This Cookbook gives a step-by-step guide to operation of PHARO at
the Palomar 200-inch Telescope. For a description of the instrument
and information needed for advance planning of an observing run, go
back to the PALAO
homepage and click on the Overview button. As of this update (2008)
an observing run with AO divides the labor in the following way. The
telescope operator is responsible for the telescope and Natural Guide
Star (NGS) operation. A visiting engineer is responsible for the
Laser Guide Star (LGS) system. The observer is responsible for, in
this case, the PHARO the science camera. Use the web pages as an
introduction, and the support engineer will give you instruction on
the system to fill in the blanks for PHARO when you get to Palomar.
The AO manual is intended as a reference for AO operation, and may be
read by observers interested. However the observer’s responsibility
Despite its intended real-time usefulness, the Cookbook contains
nuggets of instrument theory here and there that might require some
time to digest...so of course a thorough reading in advance of the
observing run is recommended to avoid unpleasant surprises at the
telescope. (Failing this, telescope operators carry extra Kleenex).
Get to the telescope no later than 2:00 pm on the afternoon
of your first night to make contact with the set-up crew, and the
support engineer. The set-up crew goes off at 4pm, and the support
engineer is on from noon to 10pm. The support engineers share an
office just off the control room for the 200-inch.
The PHARO computer is Ezra2. Ezra2 is old and slow. A PC
replacement is in the works, but when it will become available is
unknown (as of 4/2008). The data directories for Ezra2 are /scr1
and /scr2. The
data directories hold 150 Gb, each full array PHARO differenced image
is 4 Mb. DO NOT take images while sftp’ing data. Try to sftp
during the daytime hours while you sleep.
The day crew will handle all liquid nitrogen fills for the
PHARO dewar. The PHARO dewar holds for 24 hours. If you are
interested, ask one of the daycrew to show you how to check PHARO's
temperature. This is a quick and simple way to check that the dewar
Control is the AO control computer.
The daycrew may already have this up and running when you arrive.
Initial system alignment and
checkout tasks will be carried out by the support engineer, and are
listed in the palaosetup html. Go to PALAOSETUP
if you are interested...
Aligning the pupil and determining the centroid offsets, are
also undertaken by the support engineer in the afternoon. These
tasks are very important for optimzing the AO correction. If there
is time, it useful (but not critical) to see how this is done.
An excellent time to review the PHARO GUI is after the system
is tuned up, and before you go to dinner. Have the support engineer
show you all the PHARO functions. Controls include; camera,
dithering, filters, slits, apertures, masks, nods, image
diagnostics, and error display. All these will be covered below.
There is also the link to the Cornell PHARO page on the
Obtain (PHARO) dark(s): use the internal PHARO dewar
block...i.e., set the Lyot filter wheel to BLOCK. It's best to have
darks for each integration time you plan to use during the run. Take
a few to be safe and perhaps to allow median filtering.
Then go to dinner.
Open some xwindow and check disk space is adequate via df
-k command; if not, scrape the /scr1 or /scr2 disk.
AO should be up and running before you start PHARO (or else
restart PHARO later) to get proper communication with TCS and
hence proper display of telescope quantities (RA, Dec, etc) in the
window at lower left on PHARO's screen. Check with the support
engineer for information on PALAO current status.
Run the program via xPHARO on a terminal window
prompt. You will be prompted for the subdirectory into which data
should be written, and the names of the observers
You are pretty much free to crash and restart xPHARO any
time; you'll just lose a couple of setups on the PHARO screen
(e.g. the green crosshairs; the display settings). You should have
the Lyot wheel BLOCKED before you crash/reenter, for safety.
Inspect the small message window that shows beneath the
PHARO GUI...be sure that current telescope positions are being
displayed, instead of just zeroes, and exit/reenter xPHARO if
there's a problem.
Take a "quick" image or two (not saved to
disk)...even with Lyot wheel blocked, you should see a little
Here is a snapshot of the lefthand monitor on Ezra2. After you
start xPHARO, and fill in the data path pop-up this is what you
Left hand monitor.
The lower left window is the xPHARO camera GUI. In this
window you will control exposures, filters, grisms, apertures,
fields of view, and macros. At the bottom of this window is an
ever changing camera status window. This is the single most
important window, there will be much more to say about this GUI
which is actually an analog layout of the PHARO optical path.
The upper left window partially hidden behind the xPHARO
Plot window, is a console status window. This window gives status
updates on mechanical movement within the PHARO dewar. It is
especially valuable for trouble shooting mechanical problems.
The xPHARO plot window is an excellent tool for determining
image quality. Used in conjunction with the tools in the xPHARO
display window, xPHARO plot can give information on, image Strehl,
uncorrected seeing, and corrected PSF/FWHM.
The right side window is, not surprisingly the xPHARO
display window. This window gives detailed information on the
entire PHARO image with it's capability to "zoom in" on
observer designated features.
As mentioned in item #1, there will be more to say about
each of these windows. It is also wise to look at the PHARO
manual, which has a link on the PALAO
The righthand monitor
Right hand monitor.
This monitor display is dominated
by the full 4 quadrant PHARO image.
On the left are 4 display and
At the top is a display that shows
the type of image (eg. Reference), what buffer is being displayed
(eg. Background), the image max/min, the "green" box
max/min, and the green box stretch mode (eg. Zoom Box).
The next box down selects the size
in pixels of the green zoom box displayed on the PHARO image (eg. 128
The next box down selects green
crosshairs and lines that are used to mark slits, spots, and target
positions. Simply click on A or B and a crosshair or line will appear
on screen. To move the fiducial markers around, all you need to do is
click and drag. The Col and Row indicators display marker column and
row pixel position on the image.
The final box at the bottom, which
is blank in this example, sometimes indicates cardinal image
direction. At the moment this is a buggy feature which is being
looked at. Re-starting PHARO sometimes restores this feature.
A section will be added to this
document at a later date which goes into more detail on GUI
selections, but for the moment refer to the above images when this
text refers to GUI commands. Also check the PALAO
link to the Cornell PHARO manual.
Obtain (sky or dome) PHARO flats
for each filter/platescale combo you'll use. For sky flats:
Use the chart of detector counts vs. time (after
sunset)/(before sunrise), found in the blue PHARO manual, as a
check. You'll have to be set up on the sky (PALAO on, open to sky,
just a flat DM) monitoring counts at least 15 minutes before
sunset for broadband (J,H,K,K',Ks); narrowband require being on an
hour or two earlier (check with crew regarding the rules of
Work fast; suitable sky brightness lasts only 20 or 30
Cycle through the suite of filters you'll do science with.
For each filter, get about 5 different 10-second exposures
that span a good range (a few tens of thousands) of counts as the
sun rises/sets. For flats through neutral-density filters, you
may have to make do with fewer counts.
Point the telescope at zenith and let Earth's rotation smear
out any stray stars in your way.
Remember that narrow-band filters may need to be done well
before sunset or after sunrise, and at longer integration times,
to get enough counts, brighter sky is preferable to longer
integration to get the counts (to avoid those stray stars).
Flats are generally believed to be constant over an observing
run, so long as nothing in the optical setup is changed.
Once you are ready do some science, the telescope operator will
want to look at an SAO star near zenith to check and set up several
things. The telescope operator will want to check pointing, get a rough
focus, and get an uncorrected seeing value from PHARO. To get the
uncorrected seeing expose on the SAO star for 30 to 60 seconds with
no AO correction. The support engineer will show you the tools for
estimating the seeing on the PHARO GUI.
When you are done with the SAO star set the Lyot
wheel to Block. It is required that the observer do this before every
After the PALAO console has been used to achieve a good lock on
the new guide/target star, most of the action occurs at the PHARO
console. Small (5″) closed-loop dithers and even
locking/unlocking the loops can be done remotely from PHARO.
Be patient, the PHARO GUI will only accept one command at a
time, and that command must complete before initiating anything new.
If you try to make a new command before a previous command has
finished, you will get an annoying error message.
DO NOT MOVE THE Shutter (located in the middle of
the PHARO GUI)...just leave it open. It could get stuck in the
closed position if moved, requiring days to fix.
ALWAYS Block PHARO's internal Lyot
stop during large telescope moves to avoid frying science CCD.
Turn off Continuous Acquisition (Cont Acq) mode before
taking an image or moving a motor.
Be very careful when using the tweak option on a motor's pull
down menu. It is possible to overwrite the motor position default
calibration file. It is better to home the motor and try the move
again. Putting the Carousel wheel in pupil mode can help
determine if a wheel has not moved correctly. Ask the support
engineer to give you an in depth education on the tweak function,
and its pitfalls.
Make small (no more than 5 arcsecond) dithers of the
telescope while loops are locked (not really for safety, but to keep
tip/tilt and SSM mirrors happy and loops locked).
There are two set of image buffers: "regular"
and quick, you take an image into one or the other with
the upper/lower rows of image-taking buttons, and you may
choose their integration times independently in the white rectangles
immediately above those. ["Regular" is not called
that...it is the row Take Src & Write Diff etc].
Quick is always a single image, and not written to disk. You
use it to scout out the situation, perhaps to check the field or
exposure time. "Regular" will take as many images as you
choose in the #Cycles box, and write them all to disk.
Use this option for your science observing.
Within those divisions, there are two more kinds of image
buffers: source/background...these are addressed with the
left/right columns of image-taking buttons, and are referred
to in the display window as Src and Bgd. If you obtain a
background on dark sky, you may display the more sensitive
Any image, regular or quick, source or background, is made up
of two readouts of the CCD chip, called "signal" and
"reference". Except in engineering modes, you will
generally want to work only with the "difference" of
REMEMBER: for objects
fainter than about 10th mag at V (current limit of PALAO's
acquisition camera), PHARO's continuous-acquisition imaging mode is
a powerful acquisition tool ...and it has a wider field of view
than PALAO's acquisition camera...
At start of setup, make sure Lyot wheel is set
to Block (as with most selections, a pull-down menu
Select plate scale (25 mas or 40 mas pixels) by selecting
25/40 arcsec field on Slit wheel, and 25/40 mas on Carousel.
Select spectral and (optionally) neutral-density filters from
the assortment offered on the Filter and Grism wheels.
The Shutter wheel should be set to open
(DON'T MESS WITH IT...it may get stuck). The shutter is intended for
future automated use with the Laser guide star upgrade.
Enter an integration time, in milliseconds, in the Quick
Time box (2000 = 2 seconds is the shortest for the full
array)...the box will alter it slightly while accepting it, to fit
chip readout specs.
When setup is all ready, move the Lyot wheel to
the setting of choice, most likely Standard Cross, which
has the basic pupil-plane (Lyot) mask designed to cover the aperture
and spiders of the telescope.
Hit the Quick Src button, and in a few seconds an
image should be displayed.
Hit the Cont Acq
button if you wish, to continually obtain new images (none saved);
remember to press it again to exit before doing other things.
The PHARO array is a four quadrant HgCdTe chip,
where each chip is 512 × 512. The CCD optimum linearity is up to
PHARO pixels are readout one row at a time.
Then a reset is applied to all the pixels in the row. A timing
discontinuity exists on chip quadrant borders, so it is wise not to
put targets on the borders.
On long dark exposures, you will notice a
circular glow at the four quadrant read points. This is multiplexer
FET glow, which increases with exposure time.
There is no shutter on this array. When idle,
the array continuously clocks and resets to a preset bias level. It
is always recommended that the block on the Lyot wheel be inserted
into the path when slewing the telescope, or when PHARO will be
idle for a significant length of time (THIS IS YOUR 4th
The bias reset is noisy. Correlated double
sampling is the technique used to remove reset noise.
The default PHARO image setup sends a
difference image to the display. This image is the difference of
the target signal minus the initial reference signal.
All exposures are sent to either the source
(src) or background (bgd) buffer. Quick exposures are for target
acquisition, exposure determination, and AO peaking. Quick
exposures are not saved to disk, unless the write FITS
button is pushed. Take & Write exposures are written to disk as
standard three dimensional FITS files (X, Y, quadrant#). No
coadding, or image subtraction is done to the images before writing
to disk. The S-B display button on the GUI DOES NOT
create a subtracted file of the source minus background files. The
displayed subtracted image is for display only, unless you use the
write FITS button.
Clicking the Detector Setup button
opens up the "Settings" popup window. In the popup, the
quadrant size can be changed, endpoints can be added, and what gets
saved to file can be expanded. There are no binning options on
The #Cycles text box is used to enter the
number of "Quick" or "Take & Write"
exposures to be taken in sequence with on button push.
The xPHARO display box can be used to get various image,
zoom modes, stretch options, Strehl analysis, photometry
statistics, statistical plots, and full field options. Some of
these were covered earlier in this document. To prevent overloading
the new observer with more detail on these options, it would be
better if the observer meets with a support engineer to go over the
entire PHARO GUI in general, and the display options in particular.
Dithering is accomplished with two steering mirrors in the
AO bench. These mirrors moving in tandem enable the AO system to
remain locked on a target while the target can be dithered in the
field of the science camera. Here are some tips for "smart"
dithering: Don't make dither moves greater than 5 arcseconds,
and wait 2 seconds before making another move to maintain a stable
lock. Moves are made manually, or with a macro, both methods are
discussed in this section.
If the observer makes a dither move too far or too fast, the
AO system will lose lock. The telescope operator will need several
minutes to restore the lock.
On the xPHARO GUI there are manual dithering buttons for the
observer to use. The buttons can move in either sky or detector
units. As stated previously, the buttons move the steering mirrors
to move the target on the PHARO array but not off the AO wave front
sensor (dithering). The steering mirror motors suffer from
hysteresis, and stiction. Fine move accuracy in a best case
scenario is around 1 pixel/.025″. These buttons are mainly
used to get fine placement behind a coronagraphic spot, or a grism
A digital fine move method exists. This method applies
tip/tilt to the deformable mirror. This method is controlled by the
AO operator/telescope operator/telescope operator. The digital moves
are once again a method to make fine moves to position a target on
a coronagraphic spot or slit. While this method is accurate, it is
limited in range 0.2″, it reduces Strehl by several percent,
and it reduces available DM stroke.
There are three nodding methods to move the telescope with
the AO unlocked (open loop). Nodding is used to make moves for sky
frames, field identification, and star searching. The three choices
are: 1) telescope hand paddle, 2) ask the telescope operator to move
the telescope with the TCS, 3) use the xPHARO mover buttons with
the system unlocked (be sure to unlock).
Macros are simple text files that can make dither moves,
take exposures, move filter wheels, and repeat sequences. Examples
of macros can be found on the PHARO computer Ezra2 one level below
the PHARO login directory at /macros. Macros are loaded with the
Load Macro button on the xPHARO GUI. Macros are run
with the (you guessed it) Run Macro button. Be sure to
add 5 second pauses after each 5 arcsecond dither move, to make
sure the steering mirrors are stable.
The slit wheel holds the two coronagraphic spots. The slit
wheel is in focus, and the spot locations can be easily identified
against a moderately bright background.
The coronographic spots are mounted on transparent calcium
fluoride substrate. The substrate is masked to match the 25″
field. There is no wider field of view advantage in using the 40″
field since the area beyond the 25″ field substrate is opaque.
The two spots are 0.46″ and 0.91″. They can be
selected using the slit wheel pull down menu.
A drawback to coronagraphy is the fact that the slit wheel is
not repeatable to better than a few pixels. Dust specks and the
spots will not line up on the same locations after slit wheel moves.
This non-repeatability makes flat-fielding a problem. There are two
choices for taking flats; 1) remove the coronagraphic spots for
flats (use the 25″ field mask), and 2) set up the slit wheel
for one spot selection, and never move the slit wheel after taking
Note items 11. through 16. in the section above relating to
dithering, nodding, and macros. The .025″ move is the finest
move that the SSM's can currently make, and this move is
unfortunately not always precise with the mover buttons.
It is reported by observers that the spots are slightly
transmissive in the K-band. If the object is bright enough, or the
exposure long enough the occulted object should be detectable with
the xPHARO line display plot. Use the line display plot to center
the target behind the spot. BE CAREFUL to not over expose the array.
Use short exposure quick sources to make sure the target is fully
behind the spot before taking long science exposures.
System flexure is an issue. Targets can drift out from behind
the spot on the order tens of minutes. The flexure effect is
variable depending on Cass ring rotation, altitude, and azimuth.
In the afternoon home the slit wheel. After homing, move the
slit wheel to the desired slit for observations. Take an image with
enough stray light to see the slit. Use the vertical fiducials to
mark the the slit location. You will need this when observing to put
your object correctly on the slit.
If the the slit does not look vertical, use the tweak option
in "current" mode to get to vertical. After tweaking, home
the Slit wheel, and then move back to your slit. Take another image
to make sure that the slit is vertical. Tweak and repeat as
Grism spectroscopy with PHARO was designed for use on the 40″
field. Make sure the Carousel wheel is in the 40″ position.
In the afternoon, home the Grism wheel, and then move to the
grism of choice. Take a test spectra of the dome and check that the
spectra is horizontal with the horizontal fiducial markers. If the
spectra is not horizontal, tweak in "current" mode, home
the wheel, go back to the grism and check the spectra. Tweak and
repeat homing as necessary.
The grism wheel is "heavy" and a bit unbalanced due
to the grism wedges. This at times can cause the grism wheel to make
incomplete moves. During observing if the spectra looks strange,
home the wheel and try again.
It is a good idea to write down the coordinates of critical
fiducial markers. If the xPHARO GUI crashes, and you need to
re-start xPHARO, the fiducial locations will be lost.
The basic spectral observing tactic goes something like...
Set the Slit wheel to 40″ field, set the
Grism wheel to a safe ND location, make sure the Carousel is at the 40
mas position, and then take a short "quick" image to
safely set an exposure value. With the camera in Cont Acq (&
AO locked), dither the object into the vertical slit fiducial
markers that you have previously set. Turn off Cont Acq,
move the Slit wheel from 40″ field to the chosen
slit, and take another "quick" image to confirm the target
location on the slit. Then move the Grism wheel to the selected
grism. Take science images. When slewing to a new location, set the
Lyot wheel to Block, set the Slit wheel to 40″
field, and set the Grism wheel to some safe ND location.
The currently available slits are; 0.13″, 0.26″,
0.52″. The available grisms are; J, H, K bands.
By Palomar custom, you have until the next afternoon to
remove your data from Ezra2, the Sun workstation that runs PHARO,
and purge the disk for the next user. It never hurts to leave a
note on the computer giving its status.
If yours is the last night of
AO, make sure you put a note on Ezra2 for the day crew, telling
them that you are still running tape backups or FTP. Hopefully the
daycrew will gently roll Ezra2 to the corner of the data room
without disturbing it.
Please share your feedback on this page or any other Palomar topic at the
COO Feedback portal
PHARO Cookbook / v 2.0
Last updated: 16 June 2015 ACM