ca-forum #19 minutes

date: Tuesday, October 8, 1996. 11:00 - 11:30

Pointing discontinuities requiring ACC reset (DJM)

This is a serious intermittent problem - every now and then an antenna will suffer a serious pointing offset. The diagnosis is difficult since the problem is still quite infrequent in its onset. At this stage the only aspect which seems well established is that an ACC reboot will cure the problem.

  • The problem has been seen, at some stage, in every ACC
  • The problem is always accompanied by a large increase in the ACC's count of encoder jumps.
  • Most commonly, it seems to have been a sudden onset (Earlier suggestions of a gentle onset have been withdrawn .. close examination shows that in those cases also the onset was abrupt). The current suspicion is that an azimuth error is the commonest form.
  • Derek has recently modified the ACC so as to improve the encoder reporting. On monday (Oct. 7) the problem appeared - all the ACC monitor points were OK - that is, the pointing parameters had not been over-written; the azimuth and elevation were correct.
  • Further changes have been made to get at the NPL-reported encoder values. (The problem here is that to counter the 1.5 degree jumps caused by mis-alignment betwen the fine and coarse encoders the ACC operates in a pseudo-incremental mode, and ignores the coarse readings if the ACC suspects them of being incorrect).

    additional note (DJM) - tuesday afternoon:

    We were afforded a rare opportunity this afternoon when CA01 was caught in flagrante delicto. The ACC reported a jump of -0.352 degrees, which we could compensate for with an identical offset in the EA pointing parameter. The jump "counts" were continuously incrementing during the entire event, but otherwise, all other aspects of ACC performance seemed okay. The jump size corresponds to an error in bit 12 (bit 0 = LSB = 0.31").

The nature of the problem should become clearer in the next few weeks.

Circular Polarisation Calibration (RS)

This item was included at the request of some observers: could we provide a simple unambiguous proof that the sign of stokes V was correct in all the CA bands, at the end of the processing chain. (This is similar to the question raised with respect to the Tied Array machinery).

The proposal which cuts to the core of the issue is to radiate circular polarised signals from the vertex roofs of two antennas - much as the VLBI community do in their calibration. This requires two synthesisers which can lock the the antennas' 5 MHz reference, and a set of helices.

Bob Sault and Graham Baines will look to this.

L-band sensitivity limits - implications for DQO (RS)

The DQO blank field, at L-band had a fairly high noise level, considerably above the thermal noise. (100 uJy/Beam, compared to the 30 uJy/Beam expected). What do we need to do to achieve thermal noise in the image?

Bob showed an L-band continuum image from a recent 6 km. run (48 hrs), which did indeed reduce to near thermal noise levels. The suggestion is that calibration is the issue. Perhaps the DQO protocol should be revised to provide closer ties between the blank field and the primary calibrator (1934 or 0823).

Frequency Switching in mosaic mode (MW)

Recent tests have shown that the digital synch. demods have cured the earlier problems which affected this class of observation - the levels swicth smoothly at the start of each new frequency cycle. A problem remains - the levels change abruptly on the last integration cycle of each frequency setting. Presumably a software matter - to be pursued.

Focus machinery update (DJM)

The necessary changes to ARRAY and to a_test have been made; the hardware is in place in two antennas. Time has been set aside this week for tests.

Ron. beresford reports steady progress on the hardware front - roughly one antenna/maintenance period, as promised.

next meeting : Wednesday, December 4.
Staff space
Public