CCD Imaging
The 3towers Observatory was
originally designed for visual observing. The optical quality of
its Meade LX200 12-inch telescope is quite good. The ten-foot
Home Dome used for the observatory is convenient for one or
two observers and offers great wind protection. Its large
36-inch shutter opening provides a 360 field of view, allowing
for up to two hours of observing before the dome has to be
rotated. A roll-off roof observatory would give a better view of
the night sky, but the homeowner’s covenants in my neighborhood
do not permit a roll-off roof observatory. The local seeing at
the 3towers Observatory is average or below average, though
there are occasional times of good seeing that permit excellent
planetary observing. The dome does not cause excessive heat
build-up during the day, and the telescope can be cooled to the
ambient evening temperature in less than an hour.
The 3towers Observatory functioned quite well for visual
observing for several years and was easily converted to CCD
imaging in June of 1998 when an
Apogee KX260
CCD camera was purchased. The Meade LX200 12-inch telescope
guides accurately enough that one-minute exposures with the
Apogee KX260 CCD Camera often show no trailing. The camera is a
delight to use, because it is reasonably sensitive, and because
its 20 micron pixels are well suited to the focal length of the
LX200 either at its normal f/10 ratio or at a reduced focal
ratio of approximately f/6.5 using a Meade focal reducer. The
one-second read out time of the camera allows one to acquire
objects and focus in near real-time. Testing has shown the KX260
camera can record recognizable 17.5-18.0 magnitude stars in a
one-minute exposure. To record fainter stars or faint
nebulosity, it is necessary to use longer exposure times.
In April 2000, a new CCD camera system was ordered for the
Grasslands Observatory, and the Apogee AP7 camera that had been
in operation at the Grasslands Observatory was relocated to the
3towers Observatory in early September 2000. While the KX260 CCD
camera is a pure joy to use, it does not have the quantum
efficiency of the AP7. The AP7 has a somewhat longer readout
time making centering and focusing more difficult. However, it
makes up for this disadvantage by permitting greatly reduced
exposure times and fainter limiting magnitudes in comparison
with the KX260. The Meade LX200 telescope can sometimes track up
to four minutes without objectionable trailing. However, it is
difficult to get consistently good tracking with exposures
longer than 60 to 90 seconds. Therefore, I usually take a series
of one-minute exposures and add them together using a median
routine in either
Mira or
MaxIM DL/CCD.
The LX200 is mounted on a permanent metal pier bolted to the
concrete floor of the observatory. The observatory floor is a
12-inch thick square concrete pad 15 feet by 15 feet. The
ten-foot Home Dome is also bolted to the concrete floor. This
design is simple and inexpensive but provides little protection
from vibrations being transmitted to the telescope by someone
walking around inside the dome. This was a problem when the
telescope was used visually and is a problem with CCD imaging.
If there is much activity inside the dome when a CCD image is
being exposed, the subsequent vibrations transmitted to the
telescope ruin the CCD image. If one sits comfortably and
quietly in a chair working on the observatory’s computer during
an exposure, no significant vibrations are transmitted to the
telescope. Thus, it is quite possible to use the system while
inside the dome. However, it rapidly gets boring if you want to
take a sequence of 30 one-minute exposures.
To do CCD imaging, I had two choices. I could either sit there
and entertain myself by reading a magazine in the low ambient
light level of the observatory, or I could go inside the house
for 30 to 60 minutes and return when the sequences of exposures
was finished. I had to be careful not to bump the telescope, and
I had to plan for one or two exposures to be ruined while
getting up to exit the observatory.
Going back and forth between the house and observatory during a
long sequence of exposures worked fine for a while. One or two
exposures were ruined, but most of the images in a given
sequence turned out fine. I could be inside doing something
else, but I did get nervous wondering how the images were
turning out and whether the telescope and computer were all
right. I used this technique of going inside for a long sequence
of exposures for a few months, but, because the observatory is
220 feet away from my front door, I rapidly got tired of running
back and forth for frequent checks on the system. The CCD camera
is light tight, and a dew shield placed on the telescope
prevents light coming off the computer screen and light cast by
a small reading light from fogging the exposures. Thus, it was
possible to sit in the observatory and quietly catch up on one’s
reading while taking a long sequence of exposures.
None-the-less, this also got tiring. Not only was it boring to
sit in the observatory reading astronomy magazines, but I also
got cold. Even in Arizona it gets cold in the winter, sometimes
going below freezing. Therefore, I decided to see if I could
automate the observatory to control the telescope and CCD camera
from inside my house, only going outside to start the system up
at the beginning of an evening’s observing run and later going
outside to shut it down at the end of the evening’s work.
Intel AnyPoint Network
When I built the observatory, I had a phone line, a coaxial
cable, and a power line extended from the house to the
observatory. If I had more foresight, I would have also run
computer network cabling to the observatory. Then it would have
been possible to network my in-house computer to the
observatory’s computer, or I could have directly controlled the
telescope and CCD camera from my in-house computer. The computer
in the house is located in my astronomy office, a spare bedroom
that has a computer, desk, shelves, and other assorted
materials, including all past issues of ASTRONOMY magazine, Sky
& Telescope issues back to January 1955, and National Geographic
issues back to 1956 as well as other assorted white elephants.
The computer is a reasonably powerful system with a 450mHz chip
and 256 megabytes of RAM. It currently runs under Windows 98
second edition and is linked to a 256K DSL line for fast
Internet connection. I wanted to link the in-house computer with
the observatory computer, but I did not want to run 220+ feet of
network cable across my front yard. It would have been an
enormous task to run it from the house through multiple cactus
patches to the observatory not to mention having to bury the
cable to protect it from the elements and preserve the
appearance of my front yard.
Fortunately, I happened to see an advertisement in Time magazine
for Intel’s AnyPoint Home Network. This product networks two or
more computers that are connected to the same telephone line. It
comes in a variety of configurations, including parallel port
and USB models as well as internal PCI cards. I originally
purchased two PCI cards, one for the in-house computer and one
for the observatory computer. The total cost was about $160, and
the system had a nominal network speed of 1 megabit per second
(1 Mbps). I had no trouble installing the cards and the
accompanying software on both computers. Amazingly, the two
computers were networked together with no problem. I could now
transfer files between the computers, and I no longer had to use
Zip or Jaz disks to move data between the computers.
The Intel software allows you to map the drives and printers on
every computer on the network to permit or deny network access
to these drives and printers. I calculated it took about 5-7
seconds to move a 512K CCD image file from the observatory
computer to the in-house computer. Today, there are several
similar or less expensive phone network systems and “wireless”
network systems available. The wireless networks use radio
transmission or infrared transmission between computers to move
data over the network.
PcAnywhere
No network system of this type, including AnyPoints, allows you
to run executable files on one computer from another remote
computer. Thus, all I could accomplish with my network was to
transfer files and to take advantage of the accompanying
Internet Sharing Software (ISS) that comes with the Intel
product. This software, when activated on both computers, allows
me to access the Internet on the observatory computer taking
advantage of the fast speed of the DSL connection on the
in-house computer.
To run executable programs on the observatory computer, I had to
purchase Symantec’s
pcAnywhere
for about $170. This software is designed for total control of
one computer by another computer either via a telephone modem
connection or via a network. I installed version 8.0 (later
upgraded to version 9.0) of pcAnywhere on both the in-house and
the observatory computers. It took me quite a bit of work to
learn how to use the software. The documentation accompanying it
is not designed for someone unfamiliar with such software. In
the process of working with pcAnywhere, I gained some
rudimentary knowledge about computer networks, learning to use
such terms as “host,” “remote,” and “TCP/IP.” I finally stumbled
on how to determine the IP address of the observatory computer
and the in-house computer. To control the observatory computer
from the in-house computer, I start pcAnywhere on the
observatory computer, which acts as a host waiting for a call
from the in-house computer. The in-house computer runs
pcAnywhere and remotely controls the observatory computer by
looking for it over the network using the observatory computer’s
IP address to find it.
Once the in-house computer connects via pcAnywhere with the
observatory computer, I have total control over the observatory
computer from inside my house. I actually see the screen
displayed on the observatory computer, and I can open and run
any programs available on that computer. pcAnywhere runs as a
window on the in-house computer. I can minimize this window and
run whatever I want on the in-house computer separately from
pcAnywhere’s control of the observatory computer.
To remotely control the Meade LX 200 telescope, I bought a Meade
mouse serial port connector that attaches the observatory
computer to the control panel of the telescope. Many programs,
such as
The Sky and
MegaStar,
have built in software commands for the Meade LX200 telescope,
allowing one to completely control the telescope with a
computer. This control includes find, slew, goto, and focus
commands. Once I installed The Sky on both the house computer
and the observatory computer, I had control of the telescope
from inside my house using the house computer to control the
observatory computer and the telescope via the Intel network and
pcAnywhere.
There were still a number of problems to overcome. First, the
network was sluggish having only a 1 Mbps speed. pcAnywhere is
very computer intensive and takes up a lot of “bandwidth.” I
could move the telescope from object to object and take CCD
images via MaxIm DL/CCD or CCDSoft from my in-house control, but
it required a significant amount of time to transfer images
between the two computers and make adjustments in The Sky or the
CCD imaging programs running on the observatory computer.
Second, I had only a two-hour window for CCD imaging before I
had to go outside and move the dome.
To improve the speed of my network, I replaced the original 1
Mbps AnyPoint network PCI cards on the observatory and the
in-house computers with new 10 Mbps PCI AnyPoint Home Network
cards that became available in late 1999. This significantly
improved network efficiency. It now only takes 0.5 to 1.0seconds
to transfer a 512K CCD image file over the network. Remote
control of the telescope and CCD camera was significantly
improved, though still sluggish due to the heavy use of computer
resources by pcAnywhere.
Telescope Control: The Sky and TPoint
The Meade LX 200 telescope is a wonderful instrument for the
money. Its ability to point and slew from one object to another
enabled me to observe far more celestial objects than I ever
could have imagined possible. Usually, the telescope would point
accurately enough to place a requested galaxy or cluster
somewhere within the field of view of a moderate power (150x)
eyepiece. Some evenings, I would go out to the telescope and use
its hand paddle to observe 60 objects an hour for a couple of
hours.
Because my LX 200 telescope has good polar alignment, all I have
to do to start the telescope up for an evening of observing or
imaging is turn it on and accurately center it on a known
object, such as a first magnitude star. Once I tell the
telescope where I am pointed initially, it is able to slew about
the sky with great accuracy as long as its power is not turned
off and as long as the RA and Dec axes are kept locked and only
adjusted with the motion controls on the telescope’s hand
paddle. The Apogee KX260 CCD camera has a relatively wide
18-minute field of view when used with an f/6.3 focal reducer
and the Meade 12-inch LX 200 telescope. Nevertheless, I found
the pointing accuracy of the telescope to be less than ideal for
CCD imaging. Often, the object I wanted was located on the edge
of the CCD image, or it might be located slightly off the edge
of the image, necessitating several minutes of work to center
the object. This centering is not too hard with the KX260
because of its one-second readout time, but it is time consuming
none the less. To overcome this problem, I bought version 5 of
The Sky and
TPoint pointing software from
Software Bisque.
TPoint is designed to analyze and improve the pointing accuracy
of computerized telescopes. It is a complex program that
corrects for a telescope’s azimuth and altitude errors. When
correctly applied, TPoint greatly improves the pointing accuracy
of a telescope.
I spent many hours learning how to use The Sky and TPoint
properly. There is steep learning curve with TPoint. Among other
things, it requires you to center the telescope on multiple
stars in all parts of the sky so that it can model the pointing
errors of the telescope. After several failed attempts, I
finally got the program to operate properly. It greatly
increased the pointing accuracy of my telescope. Now, most
objects that I seek with the CCD camera are reasonably centered
on the CCD image on the first try. I did not make a particularly
good TPoint model for my telescope. My model is based on
approximately 30 stars. I need to make a much more accurate
model based on many more stars to realize the full potential of
the program. However, I am very satisfied with it, and I greatly
recommend The Sky and TPoint.
The Sky does a superb job of controlling my telescope. Its star
maps are graphically accurate enough that I can use them to
recognize stars and objects on my CCD images. It accurately
slews the telescope, and it can even focus the telescope! When I
begin an evening of imaging, I carefully focus the telescope
with its hand paddle and the use of a
Kendrick Kwik Focus Lens Cap placed over the front of the
telescope. However, as the evening progresses, there is often a
change in the telescope’s focus. I used to walk outside to
refocus the telescope. Until March 2001, I used a
JMI
electric focuser that is powered by the telescope’s control
panel. This is quite a good focuser, and it gave me many years
of excellent service. The Sky was able to access the focuser,
enabling me to focus from inside my house. I replaced the JMI
focuser with a
Technical
Innovations Robo-Focus, because the Robo-Focus has a precise
digital readout that remains fixed even when it is turned on and
off. The focusers is precise enough that you can make
incremental minor focus adjustments that are reproducible. That
is, if you overshoot a position by 100 units, the Robo-Focus is
able to easily bring you back to where you were by adjusting the
focus 100 units in the opposite direction.
The focusing readouts I used to get with The Sky and the JMI
electric focuser were "digital" in the sense that focus
positions were represented by numbers, and I could use these
numbers to obtain a good focus. However, the number positions
are not very precise. If you overshoot a focus by say 100 units,
you can not simply back up 100 units to get back to where you
were. In addition, whenever The Sky is closed and reopened, the
focus position number goes back to zero.
With my present equipment and software setup, I now sit inside
my house and use pcAnywhere to completely control the
observatory computer. For focusing, I use either MaxIM DLL/CCD
or CCDSoft to take a series of short exposures. While the
exposures are being taken, I use the Robo-Focus software to
slowly change the focus until I once again have a sharp focus.
This is a not a quick procedure, but I usually have to do it
only once an evening, and I much prefer to do it from inside the
house rather than having to walk out to the observatory.
Dome Tracking and Automation
Once I opened up the observatory, uncovered the telescope,
started the computer, focused the telescope, and started The Sky
with Tpoint, I could go inside and operate the telescope from
the computer in the house. However, after two hours or less, I
would have to go outside and manually move the dome to keep the
shutter opening aligned with the telescope. To overcome this
problem, I purchased
Technical
Innovations dome motor and power supply for the 10-foot Home
Dome. To motorize the dome, James McGaha and I had to remove the
dome from the observatory base and lift it up to rest it on wood
blocks while we mounted and installed the motor and power
supply. This took an afternoon of hard work in the hot Sun, but
the dome motor worked perfectly the first time we tried it. A
small hand paddle moves the dome in either a clockwise or a
counterclockwise direction. It takes one-minute for the dome to
move through a complete 3600 rotation.
At first, the dome became consistently stuck at one location as
it rotated. To get it to move past this spot, I had to gently
push on the dome. Apparently, the skirt hanging down from the
dome designed to cover the dome track so dust and water would
not enter the observatory rubbed against the dome base at that
point, probably because the dome base is not perfectly circular.
I discovered that a small amount of WD 40 applied to the inside
of the dome skirt alleviated this problem. I now keep a can of
WD 40 in the observatory for this reason, although lately the
dome has been rotating without problem. The area where the dome
skirt stuck on the dome base must have worn down enough with use
to allow the dome to rotate freely.
Next, I purchased
Dome-Trak
from Technical Innovations. Dome-Trak is an infrared system that
links the telescope with the dome position, keeping the dome’s
shutter opening centered in front of the telescope. It consists
of a control module, four infrared emitter/detector pods, relay
box, hand control, cables, and Velcro bands for fastening
equipment to the telescope. The four infrared pods are placed at
900 intervals around the telescope. The Dome-Trak works by
emitting infrared beams from the emitter/detector pods. When the
telescope is properly centered in front of the shutter opening,
the beams go out the opening. If the dome is not properly
centered, one or more of the beams will reflect off the inside
of the dome and be detected by the infrared pods. This causes
the relay box to activate the dome motor to move the dome until
the shutter opening is properly centered in front of the
telescope.
The Dome-Trak works reasonably well. It follows the telescope
faithfully for several hours, especially if I decide to “work” a
constellation over a period of hours making small slewing
movements from one object to another in the same constellation
rather than jumping from one quadrant of the sky to another. If
I slew a large distance from one part of the sky to another, the
dome will usually line up with the telescope, but this may take
a few minutes while the Dome-Trak figures out where the
telescope is centered. Some evenings, I can image all night and
have no problem with the dome tracking. Other evenings, I may
have to go outside and use the hand paddle to align the dome up
with the telescope, as the dome sometimes gets lost and rotates
around for several minutes trying to find the telescope’s
position.
When I slew from one object to another, I carefully examine the
initial images for the new object to see if it is centered
correctly and to see if there is a problem with the dome
alignment. Once I determine the new object is properly centered
and focused and there is no problem with the dome alignment, I
begin a new series of exposures. At this point, I can watch the
individual exposures as they come down the line, or I can go do
something else, such as going to dinner or watching a movie. I
don’t have to stick around and monitor the system. I come back
to the in-house computer when I want to begin a new series of
exposures on the same object or slew to a new object. I no
longer have to manually transfer images from the observatory
computer to the in-house computer. I just transfer them over the
network automatically as they are being obtained.
It is fairly easy to recognize when the Dome-Trak has not
properly aligned the dome with the telescope. Sometimes, no
object is visible. All that is seen is a white, mottled image of
the inside of the dome. At other times, the image shows light
streaks from visible light or the infrared beam partially
reflecting into the CCD field of view. When I see that the dome
is not properly centered, I take a series of short exposures to
determine if the Dome-Trak will correct the situation. Usually,
within two to three minutes, the dome will be properly centered.
If the dome is not properly centered after five minutes or so, I
go outside to see what is the matter. Fortunately, I don’t have
to do this very often.
I don’t think the Dome-Trak is sufficiently accurate or reliable
enough with my dome and telescope combination to permit a
totally unattended, multi-hour slewing of the telescope about
the sky for an automated supernova search routine. An automated
supernova search routine might image 200-300 galaxies an evening
and cover a large portion of the sky from north to south and
east to west. To be effective, such a routine must move the
telescope accurately and rapidly from one galaxy to another.
There can be no lost time or lost images. Because the Dome-Trak
occasionally gets lost or requires several minutes to properly
center the shutter opening in front of the telescope, it
requires periodic monitoring. It is a very useful device, and it
has permitted me to spend most of my imaging sessions inside the
house. I like it a lot, and I recommend it for others in similar
circumstances. Nevertheless, it does have its limitations.
Conclusion: the Good News & the Bad News
The 3towers Observatory is now “semi-automated.” It can be
controlled from inside my house by remote control. This is a
good news bad news situation. The good news is that I can both
“observe” and be inside the house doing other things. The bad
news is that this cost me a lot of time and money to implement,
and it discourages me from being outside under the stars. I have
foregone the beauty of the night sky for the allure of CCD
imaging. I have traded the joy of observing faint galaxies and
nebulas at the limits of one’s dark adaptation for the glare of
a computer screen and the electronic beauty of a CCD image. But
that’s another story. |
