Minutes, CA-forum #16
date: monday, june 24, 1996. 11: - 12:30
1. Tied array - hardware update (wew)
The hardware was tested during the week june 17-21, and passed its tests. This tested the production model. Some further packaging work remains.
The VLBI experiment of the past week used this hardware, and the data will be used to define the labelling - ie, which channel is to be called Left, and which Right. This is information is needed before the modifications to CACAL are undertaken. Those changes will also need to check the spectrum inversion state, as it changes from one frequency band to another.
SCHED currently issues a warning if the schedule points to a TDA configuration file -- R.Wark will attend to this.
2. antenna 4 polarisation leakage -- (rs)
This problem (a 2 MHz ripple in the polarisation leakage term on all baselines I involving antenna 4) evaporated during recent tests. (See below (appendix C) for a brief description of the problem).
Last week the RF cables between the dewar and the conversion chain were crossed over (in antenna 4). It had been exepected that the problem would move to the "y" channel. But it vanished, and remained undetectable when the cables were restored to the original configuration.
Ron Beresford found that the connectors were dirty, which may point to the problem.
Meanwhile, the plan is to include these leakage tests in the DQO runs (about 1 hour at 8 MHz is required). Mal Sinclair will look to having the other antenna cables checked.
3. DUTC ... (wew/mw)
Warwick has suggested that the DUTC distribution be consolidated to a single operation - at the site clock. The delay control computer will extract DUTC from the time frame, and have it distributed to all the site computers (in particular, NOEL).
The same machinery should be usable at Parkes and Mopra as well, although some additional work will be needed.
W.wilson and M.Wieringa will look after this.
4. Reference pointing (djm)
Reference pointing was given a serious trial by Lucyna Kedziora and Dave Jauncey - as part of their experiment looking for variablity in a small number of strong sources. They checked the pointing every half-hour. The results are shown in figure 1. The bottom two plots show, on the left the actual pointing error, and on the right the pointing error that would have been present had reference pointing not been used. With reference pointing the mean is zero, and the rms is 5 asec/axis; without reference pointing the means are 10 asec/axis.
The upper three rows of plots show what the reference pointing machinery found (as a function of elevation and time), and the pointing error corrections used.
The results are very encouraging - the machinery works most of the time; on two occasions an incorrect correction was applied. The reasons are understood, and are now trapped. Further experience is needed, and observers are encouraged to give the system a trial.
The support machinery is quite user-friendly - the observers had no trouble using it; what is needed now is experience to guide the operational advice (flux limits; optimum schedule choices).
see also reference pointing guide
5. SAC - R.S's recent problems with attenuators
Bob found in a recent mosaicing experiment that the attenuators never settled down. It is thought that this problem has been around for a few months. Mark Wieringa will investigate.
6. DQO - status JLH/RW
see the web page: dqo status This has the results up to 6 june (inclusive).
- The blank field average
The images are available at the same URL dqo status.
At present the rms noise is still high - 0.13 mJy where 0.03 mJy are expected. (Note: this is the average over 8 DQO observations, with a useful range of configurations. The totla integration time, after all the flagging, is 12 hours).
7. focus - (DJM)
RJB report (see attachment A).
Cost to observers (MK) (see attachment B)
Software status (DJM)
- A test task is available, able to control the subreflector remotely.
- A variety of monitor points are available, but not all are reliable (see Ron's report).
- It would be useful if the subreflector positions were checked - they are thought to be at the 1990-defined X-band optimised location.
- Tests on focus setting/elevation can wait till some automatic control is available.
8. phase switching .. (RS).
Bob will trial a software phase-switching scheme (a refinement of the "de-birdie" scheme).
The motivation comes from the circular polarisation experiment, where low level problems - thought to be related to small sampler offsets - produce artefacts at the tracking centre.
The "de-phaseing" will be done in atlod (miriad).
The scheme will be trialled in late july.
Follow-up :
precision of stored positions -(wew/mjk) - installed; tested.
Appendix A. Narrabri focus machinery. (RJB).
- The hardware has been rebuilt, and now complete, and tested.
- Full manual control is available at the antenna, at the ELVEE rack in the pedestal room..
- Computer monitor control and monitor is via the dataset in the ELVEE rack and the ACC. There are still some wiring problems here, requiring perhaps 1/2 day/antenna.
- The LVDTs and read-outs have been set so that "0" is in the middle of the mechanical range (+/- 1 inch).
- The subreflectors are now set to the nominal X-band optimum position. (1990 data).
- recommendations:
- attend to the wiring problems over the next few months as time permits.
- check the calibrations, to ensure that the X-band setting really is optimum.
- provide some permanent mark on the focus structure - possibly at the LVDT=0 mark.
- provide remote computer control over the subreflector.
Appendix B. The cost of a focus error (mjk)
The loss in efficiency is roughly F*(offset/lambda)^2 dB .... (F is ~6 (+/- 2)).
E.G. : at 8.6 GHz, the gain has dropped to 95% for an 8mm displacement from the optimum axial focus setting.
The difference between 8.6 GHz and 4.6 GHz focus positions is 12.mm.
at 4.6 GHz. We lose ~3 % if we equalise the losses.
The 1990 results were confused on the 1.4 GHz to 4.6 GHz focus positions; the loss at 1.4 GHz is probably about 2% if we operate at some C/X compromise; the 2.6 GHz loss is more like 10%.
Appendix C. The polarisation leakage problem.
The cross-polarisation product (XY), at 8 MHz bandwidth, at 1.4 Ghz, have shown a significant ripple over the band - with a periodicity of 2 MHz. figure 2.
The effect was first seen only on baselines which involved antenna 5 (in october 1993); since then it has been seen in antenna 4 baselines (only). The effect is confined to L-band (~1.4 GHz).
The effect is seen only in the XY products, and not in the YX products.
The switch antenna 5 -> 4 is not associated with a receiver change - ie, two different receivers are involved.
The effect was only seen once in antenna 5; it was seen consistently in antenna 4 until last week, when it vanished.