UNSWIRF: hardware setup

The hardware components of UNSWIRF

UNSWIRF consists of the following pieces of hardware:

• The Queensgate Fabry-Perot. It covers the H and K near-infrared bands, with a 70-mm aperture.
• A Queensgate CS100 controller. This controls the spacing and parallelism of the Fabry-Perot. The same controller is used with the AAO Taurus instrument.
• The precision mounting slide for the Fabry-Perot. This includes a stepper motor to move the Fabry-Perot in and out of the beam, and various position sensors to determine the position of the Fabry-Perot.
• A mounting plate which suspends IRIS from the instrument mounting box, thereby allowing room for the Fabry-Perot and slide (and the Hatfield polarimeter modules).
• A PC computer that controls the CS100 and the Fabry-Perot slide. The computer is currently an NEC Powermate 1 (an old 80286 machine, quite adequate for the job). The PC has a National Instruments DIO-96 (96-bit digital I/O card) and an ethernet card (for communication with a Sun workstation, which in turn communicates with the VAX that runs IRIS). The PC boots from a hard disk, although it can be run purely from the 5.25-inch floppy drive (use the ``UNSWIRF bootable floppy'').
• The ``UNSWIRF Controller''. This contains drive circuitry for the Fabry-Perot slide, and interconnects between the various cables used to link the computer with the hardware.

Setting up the hardware

AAO staff tasks

• f/35 chopping secondary top-end installed (although we do not use the chopper).
• IRIS installed on the UNSWIRF mounting box, orientated so that the polarimetry modules are due north.
• A rack of UNSWIRF-related equipment (PC, CS-100, UNSWIRF controller) installed in the cass cage.
• The rack of equipment should be connected to a no-break 240 VAC power socket.
• The comparison lamp power supply needs to be connected to a 240 VAC outlet that is controllable from the console.
• An ethernet connection (10-base-2, i.e., co-axial cable) installed to the PC.

Observer tasks

1. The following items need to be located (they are usually in the Plate Room on the 6th floor, but may be in the 4th floor instrument room):
   • The Fabry-Perot (in a Radio Spares metal suitcase).
   • The F-P slide (black object with two shiny rails and a motor attached).
   • A short (20-cm) cable which connects the high-voltage socket on the F-P (the multipin one closest to the bottom) with the cable going to the CS-100. This short cable has a silver coloured connector on one end, and a green box-like connector on the other. It is often near or on the Taurus instrument electronics rack.
   • Four assorted comparison lamps and power supply.
2. Ensure that the Hatfield polarimeter slides are out of the way. This means that the two screw threads with attached butterfly nuts must be fully out. This is necessary even if the polarimeter modules are not in place. You may need to remove cables connected to the modules.
3. Install the Fabry-Perot in the slide:
   • The Fabry-Perot has a replacement value of about $70,000 and a delivery delay of 18 months, so please take extreme care with it. The surfaces can be damaged by any contact, and by any solvent including water; even breathing on the optics is potentially destructive. So, when handling the Fabry-Perot you should use rubber gloves and a mask (both are provided in the transport box). Don't worry that you look silly wearing a mask. The perspex covers on the Fabry-Perot should be left on until the last minute (but don't forget to take them off, otherwise your astronomy will be compromised...).
   • Any handling of the F-P must be recorded in the little book kept in the F-P case.
   • After putting on the gloves and mask, remove the F-P from its protective plastic bag. Identify the top and bottom of the F-P (the writing on the side of the etalon give a clue). The bottom perspex cover should be secured by black bolts, and the top one by silver bolts (the difference is important since the black bolts are shorter). If the bolts are the wrong way around, swap them all now.
   • Loosen the top bolts (the silver ones) slightly, so that they can be easily removed using fingers at a later stage.
   • Position the F-P slide on a flat surface with the thin F-P cover plate up (the same orientation it will be when placed in the cass cage).
   • Remove the bottom perspex cover from the F-P. Try to avoid exposing the etalon surface to dust.
   • Place the F-P on the slide, bottom side towards the mounting flange, with the connectors pointing out towards the motor.
   • Insert the three black bolts, loosely.
   • Now, in an attempt to make the F-P position reproduceable, gently push the F-P into the diagonal corner close to the motorised screw thread, but opposite the motor. While doing this, rotate the F-P clockwise (as viewed from above). These two motions will result in a fairly reproduceable position. Now tighten the three black bolts.
   • Inspect the F-P bottom surface with a torch, and note in the F-P book any blemishes that you see.
4. Install the slide in the UNSWIRF box:
   • There are two teflon pads on the slide, that rest on the bottom plate of the mounting box. Place the slide into the mounting box in such a way as not to dislocate these pads (i.e., lift the slide onto the mounting box, don't push it over the lip).
   • Remove the three mounting bolts from their retaining hole in the mounting box, and pass them through the slots into the slide. NOTE that there are two sets of holes that the screws can fit in, you want the ones furthest from the motor. Tighten the bolts finger-tight and then back-off half a turn.
   • Push the slide towards the cass cage door till the mounting bolts reach the ends of the slots. If the bolts do not reach the slots, either the polarimetry module or its cable is in the way, or you have used the wrong mounting holes.
   • Tighten the three mounting bolts.
5. Boot the PC computer:
   • Upon power-up the PC should boot and automatically run the UNSWIRF program (in c:\unswirf\project\unswirf.exe).
   • The PC boot process should show that the PC has located its ethernet card, and that TurboTCP has been successfully installed.
   • The PC monitor takes a while to warm-up. Its image tends to quiver for 30 minutes or so.
   • If the hard disk fails, try using the ``UNSWIRF BOOT'' floppy. (which was created by cd \unswirf\utils; makeboot).
   • To verify the ethernet connection, type ping unswirf on one of the Sun workstations. This should respond with "unswirf is alive". For reference, the PC's IP number is 192.231.166.26.
6. Cabling the F-P and slide:
   • Ensure that the CS100 controller is off, and that the UNSWIRF controller is off.
   • PREPARE TO connect the high-voltage extension cable (the one with the green box at one end) to the cable going to the UNSWIRF controller box. BEFORE CONNECTING have a look at all the recessed connectors in the green box; it is possible that one or more are pushed in, in which case you will need to remove the four slot-head screws and loosen the cable clamp, and push the connectors out from behind. When connecting the green boxes together, ensure that the protroding pins are straight (if not, push them with a screwdriver). The boxes should go together fairly smoothly, and completely. If not, you are probably forcing one of the recessed connectors out of its position.
   • Connect the 5 plugs into the sockets on the side of the F-P. The X, Y, and Z cables must go to the correct sockets (which are X, Z, and Y from left to right). All plugs fit easily if you have the orientation correct.
   • Connect the multi-pin plug into the socket near the motor.
   • Turn on the UNSWIRF controller and boot the PC, so that the slide can be moved.
   • Make sure that all cables are well supported so that they will not become tangled in the IRIS belt drive when the F-P is fully retracted out of the beam, and so there is enough slack to allow the F-P to be driven fully in. Use cable ties to secure the cables to the drive motor. Use the UNSWIRF commands ``slide out'' and ``slide in'' to check the cables. Keep an eye open for things being damaged as the slide moves. If you need to stop the slide quickly, just turn off the UNSWIRF controller.
7. Verifying the slide operation:
   • Having executed at least one ``slide out'' and one ``slide in'', type ``status'' to check the computer's idea of the slide position. You should see something like:

          homelimit =     ?; homeindex =  6579; home =  6528
          awaylimit =     ?; awayindex =     0; away =   128
     

     The numbers should be within a few units of the above. In particular, ``home'' minus ``away'' should be very close to 6400 counts. ``awayindex'' is defined to be zero, and corresponds to where the slide interrupts an opto-switch with the F-P in the beam. ``away'' is the number of stepper motor steps from the ``awayindex'' back until the motor's incremental encoder index is hit. ``homeindex'' if the number of stepper motor steps from ``awayindex'' to the opto-switch with the F-P completely out of the beam. ``home'' is the position of the incremental encoder index near ``homeindex''. The ``homelimit'' and ``awaylimit'' numbers are only set if the slide runs into hard limit switches that are beyond the opto-switches, which normally does not happen.
   • If the above behaviour is not observed, you will need to investigate the problem.
8. Setting up the CS-100 controller:
   • Controller off.
   • Protection on (this is a small switch on the back of the CS-100, about half-way up, and one-third of the way across from the left). It should be UP (if not, you risk damaging the F-P if an overload occurs).
   • Set the three fine X,Y,Z dials to zero (leave the resistive dials alone).
   • Set the three coarse settings to zero.
   • Set ``Coarse Z'' to ``local''.
   • Set the three gains to 32, and the three timeconstants to 250.
   • Set the close-loop and integrate switches down; the meter switch up (to ``error signal'').
   • Turn the controller on. The meters will probably all go off-scale (after a short delay.
   • Select ``resistive component'' on the meter switch, and adjust the three resistive (``R Balance'') dials to give 0,0,0 on the meters. This adjustment may need to repeated after the F-P and CS-100 have been on for a few hours. Note that it affects the parallelism of the etalon. For reference, the dial settings on 26 Oct 96 were 3.82, 4.27, and 3.63.
   • Select ``error signal'' on the meter switch.
   • Set the close-loop switch up. After a short delay, the meters should all stablise on range, and the ``close-loop'' lamp should light. If the etalon vibrates noisely, switch back to open-loop and investigate the cause of the problem (cabling?).
   • Set the integrate switch up.
   If the above sequence fails, you will need to read the CS-100 manual for further advice.
9. Remove the perspex cover from the top of the Fabry-Perot, taking great care not to drop the screws onto the etalon.

   Setting the parallelism

   The F-P and its controller appear to be stable enough that the plate parallelism can be dialed-up and forgotten about as long as the temperature and humidity do not change dramatically. Nevertheless, it is good practice to check the parallelism, since if wrong it will make your data very hard to analyse. The parallelism can be checked most easily during the daytime using an argon comparison lamp.

   1. Recheck the resistive balance of the CS-100.
   2. Place an A4 sheet of white paper above the F-P on the aluminium mounting flange to cover the 200mm hole (use sticky tape to hold the paper in place). The purpose of the paper is to act as a diffuse reflector for the comparison lamp.
   3. Place an argon comparison lamp in the UNSWIRF box. The lamp itself should rest on the black aluminium cover into which the F-P retracts when moved out of the beam. Use some sticky tape to stop the lamp from moving around (and possibly falling onto the F-P).
   4. Plug the lamp power supply into one socket of the mains outlet that is controllable from the console (ensure that the UNSWIRF rack is not plugged into this socket).
   5. Turn off the lights (incandescents and fluoros) in the cass cage. It is not necessary to turn off the dome lights.
   6. On the IRIS VAX terminal select METHOD 1, CYCLES 2, PERIOD 1, TIME 1.5, IME 1, ITIME 1.5, LF open, UF 1.65. These parameters are suitable for the 1.6519 micron argon line.
   7. On UNSWIRF do ``x -490 y 499 z -30''. If necessary, put the F-P in the beam with the command ``slide in slide 150''.
   8. With the lamp off, take an exposure and use it as a dark (i.e., ``use nnn dark'').
   9. Turn the lamp on. You should see an obvious increase in flux, hopefully in a circularly symmetric pattern.
   10. Now tweak the ``z'' setting (using commands such as ``dz 5'' or ``dz -5'') and note the affect on the display. As you go to more negative z, you should see a donut of light steadily moving out towards the edges of the plates. Make a note of the ``z'' setting at which the donut is at the periphery (or 90% of the way there).
   11. Use the Perl script cube.pl to set up a run file (use ``./cube.pl > /vaxdata/obsred/unswirf/yourfilename.dat'') of 16 or so images at delta-z of 2 about the line centre, and execute this file using ``unsw yourfilename.dat''.
   12. Convert the data to FITS format by running docube on the output files, then read the FITS cube into IRAF with rfits.
   13. Run the unswslope.cl script.
   14. Take the dX and dY values output by unswslope.cl, and append them to the lines $xx=... and $yy=... in BOTH arc.pl and cube.pl.
   15. Repeat the comparison lamp cube, rerun ``unswslope'', and iterate the correction if necessary. The absolute values of the X and Y slopes should both be less than about 0.01 (corresponding to a precision in the X and Y settings of about 2).

   Problems that have occurred with UNSWIRF in the past:

   • A dodgy 50-ohm ethernet terminator resulted in intermittent communication with the PC.
   • Damaged ethernet cables.
   • Dodgy high-voltage cables from the CS-100 to the F-P caused an inability to close the F-P loop. An electrical shock was experienced when touching some of the cables.
   • CS-100 loosing lock. Possibly humidity related (both too high and too low), possibly related to the above arcing problem. The only remedy appears to be to open the loop, power-cycle the CS-100, and close the loop.
   • The etalon drifts in both Z and parallelism about a factor of ten more than it should. Correlated with humidity/temperature. The settings are also sensitive to the resistive balance, and may be affected by gravity (i.e., telescope position).
   • A loose connector to the National Instruments 96-bit DIO card in the PC caused unreliable operation of the CS-100.
   • Contacts pushed out in the green box high-voltage connector.
   • DITS_NETSTART failed on the VAX due to software modifications.
   • Offset run files failed due to software modifications.
   • Communication problems between the VAX and the Sun. See the earlier section on how to fix this.
   • Sometimes the NFS fileserving between the Vax and the Suns has problems. It can be fixed by rebooting the MicroVax 3800 and restarting the Observer and UNSWIRF software.

   Existing problems:

   • The UNSWIRF ``slide in'' command does not return control to the user when it should (workaround: send a carriage return after waiting 60 seconds). This may be due to a faulty limit switch.
   • Offset-run files sometimes crash with an error in END OBEY.
   • XMEM crashes about once a night.
   • Vax crashes about once every couple of nights.
   • SLEW commands from ICL often fail.

   Planned enhancements:

   • Software-operable comparison lamps and flip-mirror.

   Last updated 19-Apr-1997

   UNSWIRF: hardware / Michael Ashley / mcba@newt.phys.unsw.edu.au