2hours per
map, this can consume an entire observing shift.
|
Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://www.adass.org/adass/proceedings/adass00/P3-11/
Дата изменения: Mon Jun 4 20:11:11 2001 Дата индексирования: Tue Oct 2 05:56:43 2012 Кодировка: Поисковые слова: р п р п р п р п р п р п р п р п р п р п п р п п р п п р п п р п п р п п р п п р п п р п |
The mapping of the primary reflector surface is achieved by "holographic" techniques. Currently a 94GHz radiation source, located on a nearby Telescope (United Kingdom Infra-Red Telescope) is used to illuminate the JCMT surface. The orientation of the telescope is scanned in two coordinates to measure the beam pattern. From this it is possible to determine the path length of the radiation to, and hence the position of, each panel. A new instrument has been deployed in order to achieve greater accuracy and produce maps at a significantly faster rate. The instrument uses two main frequencies, 80 and 160GHz, and is frequency agile around these frequencies to allow for the elimination of spurious signal paths.
The James Clerk Maxwell telescope (JCMT) primary reflector is made up from 276 reflective panels. Each of these panels is attached to three stepper motors that allow the panels to be individually tilted in order to achieve the best shape for the highest antenna gain (Smith 1998).
To maintain and correct the parabolic shape of the primary reflector, a panel position detection system is employed. A 94GHz radiation source located on the hill above (at the UKIRT) illuminates the JCMT, and its receiver detects this signal. By scanning the JCMT across the beam in a raster pattern whilst moving the secondary mirror up and down, the phase of the signal is recovered. This produces a whole-dish map of the vertical position and tilt of each panel (with 3micron resolution and 12micron repeatability).
Whilst the current system provides the necessary information to maintain the surface of the JCMT, it suffers from several problems. One of the largest problems is that it takes about two hours to fully map the surface. Apart from the time aspect, this introduces errors given the fact that the temperature of the telescope structure changes over the period of the mapping. It is well known that temperature changes have a significant effect on the position of the panels. The only time when the temperature is stable over this sort of time period is the middle of the night, so to produce the best maps, normal observations have to be halted.
When moving panels, it is necessary to re-map the surface to check
that the moves were sufficient and correct, and sometimes the panels
need a further adjustment with a third map required. At 2hours per
map, this can consume an entire observing shift.
Other problems include spurious reflections from the protective Gortex membrane and other structures, as well as a growing maintenance burden from the aging system in place.
A new system described in the JCMT Holography Receiver RXH3 Preliminary Design Paper has been constructed and was commissioned in December 2000. Still undergoing tests prior to calibration, its main features are: (1) a map of the surface can be produced in about 15 minutes, (2) measurement repeatability of 5microns RMS, (3) dual frequency (80 & 160GHz) system (using two frequencies provides a choice of either fast low resolution or slower high resolution mapping, and stepping around these frequencies provides the ability to solve and eliminate signals from spurious paths like reflections from the protective membrane), (4) dual signal system, main and reference beam, which produces both phase and amplitude of the signal, akin to a holographic measurement system and (5) a VME real-time computer data acquisition system for fast mapping and time stamping for accurate registration of the data.
This is a PC with a PC-LabCard multi-IO and real-time clock card. This provides control over the PTS frequency synthesiser via a parallel interface. Frequency stepping patterns are received by the PC via an RS232 interface, and when commanded, the PC runs the given pattern continuously at a 1KHz rate. This is achieved by controlling the PTS synthesiser via a parallel interface.
This digitally controlled frequency synthesiser can produce an output in the range 250MHz to 350MHz, and when coupled to the 80GHz and 160GHz signals generators produces controllable signals in the range 80.25 - 80.35GHz and 160.5 - 160.75GHz. This unit will step in frequency by 5MHz, at a rate of 1msec as per the pre-programmed frequency pattern. It will also blank the signal during each hop in order to synchronise the data acquisition system at JCMT, much like the blanking signal in a television signal.
This is a dual frequency system with 80 and 160GHz channels orthogonally polarised by a wire grid.
This consists of two curved mirrors mounted on the backing structure of the JCMT. Their purpose is to fold the reference beam, which passes through a hole in the JCMT primary reflector, into the reference input of the RxH3 receiver.
The receiver consists of two channels, reference and signal, each with two mixers (80GHz and 160GHz) to produce a 300MHz IF. These two IF signals are then passed through a dual channel correlation receiver and each channel produces a sine and cosine output. The signal blanking is also detected and used to generate a TTL sync pulse to drive the DAU (Data Acquisition Unit).
The DAU is a VME-based 68060 computer with a Wind River's VxWorks operating system and applications coded in the Anglo-Australian Telescope's DRAMA C programming environment). The DAU consists of: (1) a VME XYCOM XY566 16 channel programmable 12 bit A/D converter, (2) a VME XYCOM XY240 TTL I/O card and (3) a VME Bancom BC635VME RIGB card.
The SYSCON (System Controller) DAU and DHS (Data Handling System) components have been written as DRAMA tasks. These tasks and the tcl Control Script communicate with each other using DRAMA messages.
RxH3 will quite literally revolutionise the maintenance of the dish as well as achieve significant improvements in surface accuracy.
The ability to map the surface in 15 to 30 minutes will permit the rapid correction of aberrant panels. It will be possible to make many adjustment iterations within just a few hours, and errors introduced by thermal drift in the current system will be minimal.
A surface accuracy of 5microns RMS will be a significant improvement over that currently achievable and will significantly improve telescope efficiency.
Smith, I. 1998, Proc. SPIE, 3351, 190