Note: This manual refers to the new or "Seitzer" autoguider.
TVGUIDER is a PC based guider running on both the 2.4-m and 1.3-m telescopes. It is a prototype for a system to completely replace the DTI boxes.
The system normally guides on the digitized video coming from the guider TV camera. Two cursors are implemented: one to mark the center of the finder field, and a second one to select guide stars. Full control is possible over the sampling of the video signal, and display contrast and brightness.
For acquiring faint objects, averaging of a 320x200 subfield is possible at 10 frames per second, or one can average a 120x120 subfield at full video rate (30 frames per second.) Pan around the main field while averaging using the cursor keys on the keypad.
IMPORTANT NOTE for the 2.4-m: the system works much better if you slow down the guide rates from the TCS default. Change the guide rates in the TCS to 1.0 for both ra and dec, and turn cos dec ON.
The old PC GUIDER system and DTI box are still available at the 2.4-m, but it does require changing the setup cabling. Let the staff know before your run if you wish to use it.
The new guider can be started or stopped using the xguider window on chichon. Parameters can only be changed from the PC keyboard.
cd c:\tvguider
tvguider
the program will start up, restore all the last used
parameters, and draw the cursors on the monitor.
To be sure the program is initialized to all default values, type:
tvguider -i
The program will prompt you for which telescope you are on,
and restore the appropriate values. Cursors are reset to
the center of the screen.
(If the sky is very bright, this automatic set may fail. Then use the '1 2 3 4' keys to select a gain - I suggest 3 for most use, 4 for a very faint star (minimum dynamic range), and 1 or 2 for bright stars. Then repeatedly hit '5' to decrease or '6' to increase the sky value. Aim for sky between 20 and 60, as determined by the intensity reading on the lower right of the monitor, or returned when hitting the space bar on the PC keyboard. If all else fails, decrease the high voltage and try the above steps again.)
The intensity displayed on the monitor is that corresponding to the central pixel of the active cursor.
Saturation is 255 - but if a few pixels in the core of the star saturate, don't let it concern you. Guiding seems to be better if a few pixels are saturated (but not a lot - if the entire box is white, that is too much. Lower the gain.)
With a 2 second average time, the system will guide on stars that are just visible in the raw video.
If you have a bright star, and the guiding fails with 'not enough counts': increase the box size. Then check to see that the sky level is below 100 units. Use the 'b' or '5' keys to adjust if necessary.
If the telescope is moving without any command from the guider, try exiting the guider program and rebooting the PC (CTRL-ALT-DEL).
This mode is essential for finding faint objects on the spectrograph slit. To use:
DO use a guide box somewhat larger than the star. Sky
is determined from the outer 2 pixels of the box.
Too small a box and if the star moves to the boundary
the guider can't follow it. Too large a box and
the S/N is reduced.
DO keep the sky level between 30 and 100. Too high a sky
level and the dynamic range is much reduced. Too low
a level and the pixel rejection schemes get confused.
DO adjust the high voltage before changing gain or offset.
DON'T expect great guiding if you change the box
size or location when guiding. Large errors
can result due to buffer changes.
DON'T turn the DTI box on.
DON'T turn the PC off or disconect the cable from the
PC parallel port while the telescope power is on.
DON'T disconnect the camera while running TVGUIDER
in any mode. The program uses vertical sync as
its time base: if it loses it, guiding will
cease and the program will tell you and will not
allow you to resume until a video signal is restored.
DON'T get confused by the difference between gain/offset and
display brightness/contrast. Gain/offset control how
the computer digitizes the incoming video signal
with the 8 bit ADC. The pixel value as displayed on
the TV is the digitized value, not display values.
Display contrast and brightness can be changed using the
h-j-k-l keys without changing the gain/offset setup.
Thus white on the screen does not mean that a particular
pixel is saturated (255 from the ADC). You could have
set the display contrast/brightness such that the display
saturates (white) at a lower DN than the ADC.
If you get totally confused, hit the 'z' or 'Z' keys
to reset everything (including the display) back to
default values. Now display black = 0 DN, display
white = 255 DN.
Hit the 'p' key to verify or change the guider parameters. If you are guiding, guiding will stop.
One can enter a new value, or just hit return to keep the old one.
The following are the parameters:
Then move the telescope a specified direction and time (10 seconds?) using the 'm' command. Measure the new position, and compute the rate.
This rate will be valid for large offsets, but probably overestimates the correction needed for small offsets.
time = fraction * (offset in TV pixels) / rate
with suitable adjustments for flip and rotation.
This is a crucial value, because if you set it too large the system never guides. If set too small, the system is always moving.
Because the 2.4-m has much faster guide motions, a time of 0.05 seconds seems to be called for.
This can also be set with the 'r' key if all other parameters remain unchanged.
I strongly suggest that you leave this option enabled unless you know what the program is doing, or you run into serious problems. Much of the code depends on the sampling of the video signal being done right.
Disabling this option is provided only as a last resort.
The following values worked in April 1997 at a declination of +32 degrees and the new 135 mm guider camera:
integration time 2 seconds
ra guide rate 2 pixels/second
dec guide rate 2 pixels/second
fraction of move 0.5
minimum guide time 0.2
minimum counts 500
direct imaging 1
rotation 0.0
Guide rates are different between direct imaging with
the guider probe and looking down on the slit.
If you go back to the old 85 mm camera, new values will have to be determined.
The 1.3-m balance is critical - if the telescope seems to be bouncing in ra, have the staff check the balance.
NOTE: since there is no cos(declination) on this system, if you try to guide at high declinations one will have to change the ra guide rate to a lower value, else the system will never be able to correct enough in ra.
The following values worked well in November 2006 at a declination of +32 degrees and the 135 mm guider camera direct with the default TCS guide rates at 2.0: Default is cos dec ON.
integration time 1 seconds
ra guide rate 4.00 pixels/second
dec guide rate 2.00 pixels/second
fraction of move 0.5
minimum guide time 0.05
minimum counts 500
direct imaging 1
rotation 0.0
If you go back to the 85 mm camera, or off the slit, new values
will have to be determined.
IMPORTANT NOTE for the 2.4-m: the system works much better if you slow down the guide rates from the TCS default. Change the guide rates in the TCS to 1.0 for both ra and dec, and turn cos dec ON.
The following keys are supported - other keys are ignored.
# Guiding control
s start guiding
q quit guiding
r set rotater angle
a goto average mode
# Cursor control - all guiding is done with square cursor
c toggle between cursors (active one has lines or 'hair')
+ increase cursor size (maximum 41x41 pixels)
- decrease cursor size (minimum 9x9 pixels)
# Cursor movement via keypad (NUM LOCK must be OFF!) -
# diagonal motion can be done with the 1 3 7 9 keys
5 toggle between steps of 1 and 10
i move selected cursor to center of field
# video sampling
z set values for bright stars (min gain, max range)
Z set values for faint stars (max gain, min range)
b adjust sky to 30-60 units.
1 2 3 4 select gain state 1 (low, max range)
through 4 (high, min range)
5 decrease background
6 increase background
space bar quick sky and centroid of star in guider
box using data from last TV frame
# display brightness and contrast
h j adjust black
k l adjust white
f transpose one frame and write to display.
m move telescope specified direction and time.
w write full TV frame to disk as binary file of 8 bit
integers: 512 columns by 486 rows. Useful to see
just how awful this data really is.
p set guider parameters (stops guiding)
? type this help page
x exit the TVGUIDER program. *All* guider parameters
and cursor positions and gain/offset values are
saved to the file 'tvguider.ini' in the current directory.
This file is read when the program is next started.
TVGUIDER can respond to commands received over the serial line (9600 baud, 8 bits, no parity, 1 stop bit.) At the present time, only start guiding ('B'), stop guiding ('Q'), and resume ('R') are implemented.
These codes are an historical artifact of the old xguider program.
Any function of the guider could be controlled over the serial line with a few lines of new software. Contact the author if there is something you would like done.
TVGUIDER attempts to move the star to the center of the box, not to the first position. (I'll add this as an option.)
TVGUIDER computes a centroid first, and then guides, and does not attempt to determine new centroids while the telescope is moving.
If the cursors vanish while not in averaging mode, or the board fails, or you suspect problems, exit the TVGUIDER program using 'x'.
Then type at the DOS prompt:
cxdemo iIf that doesn't work, try:
cxdemo b cxdemo biThen restart TVGUIDER.
If problems persist, check that the video cable is connected. Then exit TVGUIDER, turn the computer off for a minute, and restart.
Updated February 21, 2007 (rab/mdm)