Single Dish forum #4 - minutes
Date: Tuesday, September 17, at 11:00 - 12:30
- Multi-channel digital synch demods/samplers
- Bonn polarimeter evolution
- VSOP at mopra
- Appendix A - Polarimetry at Parkes
Multi-Channel Digital Synch. Demods/Samplers (WEW)
The correlator group is bringing on-line a set of digital synchronous demodulators (DSD) as part of the Tsys measurement machinery. The new suggestion is that these DSDs be made available for experiments other than the spectral-line (correlator) experiments - for example, they could serve as the general purpose Analogue-to-Digital units for continuum observations.
Four DSDs are housed in the one module; the communication with the outside world is via a dataset; a timing signal needs to be provided. The original design had the correlator talking to the dataset through a serial line from a terminal server on the net. The suggestion is to have a pSOS PC communicating with the dataset; the PC itself is on the net, and can be accessed from any process. An event generator in the PC will provide the timing signals (in general, locked to the correlator).
This scheme would simplify the task of the receiver group in their construction of the LO conversion rack: they would need only to provide analogue synch. demods for stand-alone operation.
The pSOS PC is very similar to the units already provided by the correlator group - the development is modest - a target date of the end of the year is proposed.
This unit is relevant to the next topic - the Bonn polarimeter, as it provides the machinery to free the polarimeter from the VMS system.
(Appendix A presents a summary of the wider issue of polarimetry at parkes).
The Bonn polarimeter is probably the system of choice for continuum polarimetry; its status with respect to the proposed conversion chain upgrades, as well as its relation to the planned computer upgrades needs to be clarified. Some background is provided in attachments A and B. To summarise:
- The unit provides full stokes from two (500 MHZ) Ifs centred around 300 Mhz.
- The unit is currently placed in the focus cabin; its analogue outputs are translated through Voltage-to-frequency units to provide a robust and reliable transport down to the control room.
- The calibration of the unit would be compromised if the bandpasses (amplitude and phase) were not stable.
Should the polarimeter be moved from the focus cabin ?
If the polarimeter remains in the focus cabin then we need to resolve the problem of providing access to the unit from all the receivers in the cabin - this requires additional LO conversion, to bring every received bandpass down to the 50-550 Mhz required. Since it is likely that not all experiments will be able to use the full 500 MHz bandwidth, some additional (switchable) filter system will also be needed. Access to the unit is of course more restricted - the antenna needs to be stowed if the unit requires any attention.
The cost of moving the unit downstairs could be high: the cable stability needs to be assessed; the cable passband is a difficulty: we need to translate every IF to a bandcentre of 300 Mhz in the focus cabin in order to ensure that all Ifs suffer the same bandshape problems, and can all use the same gain compensation algorithm.
How does the polarimeter connect to the computer system ?
In the present arrangement the polarimeter output (from the Voltage-to-Frequency units) is processed in an 11/73, and then transferred to the Vax. The gated counters + pSOS PC provides an alternative access path for the data.
- The VLBI PC - where are its boundaries ?
- Link to the antenna; to the S-2; the samplers and friends;
- The conversion chain; the receivers and attenuators.
- The circular polarisation machinery - is the correlator involved? Single dish CACAL?
The Mopra block diagram shows John's current plans for Mopra. The VLBI PC is a standard unit (largely programmed by nasa) which will supervise the data acquisition aspects of the operation. It can additionally supervise the antenna and LO operations, although this is deemed a matter of low priority as the current belief is that the scheduling will not impose many frequency changes. Two questions were discussed:
- How do we convert linear polarised IFs to circular ploarisation ?
Solution #1 is to use a hybrid - this is the receiver-specific solution. In its favour is the fact that hybrids have already been produced for several receivers. Each hybrid costs approximately $100 (hardware alone; manpower adds significantly to this), and covers a large fractional bandwidth (Each unit will cover about an octave bandwidth, but tuning to the precise band centre could be beneficial).
Solution #2 uses a mopra variant of the Narrabri Tied-array machinery, operating on the sampled data. This solution is frequency independent, and has the additional virtue of providing further software controlable gain and phase adjustments. The unit will cost $K2-3, and take about 6 weeks to complete.
Since the actual costs are comparable the Tied Array solution looks the more attractive.
- software meeds
- Control of the AT Data Acquisition System (DAS) .. Simone Magri
- Control of the S2 .. nasa/SGL/ISTS
- Control/monitor of the telescope .. (Setiobs ? mjk)
- Control of the conversion chain .. (Setiobs ? .. mjk; or VLBI observer)
- Tsys monitoring ?
- Phase calibration for the linear to circular (MOCAL .. variant of CACAL ?)
Issues for future SD forums
- multi-beam - s/w support
- aips++ and SD - offline processing; SDCalc, et al
- Spectra replacement
- Spectral calibration at mopra.
- Antenna control - ("telgo" via socket) .. vlbi; pulsars; multi_beam.
- Antenna monitor - ditto.
- DUTC - parkes, mopra.
Appendix A - Polarimetry at Parkes
The essential problem is to measure the four stokes parameters. For some experiments one can relax the scope and measure just the linear polarisation, or just the circular polarisation. The following summarises the techniques which are (or were) available at parkes.
Full stokes solutions
Almost all the receivers installed nowdays are two-channel receivers with either orthogonal linears or left and right hand circular polarisation. The two IF total powers together with the (complex) cross-correlated signal provide the information needed. The Bonn polarimeter is the recommended solution for continuum experiments (ie, bandwidths down to 100 Mhz). The correlator, with recirculation, is the recommended solution for bandwidths below 100 MHz; its particular strength, of course, is for narrow-band spectral experiments.
- Bonn Polarimeter
This unit has an input bandwidth of 500 MHz; it has two total power detectors, and two analogue multipliers providing the SIN and COS cross-products. The four outputs are derived from Voltage-to-Frequency units. (Calibration requirements - determining the zero-points of the V/F - complicates the counter machinery. There is phase switching on the SIN and COS lines, and each total power output uses two V/Fs. The bottom line is that six gated counters are needed). The unit is currently installed in the focus cabin.
- Correlator with recirculation.
Table A shows the options. The recirculation machinery has been built and tested, but has not yet been used in anger, and a modest amount of work (1-2 weeks) debugging and building the configuration files is required. The narrow-band filters are currently in production; these are unlikely to be routinely available before march 1997. A 2 MHz mode would be possible with 1-2 weeks advance warning (to build one filter). Narrower bandwidths could be made available provided there were no conflict with the VLBI requirements.
Table A - Bandwidth/channel options BW (MHz) channels recirculation factor 4. 4000 1 2. 8000 2 1. 16000 4 0.5 32000 8
Reduced measurement set techniques
Rotating the receiver platform is no longer a viable option as most receivers are mounted at some distance from the rotator centre; this excludes the traditional linear polarisation technique.
It is still possible to determine the circular polarisation, provided that the receiver IFs are set to circular polarisation - in general this means that a hybrid will have been provided to convert the linearly polarised receiver to circular.
Attachment A: Evolution of the bonn polarimeter
Discussion note - a possible evolutionary path for the Bonn Polarimeter.
The essential points are:
- It provides a flexible interface to the processing machinery --- ie, it is independent of the processing platform, and could link to any of Wirrun, yowie, yeti or the suns.
- It is based on hardware developed for the correlator synch. demods, so the additional hardware and software effort should be modest.
The suggestion:
A. Replace the counter machinery.
A suggested implementation of the correlator synch. demods (see - Digital Synch Demods ) places the hardware in a control room rack (to be near the IFs), with a pSOS PC to control the operation. The PC has an event generator to provide the timing signals, and a bus to the DSD to retrieve the counts. The PC is on the network.
- In correlator mode the correlator talks to the PC.
- In continuum mode the observing task talks to the PC.
We propose a second unit, containing just the gated counters of a DSD to serve as the back-end of the Bonn polarimeter. Ie, it would share the pSOS machine and the event generator.
The processing of the data would need to be revisited -- it could probably be handled within the pSOS PC, but that may not be the best path to follow.
B. Further options.
Should we move the polarimeter downstairs? Warwick points out that pass-band compensators have been developed for the multibeam, and these should allow the IF to be transported as analogue signals from the focus cabin.
Whichever choice is made, we should ensure that the decision is consistent with the new conversion chain and RF switching machinery that the receiver group are developing.
Should we dig deeper into the unit, and replace the V-to-F as well? ie, place the synch demods in the polarimeter? This is an option wherever the polarimeter is located, but is clearly easiest if the unit is downstairs.
Comments - R.Haynes.
One of the most important points to consider in the discussion of the Polarimeter is the effective bandwidth provided by placing the temperature controlled phase switch unit in the Focus cabin. At the moment we can achieve 500MHz of instantaneous bandwidth on the higher frequency receivers. This must be maintained. At lower frequencies the limitation becomes interference free bandwidth capacity at the observing frequency. However non-thermal emission tends to rise with decreasing frequency. Thus smaller bandwidth at lower observing frequencies works just fine.
The broad bandwidth at 6 and 3cm is fundamental. I cannot emphasise this enough. Any discussion in the direction of bringing the polarimeter downstairs must address this need. I would be surprised in one can achieve this sort of instantaneous bandwidth across co-ax cables even with frequency compensation.
Your comments on the transfer of the signals from the focus cabin to the control room seem fine. Whether you use V-F's and bring frequencies down which are then sampled with appropriate demodulation to 'dig' out the cal and signal from each of the two TP and Q and U channels or whether you digitise in the focus cabin and bring pre-processed channel data down doesn't matter.
Andrew and my original idea of going the V-F way avoided as much as possible problems of co-ax cable bandwidth and two many digital signals and mixing signals flying around up in the environment of the focus cabin.
The polarimeter system as we have it at the moment is a remarkably stable instrument with very low gain and temperature drift problems. John Murray's stabilisation of the polarimeter, and the approach we used to switch both total power and Q and U signals makes the whole system very stable.
Whatever system of signal transmission and demodulation we go over to, we must maintain and gain and drift stability characterisation we have at the moment.
If we were considering an enhancement in the system then I would encourage the buidling of a second polarimeter. It has always been a frustration to me that we could not simultaneously observe two bands in full polarimeter mode. In fact Andrew and I built into the transmission system the capability of bringing another set of 6 signals down (4xTP - switched + Q & U) in case we could ever develope a second polarimeter.
There are excellent reasons for having a dual polarimeter system at Parkes which would allow the analysis of two channels - eg for rotation measure & spectral studies of objects.
The use this last weekend of the Bonn Polarimeter system again showed us just how good this system is. It produced beautiful results again (this time at 20cm).
Next meeting: Tuesday, November 19, at 11:00