Automated Astronomical Site-Testing Observatory Working Group 6th Meeting Friday, 26 July 1996 at 10am at UNSW in Sydney. /Note: these minutes are preliminary, and do not yet include the discussions on the deployment schedule./ After a tour of the UNSW labs and workshop, the meeting convened at 11am. Attendees Michael Ashley (UNSW), Max Boccas (UNSW), Michael Burton (UNSW), Graham Davies (SEA/UNSW), Jack Doolittle (Lockheed Martin), Peter Hall (ATNF), Gary Hovey (MSSSO), Julia Hu (MSSSO), Martin Mulligan (MSSSO), Andre Phillips (UNSW), Jack Sayers (Australian Antarctic Division), Antony Schinckel (UNSW), Craig Smith (ADFA/UNSW), and John Storey (UNSW). Jack Sayers was welcomed as an observer from the Australian Antarctic Division. Apologies were received from Mike Dopita (MSSSO), whose schedule was incompatible with the revised meeting date. NISM/MISM update Boccas gave an update on the current NISM/MISM design. Most engineering details have now been finalised, and the UNSW workshop has begun fabrication. For the latest description of the NISM and MISM, see the AASTO home page (http://www.phys.unsw.edu.au/~mcba/aasto.html). Still to be resolved are what sort of position sensors to use for the NISM/MISM rotation and the CVF index: the possibilities include hall effect sensors (reliable, but stop working at low temperatures), microswitches (can work down to -90C, but we would prefer a solution without moving parts), and optical sensors (integrated units are not rated for low temperatures, but a custom-made system should work). The greatest unknown is the amount of microphonics expected from the Stirling-cycle coolers (and chopper motors). Storey described the proposed detection system, which should reject most microphonics. Davies wondered whether it would be possible to turn off the cooler for a few seconds while data is acquired. GMOUNT Hovey described progress on the GMOUNT. The printed-circuit layout has been completed for the temperature/voltage monitoring card. Design is progressing well on the axis-controller card (incorporating the inductosyn R-to-D converters and the DC motor drivers). The PC-104 computer that MSSSO has bought may not be the one going to the Pole since Real Time Devices (the manufacturer) is about to release a new version that is qualified for running the QNX operating system. Power supply issues have not yet been resolved. The current suggestion is for the GMOUNT to have its own battery, and to provide power for the ADIMM (now delayed until next year) and AFOS. Hovey requested that the critical design review for the GMOUNT be held in the week beginning September 9 at MSSSO, in conjunction with the next (7th) AASTOWG meeting. /It was agreed that this would be held over two days in Canberra./ MSSSO is experiencing some problems procuring parts. The inductosyns (11-inch diameter, 2mV output signal), preamps (tested by Farand International to -75C), R-to-D converters, and motors have all arrived. Gear design for the axes is underway now; two-thirds of the azimuth axis design has been completed. Hovey is proposing to flush the GMOUNT with a continuous supply of dry N2, to avoid ice accumulating in sensitive areas (e.g., bearings). Flat teflon washers will be used to provide seals. Hall noted that Balder make a cable specifically for carrying the signals and excitation voltages for inductosyns. The GMOUNT is expected to require about 30W during operation (although this figure is uncertain pending measurements). The most power-hungry components are the R-to-D converters for the inductosyns, which require two lots of 0.8W per axis. The DC torque motors only need 1.7W for full torque (20Nm - with an allowance of 5Nm for friction, 15Nm should reach the axes). MSSSO have designed very frugal bus interface logic for the PC-104 board. Storey wondered whether the AFOS would need a wind-sail on the opposite GMOUNT port in order to balance its wind loading. The GMOUNT is expected to deliver 30 arcseconds blind pointing accuracy, and 3 to 4 arcseconds after corrections have been applied (for inductosyn irregularities, orthogonality of axes, and flexure). Orientation of the GMOUNT will require two objects to be acquired and centered. MSSSO was suggesting a finder telescope on the port opposite AFOS. After discussion, it was proposed that the finder could be mounted on the AFOS itself. The finder is expected to be a refractor (60mm?), with an R filter (to reduce the sky brightness) and a small CCD camera (ST4 from SBIG? Davies later showed a tiny surveillance camera, available from Oatley Electronics, which may be suitable). /MSSSO will provide the finder telescope./ Tiltmeters were discussed as possible aids to orientating the GMOUNT and monitoring the GTOWER deflection during the year. Tiltmeters have 0.02 arcsecond resolution over small angles, and can be obtained in +/-60 degree versions (with resolution degraded to a few arcseconds). They can operate at very low temperatures (-100C). Hall mentioned the important of knowing the local gravitational deviation. Hovey commented that the centroiding accuracy from the AFOS would have to be of the order of 1 arcsecond if it was going to provide data suitable for determining high-order GMOUNT corrections. Any lesser accuracy would result in unstable solutions. Doolitte asked about the GMOUNT lubrication, and suggested that MSSSO contact Bill Trabucco at Stanford (@nova.stanford.edu) who has had success running an electric-field mill for extended periods in Antarctica. Hovey is expecting to use Dow Corning 321. Bearings will have their existing lubrication removed, will then be ultrasonically cleaned (after first disassembling them - Storey warned that precision bearings are destroyed by ultrasonic cleaning, unless disassembled), and then be sprayed with a mixture of molybdenum disulfide grease with teflon microbeads. AFOS Boccas showed the progress on the AFOS design (see the AASTO home page for details). The telescope tube (58cm swing radius; c.f. 60cm max for the GMOUNT) will be made of thin-walled Invar, welded on two seams, with a box-shaped central section for mounting to the GMOUNT. The 12-inch f/3.5 primary is being made from astrositall (Russian equivalent of zerodur, a very low expansion coefficient glass). The current proposal is to minimise the number of adjustments possible with AFOS (to zero?), and gluing the optics into precision-machined holders. The tube will be pressurised to about 1psi (and will vary by about +/-1 or 2psi as the ambient temperature/pressure varies), with a bidirectional blow-off valve to prevent explosive disassembly. AFOS will have a spring-loaded lens-cap which is activated by pointing towards the zenith. The details of this design need to be finalised. Storey described recent progress finding chemical desiccants that will aggressively soak up any water molecules in the AFOS. The critical parameter is the vapour pressure of the water above the desiccant when compared with the vapour pressure of ice at the temperature of interest. The difficulty is that the latter goes down precipitously as the temperature drops, and at -80C is of the order of 1000 times lower than the best obtainable dry nitrogen. Silica gel is about 1000 times too poor for this application. Two alteratives are calcium hydride (which converts water into hydrogen gas - possible embrittlement of the Invar? could be absorbed by titanium) and phosphorous pentoxide (converts water into phosphoric acid, which then coats the surface of the chemical, reducing its efficiency). AASTO web page and newsgroup Ashley reminded everyone of the existence of the AASTO web page (http::/www.phys.unsw.edu.au/~mcba/aasto.html) and the newsgroup (news::/www.phys.unsw.edu.au/aasto), and suggested that documents from MSSSO and Lockheed Martin could usefully be added. AASTO update Doolittle described the current state of the AASTO shelter and components. The contract began on 15 August 1995, and Lockheed Martin delivered the bulky components (the AASTO shelter, lifting hardware, support poles) for transport via ship on 4 December. The AASTO arrived at South Pole in February 1996. Some minor cosmetic damage was evident, since protection installed by Lockheed was removed during transport. Anchor plates and footings will need to be manufactured in McMurdo. /Doolittle agreed to organise this./ The AASTO incorporates a number of improvements over the AGO design. For example, a perspex storm-cover is provided over the windows, so that the occupants do not have to wait in the dark for arrival of the aircraft; webbing is provided between the inner and outer skins of the shelter to reduce the chance of delamination due to huge thermal gradients; the thermoelectric generator (due on 10 August from Teledyne) is now decoupled from the thermal control system, to allow easy servicing in the field without interrupting the freon loop (this operation used to take 8-10 hours, and now takes 30 minutes); MOSFET power switches are now being used instead of latching relays for the experiment switching provided by the Data Control Unit (DCU); a GPS receiver is used to supply accurate time; seven IOMEGA Jaz rewrittable optical disk drives (with removable 1 GByte media - supplied by ASA) are being used instead of the original MAXTOR 393Mbyte WORM drive (no longer available); the SCSI interface is now type II (used to be type I); the DCU uses a new computer from Winsystems with improved on-board serial and parallel interfaces (total DCU power has not increased); standard DOS calls are used to access the disk drives, making it easier to change devices in the future. Outstanding jobs to be completed on the AASTO include: purchase of the ARGOS transmitter (435 MHz with a whip antenna and no ground-plane), the propane tank for the stove, guy cables, cushions, fabrication of the exhaust stack, shunt regulator and pyrometer panel, the battery module, and the barometer for the metrological instruments (the wind monitor has arrived). /It was suggested that the wind monitor be placed on top of the GMOUNT, in order to provide wind data to the GMOUNT computer./ ASA is experimenting with a new propane tank design for the AGOs. Rather than using 10 tanks (+1 for nitrogen pressurisation), each one of which holds 100 gallons of propane, weighing 420 lb when full, 275 lb when empty, and is certified to DoT 4BW, ASA have a design using two 500 gallon tanks, 1000 lb when empty, certified to DoT 51. The new design greatly reduces the number of single point failure modes in the fuel system, and makes retrieval of the tanks in the field much easier. There are some uncertainties with transport waivers for the new tanks, and Doolittle recommends that we hold-off buying tanks until these are resolved. In the interim we can use spare tanks from the AGO program, and the existing inventory of propane in Antarctica should be sufficient for next year's operation at South Pole. The power system should be delivered to Lockheed Martin in mid-August, and tested in the lab for 30 days beginning late August. /We should consider purchasing a spare power module in the future./ The University of Maryland has almost completed the Data Acquisition Unit (DAU) - should be complete by the 1st or 2nd week of September. As requested, an additional 32 channels of 12-bit ADC data will be available. Input is via a DB-37 connector - /we should provide an interface panel/. In principle these additional data could be transmitted over a second ARGOS channel, however, we decided not to request two channels at this stage since there may be other telemetry options by the time the AASTO goes into the field. Multiplexing the data into the existing channel is only marginally possible since the channel is already multiplexed, and further multiplexing increases the likelihood of being unable to get a complete sequence out during one satellite pass. /It was agreed that the DCU and DAU should be shipped to UNSW on about 1 October, for testing with the AASTO instruments./ The shipment will occupy one cubic metre. UNSW will require a 28V and 18V DC power supply. The DCU has an RS-232 port (called the GSE port - Ground Support Equipment?), that can control the functions of the DCU (e.g., enable/disable various experiments). It is proposed that the AASTO installation team (Doolittle and ASA employees) arrive at South Pole on 4 November on the first flight. They expect to have the shelter operational at its site (between the skiway and the ASTRO building) within 7 to 11 days. /We will need authorisation for our ARGOS transmitter to operate at South Pole, and should contact ASA to arrange this./ Power supplies Doolittle gave an overview of the AASTO power system. The thermoelectric generator provides 50W of power, clamped at 28V by a shunt regulator (an active element with a current sensing resistor monitored by the DCU). If the current through the shunt drops below a threshold value, a relay opens which increases the fuel flow (by about 50%) in an attempt to restore power by increasing the burner temperature. If the voltage drops below 25V, the DCU will shut off one of the eight experiments, on a randomly selected basis. Experiments will be enabled/disabled to maintain the voltage within limits. The DAU and DCU derive power directly from the 28V AASTO bus. In addition there is an 18V battery, charged from the 28V supply using a Unitrode UC3906 IC (Hall knows of a Texas Instruments battery charger IC that may be preferable). This battery is used to supply power for the disk drives, ARGOS transmitter, TEG spark igniters and magnelatch fuel control valve. The AASTO batteries are Powersonic type PS6120, 6V, 12 AH, made in Japan. These batteries have an excellent history of operation in the field over many years. Ashley then made coffee using a Braemar 240V, 500W dripalator with integral filtration system and 2 litre capacity. Melita Mocha-blend coffee provided a rich and satisfying brew. Hovey has searched for a galvanically isolated DC-DC converter with presettable current limits, but has been unable to find a commercial unit. Davies knows of designers who could do the job. Storey suggested asking Statronics, a local firm. Storey proposed that each instrument have its own battery, charged from the AASTO bus. Each instrument would then be responsible for managing its own power: turning on when the batteries were charged, and running for however long was possible. This approach was agreed to. Software and PC-104 Ashley gave a brief run-down on the state of the UNSW software/PC-104. Two CPUs from DSP (a UK firm) have been purchased: the 25Mhz 486 (2.5W) has arrived, and the 188-based (0.3W) board should arrive any day. The PC-104 computer has been fired up, and works, although there are a few problems. Two alternative ADC boards have been ordered. RT-Kernel 4.5 (real-time extensions for DOS) has arrived, and Davies reported a high degree of satisfaction with it. There was some discussion about communication between the AFOS and the GMOUNT. This will be via an RS-422 serial link. The AFOS will decide where to go next, and can give calibration instructions to the GMOUNT. Davies described ACL (Astronomical Command Language), a public-domain control language written in C, designed for telescopes, currently in use at several sites. Hovey is going to investigate its applicability for the GMOUNT. Deployment schedule Ashley has communicated a preliminary schedule to Bob Pernic for deployment of the instruments at the South Pole. /To be completed.../ Next meeting The next meeting was scheduled for a date to be arranged in the week beginning 9 September 1996, at MSSSO in Canberra. AASTO / Michael Ashley / mcba@newt.phys.unsw.edu.au