PDF Version

 

WIYN TIP-TILT MODULE

 

USER MANUAL

 

(KPNO SYSTEM)

 

Version: July 02, 2002

Chuck Claver and Dipankar Maitra

 

With contributions from

The rest of the world

 

 

The Ring Nebula

 

[send comments on manual to claver@noao.edu]

 

 

Overview: Quick Start

 

General Characteristics:

 

 

 

Arrays:

2048 x 4096 EEV CCD; thinned , engineering grade

Image size:

2048 x 2500 @ 16 bits, plus header, overscan:~10.58 Mbytes

Pixels size:

13.5-um ( ~ 0.1125"/pixel )

Read-noise:

~4.65 e-

DQE:

86% peak at 6500Å    [find this]

Dark-current:

~5 e-/pixel/hr [find this]

Read-out time:

2.6 minutes

Cosmetics:

Fair to good; about half a dozen bad columns; a dozen or so small 1-2 pixel bad areas.

Filters:

2" x 2"

Saturation:

Typically, linear to 0.1% to 100,000 e-

Gain:

 ~1.983 e-/ADU

  

WIYN Parameters:

Count Rates:

At UBVRI=20th mag: U: 35; B:330; V: 340; R: 410; I: 225 e-/sec [find this]

FOV:

3.8'x4.2', XIMTOOL Orientation: North- up, East-left

Scale:

0.1125"/pixel at center [determine field distortion]

Image quality:

PSF quite constant across the FOV, but will vary with distance from guide star less than 10%

Artifacts:

Bright stars will show a ghost plus ~100 pixels in x from the back surface of the beam splitter.

Typical focus:

 

ADC:

Not currently used

 

 

Data Acquisition: Acquisition commands are given on computer named Navajo -- Analysis commands are given also on the same computer.

 

 

All the Commands That Are Likely To Be Needed

 

 

Observing Commands (on wiyn-ccd)

Observe

take one or more exposures prompting for the exposure type

Doobs

a script which takes flats/objects for a list of filters

Mosdither

takes (typically 5) dithered images in a single filter to fill the gaps in array

More

take more exposures just like the last one

Test

take a test exposure. The output image, test, is overwritten each time

Object

take one or more object exposures

Zero

take one or more zero (bias) exposures

Dark

take one or more dark exposures

Dflat

take one or more dome flats

Sflat

take one or more sky flats

Focus

take a focus frame

Recover

recover data (if possible) following a crash during readout

 

Exposure Control Commands

Pause

pause exposure (e.g. clouds) [do not ABORT or STOP from within pause!]

resume

resume a paused exposure [then ABORT or STOP  if necessary!]

tchange

increase or decrease the exposure time

Stop

stop an exposure (and sequence of exposures) reading out the detector

Abort

abort an exposure (or sequence) discarding the data

pictitle

change the title of the picture

 

Quick-look and Taping Commands (driftwood/pecan)

mscdisplay

display an entire mosaic frame

mscexamine

general tool for examining images

mscwfits

write mosaic frames to tape in multi-extension FITS format

 

Caution :Tape your data as you go; DLT-7000 (~250 images/tape) and Exabyte drives (~35 images/tape) are available. Write time: ~75 seconds/image to DLT, ~3 minutes/image to Exabyte. DDS-4 DAT drive available on Pecan (4m); write time ~40s/image. Please be off the computer by noon of your last day!

 

You can buy DLT and Exabyte tapes on the Kitt Peak, but bringing your own is cheaper.

  

Calibration data:

      Take dome flats or, preferably, twilight flats (night sky flats work even better!).

      Take dark exposures of similar length to your science exposures.

      Take zeroes (i.e., biases) -- Darken the dome for darks and zeroes!

 

1.      Introduction

 

The WIYN Tip-Tilt Module (WTTM) is attached at the Nasmyth focus of the 3.5m WIYN telescope.  Physically the module is installed at the “WIYN” port, which also houses the Mini-Mosaic Camera. The design permits rapid change from wide-field CCD imaging and multi-object spectroscopy to higher resolution imaging over a 5 arcmin field of view and integral field spectroscopy as science objective and atmospheric seeing dictates. As a result of the active optics system already installed on WIYN, control of the local environment and location of the telescope, the image degradation is primarily a result of upper atmospheric turbulence of which measurements indicate that for an aperture of this size image motion is a major contributor. The WIYN tip-tilt system corrects this motion in real time and has the potential of reducing median seeing of 0.8" to 0.5" at R and produce nearly diffraction limited images at H.

 

2.      WTTM Instrument Description

2.1.    Optical System

2.1.1.    Principles of Design

2.1.2.    The Optical Path

2.2.    Error Sensor

2.2.1.    Design and Optical layout

2.2.2.    How it works

2.2.3.    Sensing Focus

2.2.4.    APD Safety Issues

2.2.5.    Performance Metrics

2.3.    Science Camera

2.3.1.    System Description

2.3.2.    System Performance and Specifications

2.3.3.    FOV: Why it looks the way it does

2.3.4.    Beam-splitters for Science

3.      Software Control System

The WTTM software system operates under the Linux operating system, currently Redhat v6.2 on a dual 500mhz Pentium III CPU machine.  The computer chassis is located under the telescope azimuth skirt and is named wiyn-wttm.  Also located below the wiyn-wttm chassis is the error sensor X-Y stage control chasis.

3.1.    Real Time Linux

3.2.    Command Line Interface (cli)

            The WTTM has an extensive command line interface as documented in the appendices. These notes are extracted directly from the self-documenting code. Note that most commands accept a ‘-h’ parameter to provide a single line of help text.

All such commands begin with the ‘wttm’ identifier and usually match their function. For

example, to change the task frequencies the command is ‘% wttmSetTaskFrequencies’. Clearly, this amounts to a lot of typing so soft links are also created which usually follow the form of the command. Thus, ‘% wttmSetTaskFrequencies’ becomes ‘% wstf’ using the soft link. If in doubt, use the ‘-h’ parameter to get a one-line help text on any particular command or use ‘% wttm help’ to get a screen full of commands at a glance.

The cli commands are issued by the observer logged in the computer “wiyn-wttm” from the terminal of “Almond”. For a greater detail on the cli commands (from a programmers point of view) refer to “WTTM Users Guide” by P.N.Daly (this should be in the WIYN Control Room). However all the working of WTTM can also be controlled with user friendly GUIs developed using Labview as described below.

3.3.    Labview Interfaces

3.3.1.    WttmControlVI

This panel allows the user to send the main 9 elements to the real-time core. When these parameter values are set, the button labeled SET PARAMETERS must be pressed once to send the values to the core (otherwise they are not sent). Remember that some values can be adjusted whilst the system is in closed-loop mode and some cannot.

APD Frequency: This value sets the frequency in Hertz for the reading of the 4 APD counters and the computation of the 3 error signals, X, Y and Focus.  The system must be in idle (e.g. paused from pressing the “stop acquisition” on this GUI) in order for this value to be sent to the real-time Linux core.

APD Update: This value sets the frequency in Hertz that the APD values in the WttmGetTaskValuesVI panel are updated (more on the consequences relating to this value later).

DIO Frequency: This value sets the frequency in Hertz that the DIO task updates the position of the tip-tilt mirror via the Physique-Instrumente control chassis.  This is normally set to the same frequency as the APD Frequency.

DIO Update: This value sets the frequency in Hertz that the DIO values in the WttmGetTaskValuesVI panel are updated.  Typically this is set to the same value as the APD Update frequency.

Focus Interval: This value sets the interval in Seconds over which the focus signal is averaged to determine the incremental focus corrections once the focus control loop is locked. 

Guider Interval: This value sets the interval in Seconds over which the commanded tip-tilt mirror position is averaged in order to determine incremental X and Y guider corrections.

X and Y milligain: These values set the relative conversion factor between computed X and Y errors and the additive offset in volts applied to the current mirror tip-tilt position that generates the next commanded tip-tilt mirror position.

Z milligain: This value sets the relative conversion between the computed focus error and the additive offset in microns applied to the telescope secondary position by way of the WIYN router to the SecTilt client

REPORT PARAMETERS is a debugging aid and returns nothing of interest to the average user. SAVE PARAMETERS and RESTORE PARAMETERS save and recall these 9 items to the file /home/wttm/development/.wttmrc.

The START ACQUISITION and STOP ACQUISITION buttons are analogous to the pause and resume functions in the cli. Indeed, you can still use the ‘% wcli pause’ and ‘% wcli resume’ should you so wish.

The EXIT button in this panel is the main button that sets wttmExit to TRUE and, thus, terminates all other GUIs. Note that terminate here means ‘stops the GUIs running’ and does not imply ‘removes the GUIs from the screen’. The GUIs should be removed by the ‘% wlv stop’ command.

 

Figure 1: wttmControlVI Panel

 

3.3.2.    WttmGetTaskValuesVI

This panel is the main system monitor and shows the X, Y and Z errors as well as APD counts for each channel. The following items can be identified (from the bottom up):

LOGGING? If checked, all incoming values are appended to taskVI.dat in the data sub-directory of /home/wttm/development. These files can grow quite large and should be expunged regularly.

DIO COUNT is the number of times the DIO task has been through its programming loop.

DIO HW ERROR is a flag to indicate a DIO hardware violation detected by the real-time core. A zero value indicates no error, a non-zero value is undesireable.

DIO HW TIME is the typical time take to access DIO hardware from the real-time core expressed in microseconds.

DIO HW TIME is the typical time take to complete 1 complete programming loop (for the DIO task) in the real-time core.

APD COUNT is the number of times the APD task has been through its programming loop.

APD HW ERROR is a flag to indicate a APD hardware violation detected by the real-time core. A zero value indicates no error, a non-zero value is undesireable.

APD HW TIME is the typical time take to access APD hardware from the real-time core expressed in microseconds.

APD HW TIME is the typical time take to complete 1 complete programming loop (for the APD task) in the real-time core.

APD COUNTS is a graphical representation of the APD counts as a snapshot taken every APD update rate in Hertz. The 4 APDs are colour coded (0=yellow, 1=green,2=pink,3=purple) and explicit values are also shown in the color-keyed indicator boxes at the lower right of the graph.

X is the main graphical representation of X errors. The average error is shown in the upper panel and the variance in the lower panel. The variance is 0.0 (and hence undefined) when the number of available data points is <2. Data displayed in this graph can be changed using the ring buffer widget called XYZ Graph.

Y is the main graphical representation of Y errors. The average error is shown in the upper panel and the variance in the lower panel. The variance is 0.0 (and hence undefined) when the number of available data points is <2. Data displayed in this graph can be changed using the ring buffer widget called XYZ Graph.

Z is the main graphical representation of Z errors. The average error is shown in the upper panel and the variance in the lower panel. The variance is 0.0 (and hence undefined) when the number of available data points is <2. Data displayed in this graph can be changed using the ring buffer widget called XYZ Graph.

XYZ GRAPH controls what is displayed in the X, Y and Z graphs above. There are 4 options:

Raw APD Errors are snapshot values at the cadence of the task update rate. For example, if the APD update rate is 20 Hz and this option is selected then snapshot (raw) values are placed on the buffer by the real-time core every 1/20 s and these values are displayed in the X, Y and Z graphs.

Raw DIO Errors are snapshot values at the cadence of the task update rate [looks same as above !!! which one is correct ????]. For example, if the DIO update rate is 10 Hz and this option is selected then snapshot (raw)  values are placed on the buffer by the real-time core every 1/10 s and these values are displayed in the X, Y and Z graphs.

Accumulated APD Errors are average data at the cadence of the task update rate. For example, if the APD task frequency is 1000 Hz and the APD update rate is 20 Hz then 1000 / 20 = 50 data points are averaged and the result (and variance) displayed via the X, Y and Z graphs.

Accumulated DIO Errors are average data at the cadence of the task update rate. For example, if the DIO task frequency is 500 Hz and the DIO update rate is 25 Hz then 500/25 = 20 data points are averaged and the result (and variance) displayed via the X, Y and Z graphs. It gets more complicated than that, however, if the DIO task is not running at the same frequency as the APD task as the DIO task can accumulate APD values—check the source code for an explanation!

 

Figure 2: wttmGetTaskValuesVI Panel

3.3.3.    WttmGuiderVI

This panel shows the value of the guide signals sent to the telescope control system. It can be restarted by clicking the (local) exit button and then clicking on the usual LabVIEW run arrow. The upper panels show the X and Y guide adjustments and the lower panel the associated variance (if applicable) all in seconds of arc. The ‘Logging?’ options appends incoming data to guiderVI.dat in the data sub-directory of /home/wttm/development. If the value of guideInterval is zero (which it is by default) no data is sent to the telescope control system and this GUI remains inactive.

Figure 3: wttmGuiderVI, the guider panel

3.3.4.    WttmFocusVI

This panel shows the value of the focus signal sent to the secondary control system. It can be restarted by clicking the (local) exit button and then clicking on the usual LabVIEW run arrow. The panels on the left show the incoming raw data value (upper) and variance (lower) if applicable. The right hand side panel shows the effective target (upper) taking into account the desired Z target (if set) and the desired Z offset (if set). These values are set with the cli commands ‘% wszt -t<val>’ and ‘% wszo -o<val>’. The lower panel shows the adjustment in microns taking into account the conversion factor to microns and the Z-axis gain. The gain can be set with the cli command ‘% wsg _z<val>’.

The right hand side panel is activated by use of the lock facility. When unlocked, the graph is disabled. When locked, the graph is enabled (assuming the focusInterval is non-zero too). The lock can be set with the cli command ‘% wszl _l<val>’ where a zero value indicates unlock and a non-zero value indicates lock.

The ‘Logging?’ options appends incoming data to focusVI.dat in the data sub-directory of /home/wttm/development. If the value of focusInterval is zero (which it is by default) no data is sent to the secondary control system and this GUI remains inactive.

Figure 4: wttmFocusVI, the focus panel

3.3.5.    XyManVI: Error Sensor position control

This GUI controls the position of the error sensor ans is used to move the errors sensor to the location f the desired guide star. The current x and y positions of the error sensor is shown at the top of the panel. The user can input a pair of coordinates in the middle box and hit the GO button to send the error sensor at the requested position. The user can also ‘jog’ around some position by entering the amount to jog in the input box (2units = 1 micron, physically) and then clicking on one of the four directional arrows.

 

Figure 5: xyManVI panel, as it comes up. It comtrols the position of the error sensor.

 

Figure 6: xyManVI panel after resizing the window.

3.3.6.    WttmPublish

This panel shows the published data stream items that have changed since WTTM. This panel can be restarted ‘% wttm publish restart’.

 

Figure 7: wttmPublish Panel

 

4.      Setting-up and Obtaining Data

4.1.    Initial start-up tasks

4.1.1.    Starting the Software

Full operation of the WTTM requires the observer to use two computers viz. Almond and Navajo.  Almond is the computer used for remote observing and serves only as a display console for the WTTM computer wiyn-wttm.  Navajo serves as the data acquisition computer that runs the HARCON CCD control system.

4.1.1.1.    Configuring Almond/wiyn-wttm

The first task to do is to start up the Tip-tilt software on wiyn-wttm with its display being sent to Almond.  Start by logging onto to Almond as “observer”. The password should be the normal mountain user password and should be posted on the monitor. Once logged in, open a new terminal by clicking the right mouse anywhere on the screen and choosing the “open new terminal” menu. In this terminal window, type

almond> xhost wiyn-wttm

This allows Almond to accept X-windows from wiyn-wttm.

almond> rlogin wiyn-wttm –l root [this will be changed]

wiyn-wttm> export DISPLAY=”Almond:1”

This is a bash shell command that tell wiyn-wttm top send all X-window display from this session to Almond’s screen number 1.  Note:  On Almond the default login screen is the 24 bit display and is referred to as Almond:1, where the alternate 8 bit display is referred to as Almond:0.  At this point Almond and wiyn-wttm should be configured to start WTTM software.

4.1.1.2.    Starting WTTM software on wiyn-wttm

In the wiyn-wttm terminal window type

wiyn-wttm> wlv start

This will start the WTTM Labview software which enables the user to control the operation of the tip-tilt module. A few messages will appear on the terminal, followed by opening of five GUIs with titles WttmPublish, WttmControlVI, WttmGetTaskValuesVI, WttmGuiderVI and WttmFocusVI. The WttmControlVI  window is the most important of all these because it is through this GUI that user feeds all the inputs to control the WTTM. The rest are mainly for visualizing the outputs and checking for erroneous behaviour (if any).  The user may want to minimize the WttmPublish as it is the least needed (in fact not at all if things are going fine!). The other windows may also be moved and resized to use the monitor screen efficiently. Now type

wiyn-wttm>xyManVI &

to start the GUI to position the error sensor. Fig.5 shows the window that comes up. However the user may resize it to resemble Fig.6 without any loss in efficiency. This completes the initial startup process for the WTTM and now the observer may proceed to setup Navajo, the computer used for data collection.

Configuring Navajo

The data acquisition takes place in Navajo. Login as “wiyn_ccd”. The password should be posted on the monitor. Along with various ARCON windows, an IRAF Data Acquisition Window and another IRAF Data Reduction Window opens up. An Ximtool window also opens up which shows the most recently acquired image. Right click and (Re)start ARCON if it is not already running.

4.1.2.    Initial Configuration

In the WttmControlVI  set the APD/DIO frequencies and APD/DIO update rates as shown in Fig.1. But keep the X,Y and Z milligains to zero, to keep the loop open. Hit the Set Parameters button and then Start Acquisition. The user will notice that the APD counter in wttmGetTaskValuesVI becomes alive. However the counts in all 4 APDs is zero because during the startup, the power in the APDs is off by default (to increase the longevity of the APDs).  To turn the power in the APDs on, issue the command:

wiyn-wttm>wttm_pwr apd on .

Similarly wiyn-wttm>wttm_pwr apd off turns power in the APDs. When the power in the APDs is turned on, a brief surge of counts from each APD is seen[why ?]. After this brief surge, the counts will come down to ~2 or 3 counts for each APD. This ensures that the system is working properly.

 

4.1.3.    Calibration Data

Before/after the science exposures, the observer should take some calibration data for the CCD. This invloves taking dome/sky flats. To take the dome flats you will have to tell the observing assistant first that you intend to do so. Once you get his green signal you should right click anywhere on the desktop on Navajo and select the Domeflat menu. Once the GUI comes up[need a screenshot of that], turn the high lamp on at your selected intensity (Table 1 may be a useful starting point). Once the lamps are on you can actually see them in the TV in the control room. You may use either the observe command or the dflat/sflat and zero command to take dome/sky-flats and the bias frames respectively. Usually 5-10 biases every night, and five dome flats through each filter (aiming for a count of ~20,000 ADUs in each) should be good enough. This should flatten your data to better than 1%.

 

Table 1: Dome-Flat Lamp Settings and Exposure Times for different Filters

FILTER

HIGH LAMP INTENSITY

EXPOSURE (s)

Approximate Count (ADUs)

Harris V

2000

4

20,300

Harris R

1500

3

20,600

Harris I

1000

5

20,000

Gunn r

1800

4

20,800

Gunn I

1100

5

20,500

Gunn z

1200

4

22,300

 

1.1.    Getting on Sky

1.1.1.    Initial Focus

After you get to your field, take a quick snapshot of the field, by typing in the data acquisition window

cl> test

Do a quick focus sequence. When you take a focus frame with the ARCON software at WIYN, you typically take a short (3-10 sec) exposure of a 11-12th mag star, clock the charge down 30 rows, decrease the focus value, take another exposure, clock down the charge, decrease the focus, etc., for a series of 7-9 exposures. The frame is then read down, the the image examined with mscexam/mscfocus to determine the best focus value. Note that the double­space gap occurs after the first exposure. A sample run is shown in Fig. 6.

 

1.1.2.    Acquiring your field [details of “observe” command ? already there in MINIMO manual]

Figure 8: A focus run centered at 4800 and step size -20

1.2.    Selecting a guide star

Select a fairly bright star (mv between 10-15)near you object of interest. Try to make sure that the star you chose is not variable [why?]. Type mscexamine on the IRAF Data Reduction Window. The cursor turns to blink on the Ximtool window. Position the cursor on the star and

 

type a to get its coordinate written on the data reduction window. To get the error sensor coordinates from CCD coordinates (obtained from mscexam), one needs to go to the IRAF Data Reduction Window, or open a new IRAF xgterm on Navajo and type the following:

cl> immatch

cl> lpar geoxytran

The parameters for geoxytran should look like

 

        input = "STDIN"         In