Contents

1. Introduction
2. WFI at the AAT
1. Where to find more details
2. WFI Background
3. WFI Layout on Sky
4. Differences between WFI and OBSERVER
5. The WFI Hardware
3. Shutting Down
4. Startup
1. Start the PFU Widget
2. Start the WFI_TEL_CONTROL task
3. Start the WFI Cicada system
4. After Starting WFI/Cicada (and before observing)
5. A Tour of the WFI Cicada Interface
6. Observing with WFI and Cicada
1. Looking at your data
2. In the Afternoon
3. Start of the Night
4. In the Morning
7. Using prepared scripts
8. Writing scripts
9. Guiding
1. Should I guide?
2. The Multiplexer & MaximDL Software
3. GSFIND and the AAT_TELESCOPE task
4. Guiding - Step-by-Step
10. Saving your data.
1. Formats
2. Drives
3. Writing a Tape
4. Writing more data to a tape which already has data on it.
11. Frequently Encountered Problems, or "Things that go wrong when I do xxx ... "
1. Things that go wrong when I try to start Cicada/WFI
2. Things that go wrong when I try to take data
3. Things that go wrong when WFI talks to the telescope
   
4. Things that go wrong with PFU
5. Things that go wrong with the focal plane temperature
12. Frequently Made Mistakes
13. Frequently Asked Questions
14. Appendices
1. Twilight Flat Fields
2. Other bits of Cicada you might want to know about

1. How to use this Cookbook

Section 2 is a Short Guide to WFI and PFU and their hardware. Section 3 describes shutdown procedures. Section 4 describes startup procedures and what you need to do at the start of each night. Section 5 is a guide to the WFI Cicada control panel. Section 6 is a step-by-step guide to typical observing procedures with WFI. Sections 7 and 8 describe the use of scripts to set up more complex exposure sequences, including dithering, automatic focus adjustment, observing standard stars, etc. Section 9 describes guiding procedures.  Section 10 covers writing your data.  Sections 11-12 are the Frequently Encountered Problems, Frequently Made Mistakes and Frequently Asked Qustions pages. Some Appendices are included at the end.

What should your read? Well the lot obviously. For the impatient here are the important bits.

Impatient People Preparing for a Run :  Read Short Guide, the Guide to the WFI Cicada Control Panel, the Observing Guide and the Frequently Encountered Problems, Frequently Made Mistakes and Frequently Asked Qustions.

Impatient People at the Telescope : If you aren't going to guide, you can skip the guiding section. But read the rest. Really, you'll save a lot of trouble if you do.

2. WFI at the AAT - A Short Guide

2.1 Where to find more details

2.2 WFI Background

WFI is shared facility which was constructed as a collaboration between the Research School of Astronomy & Astrophysics (RSAA) of the Australian National University, the Anglo-Australian Observatory, the University of Melbourne, and Auspace. Its use is shared between the RSAA 40" telescope on SSO and the AAT. While the WFI mosaic itself moves between the AAT and the 40", each telescope has its own "exposure controller" (i.e. shutter and filter wheel).

Prime Focus Unit (PFU) is the AAT's exposure controller.  It contains a six position filter wheel able to hold filters up to 165mmx165mm in size, and 10mm thick. It also contains a two-travelling-blade shutter, able to uniformly expose the entire focal plane at better than 1% for exposures longer than 2s.

The entire 128Mb WFI 8K image is read from the CCDs using 8 parallel controller channels, displayed in real time, and transferred to disk in 58-60s (depending on computer load). The data are written to disk as multi-extension FITS files.
 

2.3 WFI Layout on Sky

2.4 Differences between WFI and OBSERVER

• OBSERVER will not let you delete or over-write your data. But Cicada will! So be very careful with those 'rm' commands, or changing the run number to overwrite existing files. Once gone they are gone for ever.
• When Cicada takes a "Display Only" exposure it does write a temporary file, which you can save if you want. But that temporary file is immediately over-written by the next "Display Only" exposure. You don't get DUMMY files from a to z as with OBSERVER. (If you don't know what all that means, then the difference is not important to you!)
• Only WFI file headers are automatically archived (under OBSERVER both headers and images are archived). WFI images must be archived by hand by Observatory staff. So you must not delete your data until you've been told by Observatory staff its OK to do so. With 140Gb+ of space available now, this should not be a problem.

2.5 The WFI Hardware

WFI sitting in the Prime Focus cage. Viewed when the  telescope is at Prime Focus access.	Annotated detail of the PFU and WFI.

PFU Filter loading door, showing bar-code reader used to identify filters when the filter wheel is initialised.	PFU Filter door open showing a filter mounted in the wheel. Each filter sits in a small metal holder, from which it is never removed. The thing white strip to the left of the filter is the braced mounted on the filter holder.

3. Shutting down

Shutting down CICADA

First exit the WFI window, and then exit the Cicada window. If the Ximtool window doesn't vanish, then close that as well. Then type cicada_cleanup in the window where you initially started cicada.

Shutting down PFU

You should not need to do this. You can shut down WFI and re-start it without re-starting the PFU task.

Should you need to shut down the widget, just use the "Exit" button. Then type "exit" to exit drama, in the aatpfu terminal window from which you started the widget, and "exit" again to log you completely out of aatpfu.

If for any reason you need to power off aatpfu, then it is very important to shut it down first, by typing "shutdown -h now" instead of the second exit. Once the computer is shut down you can turn it off.

Shutting down the WFI_TEL_CONTROL Task

You should not need to do this. You can shut down WFI and re-start it without re-starting the telescope interface task

If you do want to re-start it, just hit <Ctrl><C> in the wfi_tel_control window, and then log out of aat40a.

4. Startup

4.1. Start the PFU Widget

The PFU is run from a Linux PC  (aatpfu) mounted in the prime focus cage. A task on this PC allows both the WFI CCD controllers to run the shutter, and for the WFI Cicada system to run the filter wheel. This task should be running before before you try to start Cicada running. (In debugging or set up this may be done by a keyboard and screen plugged directly into aatpfu). In normal operation this is done by remotely connecting as follows

1. Start a terminal window and connect or telnet to aatpfu.aao.gov.au (or 192.231.166.31).

On the usual x-terminal used for this (aatxtl) this will look like so...

2. If you are on a Sparcstation, allow aatpfu to pop up a window by  typing xhost aatpfu or xhost 192.231.166.31
3. Login using the username and password located on the wall near aatssx.
4. Type drama.
5. Type cleanup.
6. Type dtk -t tk2.tcl - you should see a widget like the following appear (though the filters slots will be blank).
7. Then click the buttons to "Initialise shutter", "Initialise Filter" (this takes a few minutes), and "Read Filter". Once the filter wheel has finished initialising, PFU is ready.

Sometimes the "Initialise Filter" "Read Filter" will fail to read one or more barricades when the telescope is at prime focus access. Ask the support astronomer, afternoon technician or night assistant to move the telescope to the zenith and repeat the Initialise Filter (followed by the Read Filter). If its still failing you may have a low pneumatic pressure or filter blockage problem.
 

4.2. Start the WFI_TEL_CONTROL task

This task mediates communication between WFI and the AAT telescope control system (TCS). Connect to the AAT40A from an X-terminal (you must use an X-terminal for this).  Then login using the username and password on the wall near aatssx, and start wfi_tel_control.

40a>  wfi_tel_control

This task will then wait for requests for telescope information from the WFI Cicada system. You should see positions etc. regularly spat out here at the end of each exposure (or if the AAT_TELESCOPE task is updating).
 

4.3. Start the WFI Cicada system

Login to the WFI console (aatssx) - usernames and passwords are on the wall.

Then at a unix line prompt type

> cicada_cleanup
> cicada

In a few moments a tiny little obscure looking widget will pop up near the top left of the console screen. This is the Cicada panel.

On the "Start Observing" pull down menu at the left you'll find a few options. Choose "WFI" and the WFI panel will appear. (The other option "WFI_BottomStandalone" should only be used if one of the controllers fails and you can only read 4 CCDs instead of the full 8).

If all goes well, and you have been able to correctly start up the TEL_TASK and the PFU widget before starting Cicada. If you get a nasty red message about a failure of the filter task, then it may mean you haven't started the PFU widget. See the things that go wrong when starting Cicada/WFI.

The first thing you may notice is that the "Open Shutter/Close Shutter" button at the bottom is yellow and says "Close Shutter". For some reason the WFI task starts its idea of where the shutter is,  in a confused state. Click on this and it will clear itself.

This panel is the one you'll use to control WFI almost all  of the time. Its pretty straightforward. But  before you start observing there are a few things you need to do, so we'll cover them first.  You should make sure you do the following well before you want to take your first flats in the evening - if there's a problem starting cicada its easy to lose 10 minutes, which can really stuff you up, if the sky is rapidly fading.
 

4.4 After Starting Cicada (and before observing)

After starting Cicada (and before you start observing) you need to do the following :
4.4.1. Update the location the data is written to, the file prefix, and the starting run number. You must do this correctly, otherwise your data can't be properly archived.

Click "Options" "Preferences" to bring up the preferences dialogue. Then choose "General"


 

1. Set the Prefix: to the appropriate date.

You must use the nnmmm (i.e. 05mar, 30dec etc.) format shown above.
 
 
2. Set the Directory:  to the appropriate date.

You must use the /wfi_{raw,data}?/cicada/YYYYMMDD/A format shown above.
 

There are four disks available for your data /wfi_raw1, /wfi_raw2 (both 70Gb), /wfi_data1, /wfi_data2 (both 18Gb). Mostly you'll use wfi_raw1 and wfi_raw2. Other than choosing the disk you want, you must follow the date format shown above, and you must end it with capital "A". This is again to mimic the archiving used on the AAO's in-house observing system, so the night assistant can archive the file headers and make an observing log.

Once the location for files has been set it will stay that way until you change it - so this only needs to be done once at the start of each day's observing.
 
 

The rest of the options here (and in the other "Preferences" categories) can be changed as you wish, but the defaults will generally be the most useful. Use 'Ok' to dismiss this window when done.
4.4.2. Start & Update the AAT_Filter Task
Under the "Instrument Controls" pull down you'll find two options. One is AAT_FILTER task. Start this and the filter panel will appear. By clicking on the filter buttons here, you will send commands to the PFU, and the filter will be changed.

It is important to note that although the PFU reads bar-code identifications from filters, and passes those filter identifications back to Cicada to get included in file headers, Cicada itself pays no attention to these names. The only filter names Cicada knows about are the filter names hard coded into the AAT_Filter task.

So, if you change the filters in the PFU, therefore, you not only need to reinitialize the filter wheel of PFU, but also update the AAT_FILTER parameters, and then re-start CICADA. Do this as follows.

Under "Options"->"Hardware Configuration ... " (this only works when running Cicada logged in as wfieng. If you're logged in as wfiobs, you'll need to exit, and log in as wfieng to make this change). Use "Category:" at the top to choose Filters, and the select AAT_FILTER from the list at the left.

Please be very careful. You have intimate control over the hardware configurations of WFI here, and can really screw things up. Only make changes in the filter boxes as described below. Change nothing else!



You'll see the filter names in 6 boxes. As shown above B is position 1, U is position 2, g is position 3, r is position 4, i is position 5 and z is position 6. You need to insert the filter names of the filters loaded in the PFU in each box. It is strongly recommended that you use the same formatting as used in the PFU (which comes from the bar-coding). So for example if we changed the g for V and r for R filters above, we'd enter the following new names (ignoring the '#' in the PFU names)

 

1   U_48_(WFI_Schott)	2   B_49_(WFI_Schott)	3   V_50_(WFI_Schott)
4   R_51_(WFI_Schott)	5   i_91_(WFI_SDSS)	6   z_92_(WFI_SDSS)

 
Now you must accept these changes with "Ok", and then exit WFI and then Cicada, and re-start Cicada and then WFI again.


4.4.3. Start the Telescope Control Task
 

Most observers won't need this panel, unless they are guiding.

Under the "Instrument Controls" pull down you'll find two options. One is AAT_TELESCOPE task. Start this and the telescope control panel will appear. This panel enables you to load co-ordinates, offset and slew the telescope. It is generally not needed unless you are guiding, in which case it is essential (for applying offsets to move guide stars into guide CCDs).

If you do need to use it, click "Start update". This will regularly poll the telescope for position information (you should see this in the TEL_TASK window you started on aat40a).
 

4.4.4. The Message Window
You don't need to start this one. It will appear as soon as the WFI task starts doing things (like communicating with the controller, changing filters, etc.). You can look in this window for regular updates on what Cicada thinks its doing. One trick to it, however, is that it usually stops updating such that it automatically scrolls to the bottom when a new message appears. So you have to scroll it down with the scroll bar itself.

4.4.5. The Ximtool Window
Again, this window should automatically appear as soon as you try to take data. You can make it appear manually with "Tools"->"Show Ximtool" pull down. This Ximtool task is a modified version of the IRAF ximtool task, which deals directly with Cicada to display CCD data as it is read out. You can't use this display for regular IRAF use. In fact if you try to start IRAF from another window when logged in as the same username running CICADA and start another ximtool, the two will often interfere with each other.

Don't run IRAF with ximtool from the same username as is running Cicada.

The ximtool window operates in a fairly self explanatory fashion. You can zoom in and out, pan, adjust the grey scale as usual with ximtool. One thing to beware of is that the data values reported by ximtool are only approximate at best, and that the grey scale limits are different in every CCD. So beware, just because one CCD looks darker or lighter than the others, it may not actually be. The grey scale limits could just be screwed up.
 
 

4.4.6. The Exposure Time Counter
Again, this window should automatically appear as soon as you try to take data. You can make it appear manually with the "Tools"->"Show Counter" pull down.

5. A Tour of the WFI Cicada Interface

Most of the control of WFI operation takes place from the WFI Panel. The major controls are highlighted below.

• Data location - this is the directory you set at the start of the night. You can't change it here. This is just for your information.
• Run number - this is the run number that the next file created will have, unless you've started an exposure in which case its the current file. You can change this at any time by simply typing in the text box. If you try to over-write an existing file, you'll get a warning asking you if you really want to do that. If you say yes, the data will be eliminated
• Object name - You can change this at any time by typing in the text box. The Object Name is picked up by Cicada at the start of an exposure, and again at the end of the exposure if you hit return in the text box, so you can update this after an exposure starts. You can select from a list of previous object names using the pull down menu on the right.
• Filter - the filter name for the current filter position. This is the name of the filter as Cicada knows it from its hardware configuration. If you haven't updated this hardware table to match the filter names as reported by PFU, then it can be incorrect.
• Exposure Type
• Object - science exposures
• Flat - calibration exposures, with a header keyword to indicate they are calibrators.
• Bias - a zero second exposure in which the shutter is not opened.
• Dark - an exposure in which the shutter is not opened.
• User - a normal exposure with the image type keyword set to the user defined name.
It is very important for processing this data that you make sure to use the FLAT exposure type for taking flats.
You should also beware that Cicada remembers the exposure time for each object type separately. This can sometimes confuse you if you change the exposure time for the next exposure, before changing the object type and get an unexpected exposure time resulting.
• Save & Display / Display Only / Save Only / Save Image
• Save & Display means data is displayed in the cicada ximtool as it it read out and saved to disk. This is the default.
• Display Only - data is saved in a temporary file and displayed. Use for checking light levels in flats, checking pointing, etc.
• Save Only - because there is no penalty in displaying the data as it is read, there is really no reason to use this.
• Save Image - turns the last temporary file into a 'real' data file. Note however that the telescope and other information in the header is acquired when the temporary file is turned into a real one, not when it is read from the CCDs. So the headers of these files should be treated as suspect.
• Exposures - the number of exposures to do one after another. This is not the exposure time.
• Duration -the exposure time. This must be an integer (if you want fractional exposure times you must set the exposure time units to milliseconds using "Options"->"Exposure Time Unit"->"msec").
• Expose - start the requested exposure(s).
• Open Shutter / Close Shutter - manual shutter control without doing anything to the detector. This is necessary for the acquisition of guide stars (the guide CCDs are behind the shutter).
• Abort - BEWARE. In an ideal world this would terminate an exposure. In practise, it rarely works, and you should never try to abort during a readout. It  frequently results in the crashing of cicada and a requirement for repeated restarts. If you need to abort an exposure, series or sequence see the Aborting FAQ.
• Status - information on disk and memory space, the focal plane temperature and remaining time in an exposure. The nominal focal plane temperature is 183K. As long as the temp is within +-0.5K of this temperature the mosaic should be fine. If it goes outside this range, advise your support astronomer or afternoon technician.
• No pre-flush / No readout - If you check these radios, then each exposure will merely open the shutter. You can use these to do a 'multiple' exposure for acquiring a focus sequence. Check both the radios.  Move the telescope focus to the first value. Press the "Flush" button, and then "Expose".  Then move the telescope on the sky, and adjust focus, and press Expose again. Repeat as many times as you like (usually putting a double step on the sky before the last exposure), and then press Readout. Make sure to turn the radios off for the next normal exposure!
• Comment - this field is truly useless. What you type here (or pull down using the menu on the right) is added to the file header with the FITS keyword  COMMENT. Unfortunately, the FITS keyword COMMENT is routinely used for storing all sorts of crap in the Cicada FITS file headers. So if you put information you really want to remember in here, it is almost impossible to winnow from the dross of all the COMMENT keywords in your files later on. If you want to record comments that can be easily recalled later,  pop the FITS Keyword window. You can create your own unique FITS keyword here, and insert comments into it.
• Pop-up FITS Keyword Window - pops up a window allowing you to insert custom FITS keywords into subsequent file headers. This is more useful than the main Comment field (see above).
• Pop-up Regions Window - the regions window allows you to change binning, select or de-select individual CCDs, choose from a range of predetermined smaller windows, or define your own windows.

 

In understanding the way WFI regions are defined it is important to realize that as far as this window is concerned, WFI is actually a mosaic of eight 2111x4141 pixel detectors, which make up a 8422x8282 pixel mosaic. In fact only 2048x4096 of the pixels in each CCD are active (i.e. detect light) - the rest are overscan regions. However, windows are simply defined in the 8422x8282 pixel plane. If you define a new window and it happens to include pixels which would normally be overscan, then you'll get overscans. If you don't, you won't. Because having overscans is a good idea, user defined windows should be created with some care!   The Region Diagram shows you graphically the layout of the currently selected window for readout. It also allows you to define new regions. The User Defined section allows you to put the pixel values in for new windows. You can put in up to two regions. You'll get the union of those two regions in the output file. If you turn on the radio for region 1, you can then drag out graphically the region in the diagram above - the pixel values will update.  You can also mark regions by dragging them out in the ximtool, and grab them into this window with "Actions"->"Read regions from imtool". The button at the top of the WFI panel does the same thing. The Pre-defined section is the most used. Mostly to read the whole mosaic! You can also chose the central 100x100, 512x512, 1024x1024, 2048x2048 or 4096x4096 pixels of the mosaic, or the central 100x100, 512x512, 1024x1024 or 2048x2048 pixels of each detector. You can also turn on and off individual CCDs. Finally, the Binning & Overscan section lets you change the binning (1x1 and 2x2 are really the only useful ones). Do not use the X overscan or Y overscan panels. You already get overscan by default. Note that reading 1 CCD takes the same time as reading 8. All you save is disk space. But changing binning, windows and number of CCDs adds greatly to your data reduction (most systems have trouble with windowed data - you need calibration files for every window, you have to split the different sized files up, its a pain). It is almost always not worth the size saving of different binnings or windows, compared to how much easier it is to reduce a data set which is all the same size.

6. Observing with WFI and Cicada

• All the filters have different focus values. When running WFI from a script, this can be dealt with by the script. But when observing "one frame at a time" from the WFI panel, you need to remember to get the night assistant to manually change telescope focus when you change filters.
• You will save yourself an enormous amount of trouble in data reduction if you choose a window and binning to use at the start of the night, and just use that. The time saved in readout from windowing or binning is rarely justified by the extra hassle in data reduction.
• Do I need to dither? Not really. filling gaps is only cosmetic and complicates reduction.
• Do I need to guide? Only in U or narrow band.
• How long can I expose? The AAT tracks well up to 10 minutes.
• What should I use as the "saturation" limit?  Keep counts on targets for which linearity is critical below 30,000 adu above bias (or 34,000 adu).

Each of the detectors reaches full well before the 65535 A-D limit at around 80,000 electrons. Because all the detectors have slightly different gains, these full-well limits correspond to different numbers of ADU (ranging from 45,000 to 60,000 adu in raw counts). Near these full-well limts the detectors are quite non-linear. Hence the rule of thuimb above.
        ______________________________________________________
         CCD  Satur'n in       Bias   Satur'n  Gain   Full Well
              (adu above bias) (adu)  (raw adu)(ph/adu)  (kph)
            1    56,000          3450    59,450  1.45    81.2
            2    53,000          4100    57,100  1.71    90.6
            3    42,000          3992    45,992  1.96    82.3
            4    52,000          4549    56,549  1.72    89.4
            5    44,000          2848    46,848  2.02    88.9
            6    56,000          3368    59,368  1.67    93.5
            7    54,000          4002    61,000  1.88    101.5
            8    55,000          4514    59,514  1.69    92.9
 
• Should I bin? 450s is ~ half well in iz. So binning is not a great idea here.
• What calibrations do I need? You need to ensure you acquire the following calibrations
1. Bias frames (at least 5)
2. Dark frames (at least 5, preferably 9) of an exposure time similar to that used for your science exposures.
3. Dome flats for R,r,i or z imaging (at least 5 for each band)
4. Twilight flats for all filters (at least 3, preferably 5 at each band)
5. Sufficient dark sky observations at different telescope positions to construct a fringe from for R, r, i and z bands.

6.1 Looking at your data

It is very important that you not create extra ximtools on the workstation, from the same account as cicada was started from (they get confused and conflicts arise, and you'll end up killing cicada). So to use IRAF to examine your data. (Examining your data from another workstation is not recomended either, as that just means aatssx has to NFS-serve the data accross the network)l.

1. Move to another workspace by clicking in the CDE control panel.
2. Start a terminal, and login in the aatssx as wfiobs (not as wfieng) - passwords are on the wall next to the

telnet aatssx
username: wfiobs
password:
3. Start an xgterm

xgterm &
4. Start IRAF

%cl
cl>mscred
5. Start xtimool

!ximtool &
6. Move to the data directory. For example

cd /wfi_raw1/cicada/20020223/A
7. Display your data one CCD at a time

imexam 23feb0010[6]

Notice that in the example above, we have not used mscexam to display the whole mosaic, but imexam to display a single CCD. Examining the entire mosaic is not recommended, as the large amount of memory this uses can cause resource problems for cicada. We usually use CCD6 for focussing, and single CCD observations.

6.2 In the Afternoon

1. Go through the Startup procedure above to get WFI ready for observing later at night (including changing filters and updating the WFI filter table if required).
2. You can have the lights in the dome turned off after 4pm in the afternoon (they must be turned on for visitors to the dome prior to that). You should use the time after 4pm to acquire
1. bias frames,
2. dome flats, and
3. dark frames.


To acquire domeflats, you need to get the dome rotated to 0 degrees, and the windscreen rotated to 21 degrees. This puts the white patch above the telescope at the zenith (ask the support astronomer or afternoon technician to do this for you). Have the support astronomer or afternoon technician turn off all the lights in the dome. Then go into the dome and check all lights are turned off.

On a green panel to the far left of the bank of terminals in the control room, is a HPM 2000 VA light dimmer. Turn on the switch below the switch marked "BY PASS" ("BY PASS" bypasses the dimmer, turns the dome flat lamps on full). The dial will now control the brightness of the one of two lamps which can be plugged in to illuminate the dome up at Prime Focus Access. The bright lamp (rather than the desk lamp) can usually be used for dome flats - but you should point the lamp away from the white patch so it is indirectly illuminated.

Take Test Exposures : Use the Regions Pop-up to select a 100x100 pixel window in the centre of each CCD. Also set the WFI Panel to "Display Only" and set the Exposure Type to "Flat". Now take a 1s exposure, and look at the brightness level of the image. Now adjust the brightness of the  dimmer up or down so that you can get ~30,000 adu in each CCD in 5-10s. (Dome flats shorter than 5s are not really recommended).  Be careful that you don't over-expose the detector. When saturated, the counts "wrap around" to small numbers, and you can get very confused. Always start with 1s exposures and the lamp dim, and then work up to the correct time and dimmer setting.

Take your Dome Flats : Once you've got the exposure time and dimmer right, turn off the "Display only", put the Regions window back to the Full Mosaic,  set the number of exposures to 5 (or more), and press "Expose".

Then change to the next filter and repeat.
 
 

3. If you observed last night, check that last night's data has been written to tape correctly. If not, start it writing again when you go to dinner.

6.3 Start of the Night

1. Twilight Flats : After dinner you need to get ready for twilight flats. You need to be ready to observe on sky for flats by around 30 minutes after sunset in the evening (or before sunrise in the morning). If observing in the U you need to start about 15 minutes earlier (or later).

 

 
 
 
 
 
 
 

When the night assistant opens the dome, have them slew to a blank sky field. Then proceed as for dome flats with test exposures (small windows, display only, exposure type=flat). Again aim for exposures of 5s or longer. Remember to turn the Display Only off and the full mosaic on when you are ready proceed after a test. Between each exposure on sky you'll need to increase the exposure time by a factor of about 1.8, and also offset the telescope by about 20s North-South (North-South is recommended so E-W bright star trails are less likely to overlap).

Keep going through your filters until the sky gets too dark (exposures of longer than 120s are pretty much useless as twilight flats).
 

2. Pointing : the night assistant will then want to observe one or more SNAFU stars. These allow us to check the telescope's pointing. The default is to set the AAT's Reference Axis to point stars at the very centre of the mosaic (i.e. the join of the four inner CCDs). We usually then set up Axis A, to point at the middle of the lower half of CCD6.

 

 
 
 
 
 
 
 

Use the Reference Axis for survey programs, and Axis A for programs where you want to observe a single object, and make sure it doesn't land on a gap between the CCDs.
 

3. Focus the Telescope: After taking flats, and a few pointing checks, it is time to focus the telescope. The focus offsets between the WFI filters are well known. You only need to focus the telescope in one filter. To take a focus frame we manually take a multiple exposure (the WFI focus function does not work at the AAT) around the expected focus value (seven steps of 0.1mm is usually about right).
1. You can usually focus pretty much anywhere on the sky. WFI usually sees lots of stars in 10s almost anywhere. Set the exposure time to 10s and "Save & Display".
2. Check the No pre-flush / No readout  radios (each time you press expose now, you open the shutter, but don't readout the detector).
3. Have the night assistant move the telescope focus to the first value in the sequence.
4. Press the "Flush" button, and then "Expose".
5. When the exposure finishes, the night assistant can move the telescope on the sky, and adjust focus to the next value.
6. Expose again. Goto step 5 as many times as you like (usually putting a double step on the sky before the last exposure)
7. Finally press Readout. You can then examine the image in the ximtool, or re-display the CCDs one at a time in IRAF to determine the best focus.
8. Make sure to turn the No pre-flush / No readout radios off for the next normal exposure!


At the end of the February 2002 run, a script for focussing was created - see below. It has not been tested in anger, but should work.

1. Start Observing : you're now ready to start observing. You know where your objects will appear on the CCD, and what is the best focus for each filter.
• You can observe by having the night assistant move to each target, manually apply offsets and focus  offsets, and manually start exposures, repeating them in sequence if desired.
• Or you can use some of the prepared observing sequences to chain together observations of different times in different filters with offsets and focus offsets taken care of - see below.

1. Photometric Standards: We recommend the use of Landolt standards from  Landolt 1992 AJ....104..340L .  (An ASCII version of the catalogue is available here). Many of the fields in this paper can be used to put standards in all the CCDs simultaneously, but they can be a bit on the bright side. The following fields are used by the WFC survey on the INT. These fields may need to be defocused in use at the AAT.

sa92-335   00 55 00  00 44 13 J2000
sa95-330   03 54 31  00 29 05 J2000
sa98-670   06 52 12 -00 19 17 J2000
sa101-427  09 57 26 -00 17 16 J2000
sa104-334  12 42 21 -00 40 28 J2000
sa107-602  15 39 19 -00 15 29 J2000
sa110-361  18 42 45  00 08 04 J2000
sa113-163  21 42 35  00 16 46 J2000

Alternatively, the Landolt list contains a number of fainter sequences clustered into regions of ~5 arcminutes in size, which will fit into one CCD. These can be observed for somewhat longer exposures and with less defocus. Click on the links for finding charts (from the ESO La Silla EMMI calibration plan documentation).

TPhe        00:30:13.5 -46:30:15  J2000
RU149       07:24:15.5 -00:32:02  J2000
RU152       07:29:57.0 -02:05:17  J2000
PG0918+029  09:21:32.0 +02:47:02  J2000
PG1047+003  10:50:10.5 -00:01:15  J2000
PG1323-086  13:25:51.5 -08:50:15  J2000
PG1633+099  16:35:34.0 +09:46:22  J2000
MarkA       20:43:59.0 -10:46:40  J2000

You should not take photometric standards with an exposure time shorter than 5s, otherwise your photometry will be compromised by atmospheric scintillation.

6.4 In the Morning

If you didn't get all the bias frames or darks you needed you can start these running before you go to bed.

And finally, you should start writing last night's data. If you started some darks running, you can put the tape saving operation to sleep, so it will start after the data taking is finished.

7. Using prepared scripts

Unfortunately, these must be written in TCL, which is the worst programming language I've ever used (Microsoft BASIC on my TRS-80 was more convenient). Its a complete dog.

So that astronomers don't have to learn how to use TCL, we have tried to prepare a few pre-packaged scripts which can be straightforwardly modified to fulfil most observing requirements. These scripts are stored in /home/aatssx/wfieng in read only format. You should copy a script you'd like to use to a new name, make it writable, and then edit it.

AAO_offset.tcl    - a script to take an imaging data sequence, with offsetting and focussing.
AAO_std.tcl       - a script to take observations in multiple filters of std stars
AAO_linearity.tcl - a script to take a sequence of data for doing linearity and shutter timing calibrations.

cd /home/aatssx/wfieng
cp AAO_offset.tcl YOURNAME_offset.tcl
chmod +w YOURNAME_offset.tcl
nedit YOURNAME_offset.tcl (or whatever your favourite editor is)

The left one starts the file chooser to select a script. The right one runs the last script again.

7.1 AAO_offset.tcl

You can look at a copy of the script file itself (as of 24 Feb 2002) at AAO_offset.tcl to get a feel for what it looks like. However, the version in use at the telescope may have been updated, so bear that in mind. As described above,  you chould copy this file to your own version, chmod it to make it writable, then edit it. There are seven (7) locations you need to change.
 

#
# (1) !!!***!!!  Set number of exposures and exposure times.
#
   set times  { 10 20 30 }

#
# (2) !!!***!!!  Set offsets for each exposure
#
   set offra  { 0 r5  12 0 0 0 0 0 0 0}
   set offdec { 0 r10 r5 0 0 0 0 0 0}

#
# (3) !!!***!!!  Set filter for each exposure
#
   set filter { 2 1 3    }

#
# (4) !!!***!!!  Select type of offset.
#
# Uncomment this to offset cumulatively
#   set offstr "OFFSET"
#   set fitsk3 "CUMULATIVE OFFSET"
# Uncomment this to offset from the position the telescope last slewed to.
   set offstr "OFFSET/SL/WAIT"
   set fitsk3 "OFFSET FROM BASE"

#
# (5) !!!***!!!  Setup focus base for this run.
#
#  Absolute V focus measured 9-13 August 2001).
#   set vfocus 39.65
#  Absolute V focus measured 14 Feb 2002 ).
   set vfocus 39.15
#
# (6) !!!***!!!  Setup focus offsets for the loaded filters.
#
   set focoff { 0.65 0.65 0.0 0.0 0.25 0.15 }

#
# (7) !!!***!!!  Select Window.
#
# Uncomment for Full Mosaic
#   set window "DO_SAVE=1 REG_WINDOW=\"REG_1x1_8422x8282+0+0\" XO=0 YO=0 WIDTH=8422 HEIGHT=8282 NREGIONS=1 CCDMASK=65535 CAMERA_COORDS=1 RCF=1 CCF=1"
# Uncomment for Full Mosaic Binned by 2
   set window "DO_SAVE=1 REG_WINDOW=\"REG_2x2_8422x8282+0+0\" XO=0 YO=0 WIDTH=8422 HEIGHT=8282 NREGIONS=1 CCDMASK=65535 CAMERA_COORDS=1 RCF=2 CCF=2"
# Uncomment for single CCD6 (unbinned) only
#   set window "DO_SAVE=1 REG_WINDOW=\"REG_MASK_1x1_4\" XO=0 YO=0 WIDTH=2098 HEIGHT=4136 NREGIONS=1 CCDMASK=32 CAMERA_COORDS=0 RCF=1 CCF=1"
# Uncomment for testing - small windows, no save.
#   set window "DO_SAVE=0 REG_WINDOW=\"REG_CEN_CCD_1x1_512\" XO=793 YO=1812 WIDTH=512 HEIGHT=512 NREGIONS=1 CCDMASK=255 CAMERA_COORDS=0 RCF=1 CCF=1"

After changing all of these to the values you'll wnat bto use, you are stongly recommended to test your script out in the afternoon or evening before you start observing (do this by selecting the test window above, and some short exposure times). Then look at the Cicada Message window to check it is doing what you think you wnat it to do.
 

7.2 AAO_std.tcl

You can also create a script to put the same std in every CCD. Just insert the appropriate ra/dec offsets for the declination
# dec=0
set offra = { -475.2  -475.2  -475.2  -475.2   475.2   475.2   475.2   475.2 }
set offdec= { -726.7  -242.2   242.2   726.7  -726.7  -242.2   242.2   726.7 }
# dec=-20
set offra = { -503.7  -503.7  -503.7  -503.7   503.7   503.7   503.7   503.7 }
set offdec= { -726.7  -242.2   242.2   726.7  -726.7  -242.2   242.2   726.7 }
# dec=-40
set offra = { -622.5  -622.5  -622.5  -622.5   622.5   622.5   622.5   622.5 }
set offdec= { -726.7  -242.2   242.2   726.7  -726.7  -242.2   242.2   726.7 }
# dec=-60
set offra = { -950.4  -950.4  -950.4  -950.4   950.4   950.4   950.4   950.4 }
set offdec= { -726.7  -242.2   242.2   726.7  -726.7  -242.2   242.2   726.7 }
# dec=-70
set offra = {-1387.6 -1387.6 -1387.6 -1387.6  1387.6  1387.6  1387.6  1387.6 }
set offdec= { -726.7  -242.2   242.2   726.7  -726.7  -242.2   242.2   726.7 }
# dec=-80
set offra = {-2737.2 -2737.2 -2737.2 -2737.2  2737.2  2737.2  2737.2  2737.2 }
set offdec= { -726.7  -242.2   242.2   726.7  -726.7  -242.2   242.2   726.7 }
 

7.3 AAO_linearity.tcl (also use this to get data to check shutter timing)

This script is easy to use - decide whether you want to look at linearity (ie high count rates) or shutter timing (low count rates), and uncomment the relevant lines.

# This is the list of exposure times IN MILLISECONDS
# which will be performed. Choose between them for linearity or shutter timing.
#
# Linearity
#
set times { 0 0 0 0 0 10000 10000 10000 5000 7000 9000 10000 12000 \
            15000 18000 10000 20000 24000 28000 10000 30000 35000 38000 \
            10000 43000 47000 48000 10000 50000 53000 10000 58000 10000 \
            65000 70000 10000 10000 10000 0 0 0 0 0 }
#
# Shutter timing.
#
#set times { 0 0 0 0 0 10000 10000 10000 5000 7000 9000 10000 3000 \
#            2000 1000 10000 700 500 400 10000 300 200 100 10000 80 \
#            70 60 50 10000 10000 10000 0 0 0 }

1. Point the telescope at the domeflat screen (get the night assistant, afternoon technican or support astronomer to do trhis for you) and open the mirror covers and prime focus dust cover.
2. Turn on the domeflat lamp
3. Set the FILTER wheel to the filter of choice (gVRr are good for this) and adjust the flat field lamp, to give around 13,000 adu in10s, and around 50,000 in 50s.

The data acquired can be calibrated as a function of time using the 10s exposures, and then you can examine the adu per unit exposure time to determine the detector's linearity.
 

7.4 AAO_focus.tcl

#
# (1) !!!***!!!  Set exposure time for each image.
#
   set times  10
#
# (2) !!!***!!!  Set offsets on sky in arcsec & rough declination.
#                Suggest offset should be > 10
#
   set offsky  15
   set dec     -30
#
# (3) !!!***!!!  Set focus start, step, and number of exposures
#                7 steps at 0.1 spacing should work well.
#
   set focnexpo 7
   set focstart 39.15
   set focstep  0.1

The data file created will have focnexpo images, in a sequence staring from focus value focstart, with focus step focstep and spacing on the sky of offsky arcseconds (assuming you've put the declination in correctly!). The sequence will appear vertically on the display, with the first focus value first (and a double spatial step at the end). The script calculates the spacing in pixels as it will appear in the mosaic and puts all this information into the header (FITS keywords FOCNEXPO, FOCSTART, FOCSTEP, FOCSHIFT).

In principle this script should create images which can be processed by the IRAF task mscfocus. The first image should be at the 'top' of  a sequence, and the gap should be at the bottom. You should be able to use mscfocus to select the top image, and it will find the rest, fit to them and plot up FWHM information. MSCFOCUS may not expect a gap at the end (I can't tell because the NOAO Mosaic documentation is silent on exactly what the telescope does),  in which case the code to put a double step at the end should be removed.
 

8. Writing scripts

You have to write these in TCL. It sucks. Big time.

You can use the scripts above as examples, and should consult the relevant section of the RSAA CICADA Users Manual.

9. Guiding

The display  of this PC is mirrored via a "VNC viewer" protocol over the AAT's ethernet, to another PC in the control room (labelled the "WFI Autoguider Control" PC). This link is functionally quite slow, so don't go wild with mouse motions.

It is almost always necessary to adjust the telescope pointing slightly so that a guide star will appear in a guide CCD. This can he done with the GSFIND and AAT_TELESCOPE tasks of Cicada.

In the sections that follow we first discuss whether you need to guide, then describe the MaximDL software, and the GSFIND/AAT_TELESCOPE tasks. And finally, we provide a step-by-step guide to guiding.

9.1 Should I guide?

So there is no read-noise penalty in combining unguided 5 minute exposures. In fact, at R, i and z, many observers won't be able to exposure for longer than 450-600s anyway (especially if observing binned) since that's the time for the sky to reach near half-well depth.

You can also only guide (easily) with the telescope in the Reference Axis, so that GSFIND can understand the telescope co-ordinates.

The bottom line is that, except for U or narrow-band observing, there is no need to guide, and the overheads and pain guiding adds is not worthwhile.

9.2 The Multiplexer & MaximDL Software

Getting started.

• If necessary log into the WFI Autoguider Control PC in the AAT control room. The username and password are on the wall next to aatssx.
• Click on the "Shortcut to vnc viewer" icon on the desktop
• Check the IP number to connect to is correct (see wall next to aatssx)
• Use the password on the wall next to aatssx to connect.

the Multiplexer selector "WFI Guider Selector"
and "Maxim DL" .

Multiplexer

The only thing you can do in this window is select which of the guide CCD the SBIG controller talks to. There are two ways to do this. There is a pull down menu, and a graphical interface.

Maxim DL

We only want to use it to guide our telescope, so the only tabs in the Maxim CCD window you need to use at all are "Setup" and "Guide".

Setup
As shown above, the Setup panel should be used to tell MaximDL that both the CCD and Guider are controlled by SBIG ST-5C controller. This should already be the case, when you start, if not set both tabs to this value.

Guider
The Maxim software has no knowledge of how the AAT works (i.e. how X-Y on the image corresponds to ra,dec or the relevant scales). It simply sends signals to relays, which pretend they are using a guiding paddle to guide the AAT. So at the start of each night, you need to "calibrate" the guider (i.e. work out all the above parameters).  MaximDL does using its Calibrate function - if you put a star down a CCD and select Calibrate, the press Start. MaximDL will drive the AAT, and work out which directions are N-S-E-W etc., and what the appropriate scale factors are. Make sure you put in what the declination was for the calibration. You only need to do this once per run.

You can use this tab to either take individual exposures (for acquisition) when the Expose radio is selected, perform the calibration (when Calibrate is selected), or you can guide (when Track is selected). In each case you start by pressing the Start button (and can then stop guiding by pressing the Stop button). The exposure time used for each is set by the Exposure. You also need to set the Declination, when Guiding so the appropriate Cos(dec) factor can be applied.

9.3 GSFIND and the AAT_TELESCOPE task

Start GSFIND using "Tools"->"Start GSFIND". Then it is usually most convenient to move this window to another workspace using the "Occupy Workspace" function of the window manager.

Start AAT_TELSCOPE using "Instruments"->"AAT_TELESCOPE". This has to stay in the same workspace as cicada. When it appears press the "Start update" button.

AAT_TELSCOPE

The AAT_TELESCOPE panel has been discussed a little earlier. It provides a graphical interface for sending commands to the telescope (via the TEL_TASK). It is very important that this task, GSFIND and the AAT all have their co-ordinates in the same equinox. Check the TEL_TASK window on the X-terminal. The lines of positions should indicate whether the data is in 1950.0 or 2000.0 equinox. Then use "File"->"Preferences" in the Telescope Control window (left) to set this task's equinox to B1950.0 or J2000.0. Finally you must use the radio buttons in GSFIND to set its co-ordinates to B1950.0 or J2000.0 (see below). All three can be one or the other, but they must all be the same.

GSFIND

Once the AAT_TELESCOPE task is updating, the GSFIND task can be used. Before you do anything, make sure the equinox GSFIND is working in is the same as that used by the TEL_TASK and AAT_TELESCOPE (above). You can use the radios to choose between B1950.0 and J2000.0. Then, if you click on the button, GSFIND will load the current telescope co-ordinates, and display the GSC stars present in the region of the sky where WFI is currently pointed.

You can select the magnitude ranges in which GSFIND display targets, whether it displays galaxies, whether it displays a circle at the field centre.

You can also grab the WFI mosaic outline in the window and drag it around until a guide star lies in one of the CCDs. After which you'll see something like the image below. Notice the WFI mosaic is no longer centred on the original field centre (marked with the red cross hairs), and that the Field and Detector co-ordinates are now different.

Now if you go back to the AAT_TELESCOPE Telescope Control task, and click on "From GSFIND" the new mosaic co-ordinates will be loaded into the co-ordinates window of the panel, and highlighted.

Now if you press "Track" the telescope will be moved to the new co-ordinates, and you should be able to open the shutter and detect the guide star in the indicated guide CCD.

9.4 Guiding - Step-by-Step

1. Have the night assistant load a guide star rate of 1"/second into the AAT TCS.
2. At the start of the night either perform a guider calibration, or check that the manual calibration numbers above perform adequately.
3. Make sure the TEL_TASK, AAT_TELESCOPE and GSFIND tasks all provide/expect co-ordinates in the same equinox.
4. Slew the telescope to your next target.
5. Make sure the AAT_TELESCOPE task has the current telescope position (i.e. its updating).
6. Use the  button to load the AAT_TELESCOPE co-ordinates into GSFIND and display the available guide stars.
7. Drag the field around until a guide star is in a CCD.
8. Use the "From GSFIND" button in AAT_TELESCOPE to get the new mosaic co-ordinates, and select them.
9. Use the "Track" button in AAT_TELESCOPE to move the telescope to the new co-ordinates.
10. Use the Guide Star Selector on the PC to select the correct guide CCD.
11. Open the PFU shutter.
12. Select 'Expose' and press 'Start' to take a test exposure to ensure you can see the guide star, and select an appropriate exposure time.
13. Select the guider pixel position, using 'Auto' or by selecting a pixel in the image display.
14. Select 'Guide' and press 'Start' to start guiding and ensure guiding is working.
15. Once satisfied guiding has successfully started, press 'Stop', and close the PFU shutter.
16. Now start your exposure. Once the PFU shutter has re-opened, start guiding again. While MaximDL should recognize when the star vanishes and stop guiding, sometimes this doesn't work. If taking more than one exposure at the one guiding position, it is recommended you stop guiding when the shutter is closed.

---

10. Saving your data

WFI data files are large, so they can currently only be stored on tape. Although the Observatory will archive all WFI data onto tape and retain copies at the AAT, requests to extract data from the archive by observers who haven't bothered to make their own copies are not likely to be processed particularly rapidly.

Writing data to tape takes up a significant quantity of the aatssx resources. Do not write data and attempt to observe at the same time. You should aim to start your data tapes writing at the end of each night before you go to bed.
 

10.1 Formats

1. DLT4000 (Digital Linear Tape)
2. DLT7000 (Digital Linear Tape)
3. DDS4 (Digital Data Storage) version 4
4. DDS3 (Digital Data Storage) version 3

Smaller tape formats like DDS2 or CDROM are not useful. You can only fit 5 WFI images on a CDROM!
 

10.2 Drives

You should save your data to tape using the drives on aatssx. Using any of the other drives on other machines will just grind the entire network to a halt. The drives available are

• DLT7000 drive - will write DLT7000 tapes which won't read on a DLT4000 drive.

/dev/rmt/1n - DLT7000 (35Gb) uncompressed.
/dev/rmt/1un - DLT7000 (70Gb)  for 2:1 compression (see below)
• DDS4 drive- will write DDS2, DDS3 or DDS4 tapes.

/dev/rmt/0n - DDS4 (20Gb), DDS3 (12Gb) and DDS2 (8Gb) uncompressed.
/dev/rmt/0un - DDS4 (40Gb), DDS3 (24Gb) and DDS2 (16Gb) for 2:1 compression (see below)
(the drive will automatically recognise whether you've inserted a DDS2, DDS3 or DDS4 tape).

10.3 Writing a tape

cd /wfi_raw1/cicada     (or wherever your night's data is wfi_raw2,wfi_data1 etc.)
setenv TAPE /dev/rmt/1n    (or whichever of the drives above you wish to use)
tar cv 20020220       (or whatever your night's data directory is)

mt rew
tar tv

mt rewoffl

10.4 Writing more data to a tape which already has data on it.

Suppose you have already saved one nights data as a tar save set. You need to space forward by one file, and then write tonight's data.

cd /wfi_raw1/cicada     (or wherever your night's data is wfi_raw2,wfi_data1 etc.)
setenv TAPE /dev/rmt/1n    (or whichever of the drives above you wish to use)
mt fsf 1
tar cv 20020220       (or whatever your night's data directory is)

10.5 Delaying a tape write until observations have finished.

cd /wfi_raw1/cicada     (or wherever your night's data is wfi_raw2,wfi_data1 etc.)
setenv TAPE /dev/rmt/1n     (or whichever of the drives above you wish to use)
sleep 3600 ; tar cv 20020220  (where the sleep command waits for the indicated number of seconds)

10.6 Tape charges

11.Things that go wrong ...

Things that go wrong when I try to start Cicada/WFI

Unfortunately, this one is not as rare as we'd like. Usual problems are

Filter task failure
This is not a disaster, and usually means you haven't started the PFU task. Try starting it, and then using the AAT FILTER TASK to move the filter wheel. If this works the problem is fixed.

If this doesn't fix it, then you probably have a serial link problem between aatssx and aatpfu. Get your support astronomer or afternoon technician to help here.

Failure to communicate with controllers
If you've had a failure of an image read, it is quite common to not be able to re-establish communication with the controllers after killing Cicada and re-starting it. You need to reedit WFI, then Cicada, use cicada_cleanup, and start again. Please use patience here. It often takes several iterations. It may help to do the cleanup, start cicada, and then exit it without trying WFI, the cleanup again, start Cicada and then start WFI (anecdotally I have seen this help a few times).
 

Failure to communicate with ximtool
If you've had a failure of an image read & display, it is quite common to not be able to re-establish communication with the ximtool task after killing Cicada, and ximtool and re-starting it. You need to re-exit WFI, then Cicada, use cicada_cleanup, and start again. Please use patience here. It often takes several iterations. It may help to do the cleanup, start cicada, and then exit it without trying WFI, the cleanup again, start Cicada and then start WFI (anecdotally I have seen this help a few times). When trying to see whether communication with ximtool has re-started it will help to put in some tiny windows, and use "Display Only".

Things that go wrong when I try to take data

Readout of the mosaic does not complete
Communication with ximtool is lost

Both of these will usually result in nasty big red error boxes appearing. Unfortunately, both are not as uncommon as we'd like. In many cases, there will be no obvious cause for the problem, and your only fix will be to exit Cicada and restart. Sometime these problems do have causes. The most likely causes are

• Inadequate resources - you have too many windows, widgets or tasks running (like saving data to tape while observing). Try to keep your activities on aatssx to a minimum. For example, never run netscape on aatssx. Also if examining your data in a seperate IRAF window, just look at a single CCD at a time - don't attempt to look at the entire mosaic (ie use imexam, not mscexam).
• You have started another IRAF task and popped up an ximtool window, from the same account as is running Cicada. This is known to cause conflicts between the two ximtools. Never run IRAF and ximtool windows from the same account as is running Cicada. If you are running Cicada as wfieng, then do your data analysis as wfiobs.

Things that go wrong when WFI talks to the telescope

Occasionally the WFI_TEL_CONTROL task which runs on the VAX and sends commands back and forth between Cicada and the AAT TCS will die .... usually with error messages indicating that the VAX's access has been violated.

This is easy to fix. On the x-terminal where the WFI_TEL_CONTROL task was running (and all the error message now are) simply type

40a> wfi_tel_control

Then on aatssx, you need to flush the serial link between aatssx and the VAX, which is currently hung. To do this type

[wfieng]/home/aatssx/wfgieng: echo WHERE > /dev/ttya

After doing  this you whould see a 'WHERE' command and the telescope's repsonse to it on the xterminal where wfi_tel_control is running. Hey presto you are back in business.

Things that go wrong with PFU

The filters don't get read properly when I initialise the filter wheel.
Sometimes the "Initialise Filter" "Read Filter" will fail to read one or more barricades when the telescope is at prime focus access. Ask the support astronomer, afternoon technician or night assistant to move the telescope to the zenith and repeat the Initialise Filter (followed by the Read Filter).

When I try to move to a new filter the filter wheel fails to complete the move
PFU's filter wheel has proved to be remarkably robust. We've only come across two reasons why a filter move has failed

1. The building gas pressure, which operates the pneumatic detents on the filter wheel, is too low. This is the most common problem. If a wheel move fails this will almost always be the problem. See your support astronomer, night assistant or afternoon technician for help.
2. The filter is too thick (or the screws holding it in project above filter holder) and the wheel scrapes or jams. This is a mechanical problem that you can only diagnose by trying to manually move the filter wheel while the telescope is at prime focus access.

Things that go wrong with the focal plane temperature

The booster heater power supply is not working
A small booster power supply for the heater (on the left side of WFI) is required to raise the focal plane temperature from liquid nitrogen temperature to 183K (the mosaic is operated so warm because of the poor serial charge transfer efficiency of two of the CCDs). Sometimes the power cable on this power supply comes loose. Remove the cable and re-seat it.

12. Frequently Made Mistakes

• Absorting an Exposure. Sadly Abort rarely works. Well it does work, in the sense that your exposure will end. But Cicada will almost always be left in a state such that the next exposure fails, and you have to restart it. Never abort during a read, unless Cicada is in a state such that its going nowhere, you know you'll have to restart it anyway.
  
  The following notes are for stopping exposures while the shutter is open. The SSO 40" notes indicate that Pausing an exposure is itself as dangerous as Aborting. A few experiments at the AAT don't reveal it to be that dangerous, but you have been warned .... the safest situation will always be to let exposures run out .....
• If you have started a single exposure and you know it is incorrect, and want to stop it, you should
1. Pause the exposure.
2. Readout the exposure.
3. Let the data read out.
4. Decrement the run counter (if you want to overwrite the duff file you've just created). You'll get a yellow warning pop up on the next exposure to warn you about writing over the existing (presumably crap) data file.
• If you have started a series of exposures (ie Exposures: >1, not a script), you can't abort the series. However, what you can do is shorten all the exposures.
1. Pause the exposure.
2. Readout the exposure.
3. The remaining exposure in the sequence will read out immediately (ie be 0 seconds long).
4. Decrement the run counter (if you want to overwrite the duff file you've just created). You'll get a yellow warning pop up on the next exposure to warn you about writing over the existing (presumably crap) data file.
• If you have started a script that will take more than 10 minutes to complete
1. Pause the current exposure
2. Abort the exposure (ie only abort the exposure while the shutter is open, never during a read).
3. You will probably get nasty red windows saying the script has failed.
4. Before going on, make sure you take a short test exposure (small window, display only) to see if Cicada has survived and can take data. If not restart it.
• Trying to use Cicada to control the telescope, when the telescope is not running and in computer mode. The offset task used from cicada waits for the telescope to reach its requested offset (or slew). If the TCS is not in "computer" mode (or not running) it ignores the requested operation, so the task waits indefinitely. Eventually it times out, but well before then, cicada times out, and gets its knickers in a twist. You can end up having to restart cicada.
• Trying to start the AAO_offset.tcl script,  when the telescope is not running and in computer mode. This is the ame problem as above, since these script try to do a telescope offset.
• Running too many other processes on aatssx, and depriving Cicada of the resources it needs, causing it to crash during a read.  Don't write tapes, run netscape, or serve WFI images to other machines while observing.
• Having the run number come up as the last run from last night, instead of the first run from tonight. Sometimes you'll set the run number back to 1 at the start of the night, and then have Cicada fail you've actually created your first data file for tonight. Then when you exit Cicada and restart, it reverts to the last run number for last night. Always check the run number after starting up Cicada.
• Starting a science exposure, after taking a test exposure, and forgetting to turn Display Only off. Everyone does this once or twice I'm afraid. Just remember, don't abort the exposure. You'll only end up having to restart Cicada.
• Starting a science exposure, after taking a test exposure, and forgetting to turn windowing off. Everyone does this once or twice I'm afraid. Just remember, don't abort the exposure. You'll only end up having to restart Cicada.
• Taking single exposures and forgetting to adjust the telescope focus for the filters. Another 'duh' moment.
• Massively saturating a flat field, and thinking the numbers are 'real'. The WFI CCDs mostly reach full well, before reaching their 65535 analogue-to-digital conversion limit. This means that when the saturate, you can get very weird effects. In fact you can get 'smooth' looking images with quite low count rates. If you don't know what the flats are supposed to look like (see the WFI/PFU Performance Information page), you can think your getting flat field data, but its crap. To avoid this always do flat field test exposures by working up from faint lamp / light levels (if possible). And have a look at what the flat field images should look like.
• Forgetting to use an axis and so observing in the wrong place. If you're doing a survey program, you'll usually want to put your survey field centres in the centre of the mosaic (ie on the central gaps). This is where we define the telescope's Reference Axis to lie. If you're observing a specific object, you'll usually want to put your object in the middle of a CCD. So we usually define Axis A to lie in the middle of the bottom half of CCD6 for this. The telescope always drops into Reference Axis when it slews to a new position, so you have to remember to switch the Axis A after every slew when you want to put your target in the middle of a CCD.
• Thinking you have to guide. See the discussion above. Usually, you won;t need to.
• Writing your data to a tape you can't read. There are several formats you can write your data in. But you need to ensure you have access to whatever format you choose when you get home.

13. Frequently Asked Questions

• How do I abort an exposure? Try to never do this! If you have to interrupt an exposure, series or sequence see the discussion above.
• Do I need to dither? Dithering is usually done with WFI to 'fill in the gaps' between the CCDs. For survey programs, these gaps represent an insignificant fraction of the array area, so filling them in really is not important. Moreover, your reduction gets much more complicated. However, if you are observing a single field and want to make dark sky flats or fringe frames, or if want to make pretty pictures, then you will need to dither.


The AAO_offset.tcl script which is provided for sequencing science exposures will allow you to create a dither pattern, or allow you to tell it to create its own random dithers within a box you specify. Experience in the infrared has shown that random dithering works well and its means you don't need to create lots of different dithering patterns for lots of different numbers of dithers.

• Do I need to guide? See the discussion above. Usually, you won't need to.
• Should I bin? Binning (only x 2 is really of any use) will save you read-time, and when operated binned the Cicada system is a little more robust (it places less demands on aatssx's resources to read and write 32Mb images instead of 128Mb images). However, your sky per pixel will be four times brighter, so in i and z, your maximum exposure times will be shorter. And you'll have 0.46" pixels.
• What standards should I use? See the discussion above. We usually use Landolt standards for WFI observing, but if you want to calibrate your data onto the SDSS system, you'll need to track down some SDSS standards.
• Where can I take sky flats? See the list of blank fields below.
• Why does CCD1 look so dark in my bias frame? Is it dead? No, this is an effect of the adpative scaling used by the ximtool. Different grey scale limits are chosen for each CCD. Because CCD1 has a few very bright columns, it gets allocated quite different scaling limits to the other CCDs, and so looks much darker.
• What are all the various cosmetic features?
1. CCD1 has a few very bright columns on its right edge (data in this region is useless), and a generally 'washed out' appearance to its flat field.
2. CCD2 has a diagonal strip in its flat field on its lower half, and an appearance like a cross between an "oil film" and a tie die shirt. This patterns flatten well.
3. CCD3 has a trapping pixel site on its left edge
4. CCD4 has a pronounced brick wall pattern, which gets more severe at blue wavelengths. It also has an epoxy void in its centre, which shows up most clearly at i and z. These patterns flatten well.
5. CCD5 has a bit of speckling accross its surface.
6. CCD6 is cosmetically very nice.
7. CCD7 has a strip accorss its lower half like CCD2, and a "scratched up" appearance in its upper right corner.
8. CCD8 also has a diagonal strip, and some very odd speckling. Its most pronounced feature is a label reading CCID20 and a serial number (backwards) at its top edge (presumably from a slipped lithograhphic mask at the AR coating phase). These features flatten well.

Appendices

---

Appendix - Twilight Flat positions

The following table is of blank fields used at the European Southern Observatory, La Silla. Observers with new blank fields they would like to make available are asked to e-mail details to cgt@aaoepp.aao.gov.au.

But be careful - some of these blank fields are blank because they are the centres of dark clouds - so they are often not that blank in i or z, and with a field as large as WFI you often see lots of stars around the dark clouds.
 

Field	RA(1950)  h  m  s	DEC(1950)   d  m  s	Comments
SGP	00:48:30.3	-28:03:01	Contains ~ 10 objects in 6.4' x 6.4'
0102-265	01:02:48.5	-26:34:25	Contains ~ 15 objects in 6.4' x 6.4'
0248-196	02:47:46.6	-19:46:11
0427-36	04:27:26.1	-36:25:16	ok - 5' x 5'
9090-7	09:09:32.6	-07:38:26
CT1	10:04:27.0	-02:19:00
1101-264	11:02:23.5	-26:43:57
CT5	12:26:08.7	-06:38:29
CT2	12:54:58.7	-02:07:04
SER1	15:13:13.9	-00:31:48
2335-40	23:34:34.6	-39:49:04
2345+007	23:45:45.9	+00:40:40
SPKS 01	12:28:04.0	-07:46:54	A bright star in the field
SPKS 02	11:07:06.0	-77:05:54
SPKS 03	12:28:38.0	-63:28:18	Not that blank
SPKS 04	15:39:43.0	-33:56:54	No stars in 5' x 5' Few in 8'x16'
SPKS 05	16:02:43.0	-45:48:18	Few stars in 5' x 5'
SPKS 06	16:25:40.0	-25:09:48	Few stars in 5' x 5'
SPKS 07	17:07:47.0	-40:34:06
SPKS 08	17:31:06.0	-25:44:24
SPKS 09	18:02:01.0	-04:32:36
SPKS 10	19:00:48.0	-37:21:18
SPKS 11	13:04:33.0	+29:50:49
SPKS 12	16:49:42.0	-15:21:00	No stars
SPKS 13	19:19:09.0	+12:22:05
SPKS 14	21:26:54.4	-08:51:41
	10:55:49.7	-77:08:36

---

Appendix - Other bits of Cicada you might find useful

Cicada is a large piece of code and very complex. For full details on all its functionality consult the RSAA Cicada documentation. However here's a few odds and ends we've found we use.

GIT - "Tools"->"Start GIT"

GIT is the Graphical Image Tool. You can use it to display images, and do various things. I have always found it to be slow, cumbersome and useless. It also uses up a lot of resources.

"Options"->"Preferences"

General - discussed above. with regards to setting file name and location parameters. There are also options here to:
• indicate whether the large time counter in a separate window appears. Turn the toggle ``on'' to use the large counter in a separate window.
• show messages window at startup will cause the CICADA messages window to be displayed and positioned below the observing window when the first status messages are received from a CICADA sub-process (i.e. there may be a delay after the observing window has appeared before the messages window is seen).
• reset the Number of exposures field to 1 after running a sequence of exposures. e.g. (with this toggle ``on'') the observer enters ``5'' in Number of exposures and selects Expose. The five exposures are run and at the end of the last one, Number of exposures is automatically reset to 1.
• toggle binning when switching between Save and No-Save modes. e.g. the observer bins 4x4 while positioning the science object on the CCD and using Display Only mode. However, the observer uses 1x1 binning during data acquisition with Save and Display mode. With this toggle ``on'', CICADA will remember the two different binning factors and toggle the binning whenever the observer switches between Save and Display and Display Only modes.
• The number of CCD flushes to perform when doing an exposure.
• X window screen numbers to display each CICADA tool - only applicable for multi-screen workstations. By setting the screen number here, each of the tools will be started on the preferred screen.
• Automated focusing parameters - this does not work at the AAT so ignore these.

Image Display and Scaling - These affect the way in which image data is displayed as they are read out:
• Clear display before next image will cause the image display to be cleared before the next image is displayed. When not selected, a following image will gradually overwrite the currently displayed image. This may be useful in focussing or positioning the science object on the CCD.
• Highlight saturated pixels will cause any pixels with a value greater than that specified in the ``saturation value'' field to be highlighted in red on the image display.  Set this to 50,000 so that pixels at CCD full well will be flagged red.
• Display distinct image regions in separate frames,... This option affects the way images are displayed from CCD mosaics, multi-amplifier CCDs and multi-region readouts. If ``on'', each separate region is displayed in a separate frame of the image tool. If ``off'' all regions are displayed in one image tool frame in ``detector'' coordinates. Leave it off.
• Use approximate integer maths for scaling... Uses shift division for placing each pixel into its display bin (for Ximtool this is a maximum of 200 levels). This, if used in conjunction with the Ultra Sparc VIS instruction set (see General options), can speed up image display by nearly 4x for most images. Leave on (the numbers are not so good, but the speed is important).
• Adaptive scaling will calculate appropriate minimum and maximum image display values as each ``chunk'' of image is read out. Depending on the settings of ``% of image to readout before rescaling'' and ``Num SDs tolerance before rescaling'' the range is re-calculated after each chunk and the entire image redisplayed. Minimum and maximum values are calculated as a weighted deviation from the median or mode after sampling a proportion of the pixels. Better performance, but perhaps poorer scaling will result from decreasing the number of pixels to sample. The min/max range can be modified by altering the contrast adjustment value between 0 and 1.
• Fixed scaling allows the observer to specify the range of image display values to be used by the image display. This may be necessary in certain situations where adaptive scaling is not giving the desired results.


Sounds Preferences - These preferences control the number of beeps issued by Cicada at certain key points during an exposure. The observer can set the number of beeps to be heard when the shutter opens, the shutter closes and when readout finishes.

Expert - These preferences control the following:

• Whether or not the observer is warned at startup if CICADA cannot allocate a colour. CICADA uses a small number of colours on the GUI for highlighting buttons and warning of error conditions. If the observer is running a colour-hungry application such as Netscape, it is possible that CICADA may not be able to use one or more colours and so the software will try to find an alternative. If an alternative is needed or no colour can be found, a warning will be displayed at startup if this preference is ``on''.
• Write an observations log. When ``on'' CICADA will write a log of each image taken showing RA, Dec, run number, exposure time, sidereal time, object. Leave on
• Show timestamps in messages window. Whether or not to show a timestamp with every line in the CICADA Messages Window. Leave on
• Show Balloon Help. Turns on and off the ``balloon help'' on toolbar items. Leave On
• Show toolbar items as text, not pictures.
• Use the Ultra Sparc VIS instruction set. For improved performance in some operations. (eg byte swapping or image display). Leave On
• Write a temperature log. Selecting this option will allow a CCD temperature log be kept for the run. Leave On

"Options"->"Hardware Configurations ..."

These panels are used to change the way the WFI, PFU, Cicada etc hardware interacts. Don't touch it.

"Options"->"Exposure Time Unit"

Cicada exposure times must be integers. You can do exposures shorter than 1s by setting the exposure time unit to milliseconds (abbreviated by Cicada as "msec" - obviously a new non-SI unit of time). In general use, leave the exposure unit as seconds.

"Options"->"Show xxx"

Pop-up the Regions, Messages, Counter and FITS Keyword windows.

"SDSU"->"Readout Mode"->"FAST"

WFI is only setup to operate in the FAST speed. Do not change this from FAST.

"SDSU"->"Show temperatures"

Pop-up the temperatures window, from which you can control the WFI focal plane temperature. DO not change any of the parameters here.

"SDSU"->all the other options

Should not be used by WFI observers!

---


This document written by Chris Tinney 19 February 2002 onwards.