information can be used. Wes (he is writing Observe) has a file format that he uses and the $SKD format (Standard Schedule File Format from SKED) should also be available. The final scheme has not been determined yet.
The MK4 field system should be adequate, with the possible addition of a DRUG like program that is run at the telescope.
(5) Calibration: Where may we find some definitions of the format Cal. date should be in?
I will take your question to refer to monitor data rather than just calibration data. We need some sort of monitor data in order to set up processing jobs. Calibration data is not really needed until after correlation.
All of the necessary monitor information can be passed in Mark III logs as long as everything is present and certain standards are adhered to regarding format and content (the proceedures make it possible for such standards to be violated). Jon Romney is working on a document to specify the information that we need. A quick look at a Medicina Mark III log by me some time ago suggested that essentially everything we need was present so there may not be much change needed.
Alternate paths for delivery of monitor data could be used but would cause use more trouble, at least for the information needed for correlation (of course, VLBA style logs would be ok). Calibration data delivered in ANCAL(AIPS)/CAL(Caltech) format can be used without much trouble if that is easier, but that requires that the observer deal with the files rather than just get the information with his/her FITS file from the correlator.
The one area where a significant change would be very helpful is in pulse cal detection. This is done at the sites on the VLBA. Pulse cal detection on the correlator will be either awkward or time consuming so if it is at all possible to add on-site pulse cal detection at the EVN stations, it would be a great help.
Craig Walker