next up previous contents
Next: Scientific highlights Up: No Title Previous: Troubleshooting and known bugs

Things that need doing

Mechanical things

Motor cables:
The method of connecting the motors to their cables is fairly flakey and prone to coming loose.

New incremental encoder:
The new incremental encoder for the RA axis doesn't work when slewed at moderate to high rates. The problem is possibly in the capacitor circuit for generating the rate and direction outputs.

Filter changer:
A robotic filter changer, Auto-Col, is currently being designed and built. This is a major challenge.

Lens cover:
We need to automate the lens cover, and provide a heating element (a light bulb?) to keep the lens warm (to stop dewing) after the cover is closed.

Computer-controlled roof:
The roof must be under computer control, with appropriate safety interlocks to prevent damage to the telescope.

Lens dewing:
The lens has often been covered in dew. One possible solution is a couple of fans blowing air on the lens continuously. These fans have now been installed, and it waits to be seen how effective they are.

Flat-fielding lamp:
It would be convenient to have a flat-fielding lamp available so that we can determine the pixel-to-pixel variations of the CCD without needing twilight frames.

Lens cleaning:
The lens probably needs cleaning once every could of years (less if we can stop dewing from occuring). The recommended solutions are lauryl alcohol, isopropyl alcohol, and acetone, with application using lint-free cotton swabs. The AAO recommend a drop of Kodak Photoflow as a wetting agent. Please do not attempt this job without speaking with Michael Ashley first.

Mirror alignment widgets:
Balancing the main mirror in its cell (so that the collimation remains precise regardless of zenith distance), is a very tedious task, that would be made much easier if a purpose built jig was constructed.

Absolute encoder problems:
The Dec absolute encoder is not properly held on its shaft, with the result that it occasionally shifts, which in turn prevents the telescope from finding the declination knife-edge.

Roof crane maintenance:
The roof crane sometimes jams after long periods of disuse. A slight tap with a hammer on one of its bearings will get it going again. Perhaps it needs some preventative maintenance, or regular movement, to keep it running...

Limit switches:
The APT's limit switches have been a cause of many problems.
  • The inner and outer limit switch positions need to be checked carefully and readjusted. In particular, the Dec south outer limit appears to be in the wrong place.

Friction disk:
The friction disk needs checking for damage, and reoiling. It may be desirable to install wipers to stop debris getting between the disk and the rollers.

Roller pressure:
It would be nice if there was a more accurate way of determining the pressure applied to the roller, so that it can be set with confidence at the beginning of each night.

Sealing the tube:
The original Baker-Nunn camera was designed to have a sealed tube, pressurised with dry nitrogen, to keep dust and condensation out. As soon as the filter wheel is finished, we should place some bags of silica gel in the tube (to absorb moisture), seal the tube, and fill it with dry nitrogen.

Motorised tilt/rotation:
There are three micrometers on the CCD mounting stage that are not automated. Adjusting them by hand is extremely tedious, so it would be nice to have computer control of them.

Accurate collimation:
There is one distance in the optical path of the telescope that hasn't been collimated: it is the distance between the corrector lens and the mirror. We should perform a series of focus tests for various values of this spacing, and see if we can get better performance. It may also be possible to change the distance between the filter and the field-flattener to increase performance. These ideas could be tested using the computer model of the APT optics. The collimation tests should wait until the IRAF autofocussing routine is complete, since we will then be able to obtain quantitative measurements of the performance.

Electrical things

Limit switch wiring:
The limit switch wiring is complex and hard to debug. It would be preferable to have a panel of LEDs showing which limit switch was hit, so that faults could be tracked down more easily.

General telescope wiring:
The recent re-wiring of the telescope has two problems: (1) many of the wires aren't long enough, so that stretching occurs as the telescope moves around the sky, and (2) the quality of some of the wire used was not very high (e.g., the limit switches are connected with twin-ax cable that can break internally without any evidence being visible on the plastic insulation). We need to improve the wiring to the mirror heater so that the cable can't become jammed as the telescope moves.

Roof crane wiring:
The roof crane and roof motors can not be used simultaneously. It would be nice if they could be.

Servo amplifiers:
We need the ability to drive the servo amplifiers manually, so that we can slew the telescope if the servo controller fails. The manual control could take the form of a potentiometer providing an adjustable voltage to the servo input, with a spring-return switch used to select manual control.

Spare parts:
We have spare parts for most of the boards now.

Electronics spares:
While we have lots of spare ICs, we have almost no common components such as resistors, capacitors, diodes, and transistors.

Electronics documentation:
We need to have comprehensive documentation for the telescope and its electronics available at SSO. This should include spec sheets for all the integrated circuits.

New control computer:
In the medium term we should look at replacing the Apple II and servo controller with a PC and a modern servo control card. A possible choice would be the cards used by the AAO for controlling the 2dF fibre positioner. This would allow optimal movement between positions.

UPS:
An Uninterruptible Power Supply (UPS) is required to allow the roof to be closed in the event of a power failure.

Lightning protection:

Noise reduction:
With all the various computer fans going in the APT observer's room, it is not particularly pleasant. It would be nice to reduce this noise somehow.

Servo-amplifier cooling fan:
The servo amplifiers need a fan blowing air over their heatsinks, otherwise they overheat. The existing fan is a Papst 110V one that I happened to have lying around, driven by a 240-110V transformer. The fan is not well mounted. This situation needs to be improved.

Software things

Weather information:
For fully automated operation, we need to be able to determine what the weather is like, and to close up if it is cloudy or raining or the humidity is too high. Here are a few possibilities:
  • Simply obtain by computer the dome open/close status of the 2.3m and AAT, and use this to decide whether to open/close the APT.
  • Make our own weather station and interface it to a computer. We already have a rather nice humidity measurer built as a vacation project by Weijian Lu working with Michael Ashley in 1994.
  • Tap into the 2.3m weather station.
  • Tap into the AAT weather station.
  • Use the CCD camera on the APT--it is fairly obvious when it is cloudy.

Precession of coordinates:
The MCCD program should have the capability of precessing coordinates, so that the ``equinox'' command actually does something besides just affecting the FITS headers.

Pointing corrections:
To improve the telescope tracking we could produce a pointing model for the telescope and apply it in software. One would first have to disentangle all the corrections built into Paul Payne's Apple code. Currently, the APT polar axis appears to be misaligned by about 150 arcseconds.

Tracking rate adjustment:
The tracking rate is currently off by about 0.3 seconds of time in 300 seconds.

Position determination:
Automatic position determination from CCD images.

Precision alignment:
Automatic precision alignment of images.

DARKTIME FITS header:
The ``DARKTIME'' FITS variable should include an allowance for the finite readout time of the CCD.

Exposure countdown:
Currently, the PC does not have a display showing how many seconds remain on the current exposure. This would be a nice addition.

Gamma-ray bursters:
Mt Stromlo is possibly getting a satellite ground station so that it can receive immediate notification of gamma ray bursters detected by a satellite. It is possible that an e-mail message could be sent to the APT from Stromlo within seconds of an alert, and the APT could then stop whatever it is doing, slew to the coordinate, and take an image.

Recordable CDs?:
The AAO has a CD recorder in Level 1 of the AAT. This can put 600 Mbytes on a CD, for a cost of something like $35, making it an interesting option for archiving APT data. We could buy own own recorder for about $6000.

Slow-scan camera:
When the APT becomes fully remote, it will need a slow-scan camera so that the remote observer can visually inspect the telescope and building to diagnose faults (e.g., if the telescope hits a limit switch and won't drive back).

Absolute encoder problems:
The absolute encoders are currently only used to determine which way to move the telescope in order to reach the zero-point calibration. The software that does this on the Apple has some bugs, which for certain positions in declination at least, will result in the telescope going the wrong way to find the limit.

Peltier cooler:
The CCD camera is cooled by a Peltier cooler, and has a temperature sensor that can be monitored with a voltmeter at the camera control box. It would be worthwhile to measure the cool-down performance (temperature versus time) for the cooler so that we can monitor it over time for degradation. The only information we have so far is from Wright Instrument's measurements, which show the cooler reaching 200K in 15 minutes from about 25C.

Remote resetting:
To allow remote operation we need to be able to cope with every conceivable hardware/software fault. At the moment, there are several things that can happen that require manual intervention before you can continue:
  • If the Apple II crashes, it can not be automatically reset. Probably the easiest way of fixing this is have the power to the Apple II controlled by a solid-state relay connected to the PC.
  • If the servo-controller crashes, it has to be reset by pushing buttons on the front panel. This could be readily controlled by the PC, or else an automatic-reset on power-up could be installed.
  • If the PC crashes you are also in trouble. The PC power could be controlled by a solid-state relay, either controlled by the workstation, or by an heartbeat circuit.
  • If the telescope hits a limit, the servo controller goes into an infinite loop, and it can be very difficult to drive the telescope off the limit.
  • If the workstation crashes, it can be set to automatically reboot itself. This doesn't always succeed (e.g., filesystems could be corrupted, and may not be able to be automatically repaired), but there is no easy way around this.

PC crashing:
The PC still hangs up occasionally. Also, the communication link between the PC and the workstation sometimes gets caught in an endless loop which can only be reset by rebooting the PC and re-establishing the APT Link window.

On-line data processing:
The DECstation is certainly powerful enough to allow the APT data to be processed on-line. The software is almost ready to go, for the supernova search program at least, but has yet to be installed.

Automatic focussing routine:
An IRAF script is almost finished that will automatically focus the APT.

Twilight frame routines:
Changes to MCCD could improve the ability to automatically acquire twilight frames.

Performance information:
The following information should be provided to observers:
  • Magnitude limit in various filters.
  • Expected sky background as a function of moon phase.
  • Linearity of the CCD.
  • Gain calibration.
  • Read-noise measurement.
  • Reproducibility of flat-fields.
  • Dark current as a function of time.
  • Precise readout times as a function of readout speed and size of the area being read out.

next up previous contents
Next: Scientific highlights Up: No Title Previous: Troubleshooting and known bugs

Michael C. B. Ashley
Wed Jun 5 14:40:19 EST 1996