The hardware of the Dwingeloo telescope will only be described briefly as the detailed description is beyond the scope of this document.
The operational control of the Dwingeloo telescope is at present done by a HP-1000 computer with an RTE-VI real-time operating system. The details of the operating system are again beyond the scope of this document. See section RTE operating system.
The disk partitions are referred to as cartridges, and certain cartridges have particular uses, see section RTE operating system.
In addition to the disk unit the computer has a 1600bpi HP tape unit (used for schedule input and data output), 3 terminals and a line printer. If these terminals have been out they may need to be reset before use, See section Procedures.
The HP-1000 computer controls and reads all hardware devices via a multiplexer, described in NFRA note 191. This is similar to, but not the same as the Westerbork multiplexer. The driver for the multiplexer is described in NFRA note 184
The correlator hardware sits inside a Faraday cage, to prevent its noise entering the receiver system of the telescope, and to prevent external noise entering the correlator. The correlator has a maximum of 1024channels (at 20MHz bandwidth). The hardware is extensively described in NFRA ITR 188 and its configuration and acquisition software in NFRA note 505. Test software is described in the draft NFRA note 531 and in NFRA note 545.
The correlator can operate in 1bit or 2bit mode, but these need different normalization and van Vleck corrections. There is a monitor channel for the total powers (as measured at the A/D converters) for both input channels (1). The correlator has two data buffers, one for measurement with the noise source off, and the other for with the noise source on.
Data is read out in increasing lag order as 1024 long integers (2). For bandwidths less than 10MHz double spectra are possible.
The correlator hardware has a minimum integration time of 13.1072 msec and a maximum of 53.6 sec; longer integrations have to be done via software.
The correlator can be configured in a number of different observing modes (See section Observing modes) with different bandwidths and either 1 or 2 inputs from IF. Crossbar switches are automatically set up by the correlator hardware. Note also that these are the limitations in the correlator hardware, and that if 2 inputs are needed then 2 IF-to-video converters need to be connected up in the correct way. This is not at present avilable.
Bandwidths are described by `bandwidth codes', which are used in the correlator setup, and are described in the section Schedule format.
Up to 20MHz there is a choice of 2 mixers (35MHz the default and 33.043MHz) and the narrow-band mixing scheme is used. For 40MHz and 20MHz the wide-band mixing scheme must be used. Note that these are the correlator setup, and that the wideband (20 and 40MHz) modes need different IF to video hardware setups and mixing schemes to the narrow band (10MHz or below) modes. See below.
The correlator is setup with various octal codes
by the program OBSRV. These octal codes are explained more fully
in the schedule format See section Schedule format.
There are 3 IF to video units available, one wideband (40MHz or 20MHz) and two narrowband (10MHz and downwards in steps of a factor 2 to 78kHz). These convert the IF signal from the frontend to a video signal which is then sampled by a 2-bit digitizer, where the total power is measured for later use. The correlator can be setup to use only the most significant of the 2 bits, i.e. in a 1-bit mode. There is a chartrecorder attached to two of the IF systems, so that the total power levels in each can be checked. This is most useful to monitor high-level interference and any major receiver malfunctions. In addition to the chart recorder there is a counter attached to one of the IF systems, so that instantaneous readout is also possible. These total power registration systems do not have any blanking for the noise source, so that when it is switched on the total power rises noticeably
The receiver works from 1375 to 1425MHz with a single linear polarization. The feed rotator system does not work. The system temperature is about 35-40K and noise step about 17K (3). With some modifications (power supply and feed) a Westerbork frontend can also be used instead of the usual frontend.
There are 3 IFs and so 3 possible mixing setups. To change IF(s) connected to the correlator you must change cables. In the following formulae `B' is bandwidth and `Fsky' the sky frequency. All formulae assume you want the spectral line at the centre of the band, and all frequencies are in MHz.
Fsynth=(Fsky -B/2 + 150)/18This uses only 1 extra fixed mixer at 150 MHz. This is described with the OBSRV code
SY=-1,150,18
section Schedule format.
Fsynth=(Fsky +B/2 +125)/18This is described with the OBSRV code
SY=1,125,18
section Schedule format.
Fsynth=(Fsky +B/2 +127)/18
The factor "18" comes from a frequency doubler outside the frontend, and the 9x multiplier in the frontend.
The receiver works from 1580 to 1725MHz with System temp about 35K. Dual linear or circular polarizations are available. The usual value for the noise step (nominally "15K") is as follows:
The 18cm receiver is in the same frontend as the 6cm receiver.
There are 3 IFs and so 3 mixing setups. To change IFs to the correlator you must change cables. In the formulae below all frequencies are in MHz, and `B' is the bandwidth and `Fsky' the sky frequency. All formulae assume you want the spectral line at the centre of the band.
Fsynth=(Fsky -B/2 -150)/18This is described with the OBSRV code
SY=-1,-150,18
see section Schedule format.
Fsynth=(Fsky -B/2 -125)/18This is described with OBSRV code
SY=-1,-125,18,
see section Schedule format.
The motors to drive the telescope are located in the centre of the building and are normally under computer control. However it is also possible to remove the connector from the motor control and steer the telescope by hand in case of emergency. There is a grey box in the telescope with thumbwheels to set the telescope speed if the telescope must be moved by hand. When the pointing is under computer control, a pointing correction table is applied giving about 1-2 arcminute accuracy.
The telescope has both mechanical and computer limits to its movement. In elevation the telescope can be set by computer between 2 degrees and 86 degrees, but the mechanical limits are 0 and 90 degrees. In practice very few observations are done at elevations below 10 degrees because of the extreme atmospheric corrections needed and ground radiation pickup at lower elevations, but the dish needs to be pointed outside the computer range in case of storm, or if the frontend needs replacing.
The storm signal will point the telescope to the zenith, from where it cannot be controlled by the software. If the elevation exceeds the computer limit then the telescope dish must be lowered using the manual override system before any schedules can be run.
In azimuth the computer limits are -269 degrees and +269 degrees, but the hardware limits are -300 degrees and +300 degrees, There is an analogue display of this azimuth value in the telescope so that the operator can see if the cable wrap limit is going to be reached. The telescope software will detect if an azimuth limit is going to be crossed, and rewind the telescope.
The 25m diameter antenna dish is made of 7.7mm wire netting, made of 0.8mm stainless-steel wire, and has a surface accuracy of about 1mm. The focal length is 12m, and the prime focus receivers are held by a quadripod. The beams and aperture efficiencies are taken from NFRA Note 275 are given below:
Wavelength Beamwidth Efficiency Gain Notes /cm /arcmin % K/Jy 21 31 64 0.11 18 31 64 0.114 6 10 40 0.071 3 5 12 0.021 3cm system now unavailable
There are 2 hardware clocks in the racks, one of which is the system clock and runs on UT. This is the clock read by the computer over the multiplexer and is used to control the observations. There is also a DCF-77 clock, which picks up a low-frequency radio signal to show MET. MET is the local time in Dwingeloo, and is exactly 1 or 2 hours different to UT depending on the season; in winter the difference is 1 hour, and in summer 2 hours. The setting of the computer clock at boot time is not critical, as the observations are scheduled using the system clock.
To point the telescope correctly the Local Sidereal Time is needed. The derivation of LST from UT is done in software, but the system clock must be correct to within 3 seconds to have the source at the centre of the beam (6). Normally the system clock is kept to within 1 second, and this is checked against the DCF-77 clock. Note however that the hardware clocks do not know the year, and this information is only available either from the schedule, or (by default) from the online system (section Schedule format).