MJI - 28/10/97
NAW - 26/11/97   (conversion to HTML and amendments)

 Startup of the observing system
 Recovering and Tips
 Telescope object catalogues
 Driving the telescope from the CLI
 Driving the autoguider from the AG CLI
 SNAFU Procedure
 Updating Filter Database
 Guide Star Selection Catalogue Programme
 Observing Log
 Using the Gantry Terminal
 Useful CLI commands
 Current Properties of the MOSAIC CCDs 
 Physical layout of the MOSAIC


1. Logon to LPSS15 as observer on system console (VT220). At prompt ">" type
>obssys          - select option (3) to bring up S5... of system
>startobssys     - brings up the DAS tasks for data capture
if all goes well you should see

dasload: Loading DAS1
dasload: Loading DAS2
dasload: Loading DAS3
dasload: Loading DAS4

if not try the following to clearup observer related processes and try again


This also the recommended way to cleanly exit from the system task.

[NB. If LPSS15 is up and running normally there is no need to do this stage]

2. Logon to LPSS12 as observer on its system console

in "pink" xterm window type

>obssys          - defines aliases etc... select option (3) ie. S5... of system
>startobssys     - fires up observing system
A series of windows will spontaneously appear including:

the talker log - gives updated diagnostics of what is going on and useful to keep visible to check on system status,

the observing log - the observing sequence log can be safely minimised,

the MOSAIC gui interface - where if you cannot type but can drive a mouse you can run the observing system from.

the original pink xterm window is the CLI (command line interface) alternative to the gui but also has enabled a large number of extra commands.

To exit cleanly from a session on LPSS12

>shutdownobssys - answering [Y] tothe prompt about the wiping the system

                     removes all observer-related DA processes running on
                     LPSS12, including the TOG AG display and DAS mimic etc..
[NB. If you are unsure about state of LPSS12 regarding leftover observing processes >cleanup in "pink" CLI window will check to see if any left and remove them.]

3. Startup on TOG X-terminal

Open an x-window session on lpss12 using NCD

logon as observer with the usual password

in xterm window type

>obssys          - defines aliases etc...
>startobssys     - fires up observing system and starts autoguider display
>shutdownobssys - finish off session cleanly, but unless you want to kill
                   the observers system too, answer [N] to wipe system 
[NB. On current NCD Xterm press "pause" key to bring up menu and then open serial connection for usual TCS control window and telnet to LPSS12 for x-window session.


First on the system computer (LPSS12) at the SYS> prompt type

shutdownobssys - and answer Y to wipe the system of observer processes

At the DAS console on LPSS15 (the VT220 above the FOX rack) type


Go to CLIP centre and press the red reset button on the autoguider VME crate (you may as well toggle the reset switch on the AG controller crate above it). DON'T do this if you are unsure about which crate it is - there is a large choice of systems you can screw up.

Next wait 1 minute for VME sub-system to come up (or walk slowly up stairs) and bring the rest of the system up in the following order:

At the DAS console on LPSS15 enter the command

   startobssys     - if you are unsure about which version type "obssys"
                     first and answer 3 for S5.... version.

and wait until you see the "dasload: Loading DAS1" ...... etc.... responses.
Next at the observers console on LPSS12 type

startobssys - ditto as above.

and wait for the SYS> prompt to come back (takes about 3 minutes).

Finally at the TOG console type


and wait for the TO> prompt. This fires up the TO's local display and the autoguider window etc..

[The only time this has failed to work so far on an otherwise running system is if the autoguide CCD controller at the PF top end has gone into another universe in which case bring the telescope to access park and reboot the bugger by power cycling it. The symptom here was that everything in the preceding recovery worked fine apart from the autoguider could not "read the CCD size" - a bit of a giveaway in retrospect !!

NB. If the link to the autoguider controller is active you could first try a soft reboot using the commands: >tip ageng, <cr>, reboot, ~. to quit. Obviously you can leave the telescope where it is to try this.]

  1. If a run task on the MOSAIC gui hangs first try the "abort" command. If this fails then go to "pink" CLI window and "kill" the wfcrun(flat|bias|glance) process. If this fails try "dasreset" if this also fails give up and restart the observing system as above.
  2. If for some reason LPSS15 needs restarting in middle of observing the command ">dasload lpss15" typed in the pink xterm window will reconnect the DAS system to it.
  3. The comms links to the autoguider and MCA go via two digiport servers, one in the CLIP centre and the other on the CASS station of the telescope. If you suspect the comms links to either the autoguider or the MCA and science controller are not working first try:
>tip meng - to connect to Science CCDs controller

This will either "connect" successfully in which case type "~." to quit or it will return with "link not available". Likewise

>tip ageng - to connect to AG CCD controller

>tip agvme - to connect to AG VME crate

will show whether or not comms link is alive or alternatively whether VME crate is stuffed.

>ping - checks if CASS station digiport server is alive >ping - checks if CLIP centre digiport server is alive

If either of these digiport servers have been powered cycled then the serving daemon on LPSS12 will need killing and restarting. First shutdown the observing system as above and then reboot LPSS12 using the captive SHUT account (usual password).

4. If the agvme link is still down it may be the VME crate, again kill all the observer processes as above and go to CLIP centre and "reset" the AG VME crate (make sure you know which crate, if in doubt get help).

5. When firing up current version of software ALWAYS bring up LPSS12 observing system first before starting AG display on TOG terminal. Likewise if problems occur requiring restart ALWAYS kill AG tasks first before wiping rest of system processes.

6. Read the "talker" messages they tell you what is happenning.

7. Read Guy's notes on recovering from a DRAMA crisis.

8. Finally if you are unsure about any of the preceding recovery procedures call for SA or DT support.


The catalogues for driving the telescope from the DAS are stored in the directory $OBSSYS/etc. There is a grid of bright pointing standards in and the active telescope catalogue is stored in

To view and manipulate the catalogue use the normal unix tools eg. "more", "cat", "emacs" or if you want to add objects to it without using an editor at the xterm CLI type

>object name ra dec equinox - eg. object JQ1 14 05 13.4 -13 0 5 J2000 >add

and then

>gocat name - to move telescope to object


Some more examples of useful TCS commands:
>guide on pixel x y         - guide on at position x,y on autoguider CCD
                              default is to specify x y = -1 -1 which
                              cause A/G to guide on current guide star location
>guide off                  - turn guiding off

>focus 45.5                 - set telescope focus

>rotate sky 180             - move PF rotator to 180 degrees [default]

>gocat 12+30                - go to pointing grid star assuming
                              is part of the user catalogue in 

>offset -440 a 440          - which offsets telescope from nominal pointing
                              centred in middle of array, to approximate
                              centre of CCD#4.  The "a" denotes offsets
                              are in arcsec.  If you prefer CCD#3 then
                              -440 a -440 will place target at approximate
                              centre of CCD.
>mikefocusrun arg1 arg2 arg3 arg4 arg5 - procedure for doing a focus run

The arguments are:

                      arg1  - no. of images to take
                      arg2  - separation of images in arcsec N-S direction
                      arg3  - focinc, ie. focus increment (0.1 average seeing
                              0.05 good seeing)
                      arg4  - focus start position (tube temperature 4 deg
                              current best focus with R filter is 45.5)
                      arg5  - integration time (best if >= 5s)
The last image is double spaced. After deciding on best focus use TCS focus command (as above) to set telescope focus. This procedure only uses CCD #3 and #4.

[There is now a TCL gui available to do a focusrun which is invoked by ... focusrun]


Some useful commands:

>autotv [on|off] - continous readout mode

>field - do a field and search for potential guide objects

>guide on star "n" - pick "n"th brightest detected object and guide

>guide off - stop guiding

>guide on pixel x y         - if you want to get exact same field on CCDs as
                              previous exposure. [NB. Sequential frames on
                              same field do this as a default since you just
                              leave the AG guiding even between exposures.]

>agwin aqu pos x y          - set centre of field window for aquisition

>autotime "t"               - set integration time "t" in ms.
If you accidently kill the AG display window you can restart it from the CLI by typing

>autodisp AUTODISP-TO & - assuming nothing else has fu'd.


Go to bright star at nominal centre:
eg. use pointing grid catalogue command eg. "gocat 12+30" in xterm session on DAS system, or use TCS to go to suitable target.

Do a rough SNAFU using the TV finder (centre on "blue" blob if its still there) and ACK. Then do a finer SNAFU with...

Either beamswitch A to place star on science device #3 and take exposure with say 4x4 binning to determine position of star. Compute offsets in arcsec to place star on centre of active area eg. pixel coordinates 256,259 for 4x4 binning - each pixel is 4x0.37 arcsec - but do nothing yet.

Or beamswitch B to place star on default autoguider window. Use autoguider 'field' command to take exposure. Compute offsets in arcsec to place star on pixel coordinates 1032,850 (for 1x1 binning - each pixel is 0.37 arcsec) but as above do nothing else yet. Alternatively use 'tv' mode to take continuous exposures and interactively position star on cross after starting 'SNAFU' procedure.

Define suitable HA/INC on TCS put in 'INC' mode and then do 'SNAFU' using handset to move star by computed amount as above. [If rotator at 180 degrees default science display (eg. SAOIMAGE) has N at top and E to left, whilst autoguider has S at top and E to left.]

Check moved correct amount by taking another exposure on either science device or autoguider. If ok press 'ACK' and you are done if not move some more and repeat check.

An extra test for correct rotator centre position is to use beamswitch A to put star on middle of device #3 and then use command "rotate sky 360" in DAS xterm window to move rotator through 180 degrees. Take exposure on #3 again with 4x4 binning and check star is close to centre of chip.


[The filters are currently stored in a special WFC filter box together with spare filter spacers.]

At xterm CLI interface type

>wfcmdb                     - enable access to WFC filter database (eg. set
                              telescope focus offsets)
The filters are identified by several methods.

A WFC filter holder number written on each one for ease of identifying when changing filters - this number has no other significance.

An ING site-wide database number not directly used by the WFC software but stored in the WFC filter database.

A unique 5 bit code on each filter holder - see below.

A name translation of the 5 bit code eg. R, U etc...

A filter position number on the filter wheel eg. 1,2.....6

If simply altering a known filter offset follow the interactive gui and "commit" the changes to the database and "save" on exit.

After exiting from gui type

>cmd MCA RESET              - to update the volatile use of the database
                              and enable the changes.
If changing a filter to another known filter see the section on use of GANTRY terminal.

To install a new filter (in a unique filter holder) you have to "read" the filter holder 5 bit code on the edge of the filter holder. 

         inside edge  |   Bits 0  through  5     |  outside edge
where a set bit is a raised part of metal and the least significant bit is to the left.

For example: the OIII filter has code 1 0 1 1 0 = 13

Update the WFC database using the >wfcmdb and >cmd MCA RESET commands.


The HST guide star catalogue to V of approximately 15th magnitude is stored on /soft/cats/hstguide.sun . This 300 Mbyte file contains the complete GSC catalogue and can be efficiently searched for guide stars for the WFC.

To run the software the following environment variables and alias have to be set up in an observing system xterm window on LPSS12: (these should be setup automatically by startobssys)

>setenv GSC_DATA /soft/cats/hstguide.sun - pointer to the HST catalogue file >setenv PGPLOT_FONT /star/etc/grfont.dat - an interim bodge >alias gsc_apm '/soft/cats/gsc_apm' - for the PE TCS version

>alias gsc_apm '/home/mike/soft/apm/gsc_apm' - if you want the latest version

with 5 beamswitch options.

The program can either be invoked using command line options such as:

>gsc_apm RaDec="10 30 0 -1 0 0" Rotpos=180 Equinox=J2000 >gsc_apm Rotpos=90 Equinox=B1950 RaDec="21 30 11.3 12 10"

or run interactively viz:


Command lines arguments are:

RaDec="hh mm ss dd mm ss"
RotPos=angle (180 default)
Equinox=J2000 (or B1950)
Beamswitch=N (or A,B,C,D)

Example command lines:
gsc_apm RaDec="10 30 0 -1 0 0" Rotpos=180 Equinox=J2000 gsc_apm Rotpos=90 Equinox=B1950 RaDec="21 30 11.3 12 10"

Ok...... do it interactively then.......

Rotator angle [180?]

Do you want to impose magnitude cut ? [N]

Do you want to precess input coordinates to J2000 ? [N]

Field Centre:(h,m,s,d,m,s)
12 0 0 0 0 0
File: gsc.out already exists - clobber ? [Y]

Catalogue Ra limits = 8963072 9001638

Possible guide stars (also listed in gsc.out)

Catalogue no. Ra Dec Magn Guide X Guide Y

49320270 11 59 44.965 - 0 25 11.57 13.6 1512.7 20.1 49320404 11 59 48.562 - 0 21 11.56 14.6 1367.0 668.0 49320532 11 59 56.012 - 0 19 22.73 15.2 1065.4 961.7 49390619 12 0 0.230 - 0 19 21.36 12.6 894.7 965.4 49390146 12 0 18.344 - 0 23 41.42 15.2 161.2 263.4

Total no. catalogue stars read in = 161

Do you want plot E-W direction reversed ? [N]

Do you want to alter default line width ? [N]

Graphics device/type (? to see list, default /NULL):

Do you want another plot ? [N]


The catalogue information is also written to gsc.out in the current directory. A plot of a 1x1 degree field with the layout of the science and autoguider CCDs superposed is optionally available.


The observing log is actually stored in /ing/etc and called something like 19970525.INT . For printing out your own version copy to a suitable directory, edit out any crap and put in any comments and then use

>a2ps -1 19970525.MINE >

if you want a readable printed copy.


For changing filters or rotating top end for cryostat fill, which should be done at rotator position of 180 or testing shutter etc..

Check RS232 cable is connected to top socket on gantry rail.

If console does not respond check communication is set for 8 bits no parity by pressing "PF3" and then using cursor keys to select COMMS option and "enter key to scroll through options for 8 bit no parity. After setting COMMS press "PF3" to go back to terminal mode.

Hit <cr> a couple of times to get DAS system login prompt.

Login as observer, using standard password.


> obssys                - sets up aliases etc..
[menu]                  - select item 8 ie. WFC option
> startobssys

To change a particular filter

GANTRY> change R - filter name or number, eg. R, 3

....wait for action to prompt.....
....turn the black knurled knob on filter access hatch to open it.... ....remove filter and change for new one.... ....turn black kbob back to close the hatch which is then locked automatically

To rotate PF top-end

GANTRY> rotate sky 180 - rotate PF top end to 180 degrees

To test shutter

GANTRY> wfcshutter open
GANTRY> wfcshutter close

Logout when finished

GANTRY> logout


The MOSAIC gui is intuitively obvious even for astronomers. However, for direct TCS control including autoguider on/off and other diagnostic commands the CLI interface is needed.

DITS (Distributed Instrument Tasking System) task command takes the generic form: >cmd [options] D-task [action arguments]

Some useful options are: -g which gets information about D-task

To reboot CCD controller:


To initialise MCA


To run observing system from CLI

>filter R - change to R filter >run exp-time ["title"] - do a science run

>bias                      - take a bias frame
>glance                    - glance frame
>abort                     - terminates active run without saving data
                             [do not use if CCDs are reading out]
>finish                    - terminate exposure and readout

>dasreset                  - restarts data aquisition subsystem after fu's
>cleanup                   - clear up after system hang up before restarting


Turbo is the recommended readout speed = total readout time of 130s for all 4 science devices.

At turbo the gain and readout noise for the active devices are:

CCD Gain e-/ADU Readout noise e- Full Well e-

        #2           1.6               14              80,000+
        #3           1.45              13              80,000+
        #4           1.4               17              60,000
The CCD broadband zero-points (ie. that magnitude that gives 1 photon/s) are:


        #2        23.9   25.3   25.5   25.5   25.0
        #3        21.5   24.6   25.2   25.3   24.8
        #4        23.8   25.5   25.5   25.5   25.0
#2 Has vertical streaking in bias frames #3 Currently has sensitivity problem and non-uniform response #4 Charge injection problem limit full well to 60,000 e-



CCD Layout and orientation for rotator at 180 deg:


          |              |   |              |
          |              |   |              |
          |              |   |              |
          |      #3      |   |      #2      |
          |              |   |              |
          |     W8-1     |   |    W12-1.1   |
          |              |   |              |
           --------------     --------------

      E                    x
           --------------     --------------
          |              |   |              |
          |              |   |              |
          |              | o |              |
          |      #4      |   |       #1     |
          |              |   |              |
          |    W5-1.1    |   |    W16-1.1   |
          |              |   |              |
           --------------     --------------
                   |             |                   
                   |      #5     |                   
                   |    W10-4    |                   
                   | Masked for  |                   
                   |   frame     |                   
                   |  transfer   |                   
Active area 2048x2048 pixels, overscan regions give 2076x2076 pixels on readout. Nominal 15 micron pixels (needs astrometric check) gives active area of 30.72x30.72mm or 12.6x12.6 arcmin per CCD.

NB. The rotator centre (o) is offset from the nominal mosaic centre (x).

#1 Dead device.

#2 Charge transfer problems but otherwise ok.

      Trimsec [13:2060,29:2076]
      Biassec [1:2076,1;28]
#3 Sensitivity problem causes apparent vignetting otherwise ok.
      Trimsec [1:2048,13:2060]
      Biassec [2049:2076,1:2076]
#4 Wafer mask 'tram lines' visible otherwsie ok.
      Trimsec [17:2064,1:2048]
      Biassec [1:2076,2049:2076]
#5 Autoguider CCD with outer half masked for frame transfer.

Note that flip/rotation of CCD pixel array is applied at readout to give correct orientation on sky - necessary because of symmetric corner readout of devices.

Capstan settings viewed with rotator at 180 deg


                |                                        |
                |                                        |
                |     8.93mm                             |
                |                                        |
                |                                        |
                |                                        |
                |                                        |
                |                                        |
                |                                        |
                |                             7.43mm     |
                |                                        |
                |                                        |
                |                                        |
                |                                        |
                |                                        |
                |                                        |
                |     7.45mm                             |
                |                                        |
                |                                        |

Astrometry of CCD layout relative to rotator centre.
Using a conventional x ---> coordinate system in the 180 rotator orientation illustrated above and defining the rotator centre as the origin:

The nominal centre of mosaic is defined to be at (0",281") or (0mm,11.38mm) from rotator centre.

#1 Unknown

#2 Centre of active area is at (449",717") or (18.17mm,29.01mm)

#3 Centre of active area is at (-398",710") or (-16.10mm,28.74mm)

#4 Centre of active area is at (-398",-148") or (-16.11mm,-5.99mm)

#5 Offset to default size frame window (400,288) centred at (1032,850)

(56",-1210") or (2.3mm,-49.0mm)

This gives TCS aperture offsets of:

Nominal to rotator centre AOXNOM = -1.3623D-3 (radians) Aperture A ie. nominal to centre of #3 TDATA(11) = -429 (arcsec)

                                         TDATA(12) =  398 (arcsec)
Aperture B ie. nominal to autoguider     TDATA(13) = 1210 (arcsec)
central default window (1032,850)        TDATA(14) =  -56 (arcsec)
Pointing model constants from 23/5/97 automatic pointing test

PCIH = -1.61615D-3
PCID = -3.46446D-3
PCCH = 5.81470D-4

using Perkin Elmer nomenclature.