Paper presented at the 190th meeting of the American Astronomical Society.
The Hanna City Robotic Observatory
Jerry B. Gunn 1269 N. Skyview Drive, East Peoria, Illinois 61611
Electronic mail: jgunn@mtco.com
www.mtco.com/~jgunn
Charles F. Lamb 10300 Smithville Road, Hanna City, Illinois 61536
Abstract. Recent development of low cost automated commercial telescopes and CCD cameras
makes possible the construction of remotely operated robotic telescope systems for under $10,000
USD. We describe the design and development of the Hanna City Robotic Observatory(HCRO),
an observatory that automatically performs photometry of stars using low cost, off the shelf
components. The HCRO can be controlled remotely or can be scheduled to make unattended
imaging observations.
1. Introduction
Several Robotic Telescopes have been described on these pages. In most cases, an existing
telescope was modified and upgraded with motors and sensors to enable unattended observations.
A history of the development of automatic observing systems is given by Richmond et al. 1993 and
we shall not attempt to repeat their fine work. Within the last five years, two instruments, intended
for the amateur astronomy market, were developed by companies and offered for sale at low prices.
Meade Instruments introduced its completely motorized LX200 telescope and accessories in 1992
and the Santa Barbara Instrument Group(SBIG) introduced its ST-7 CCD camera and accessories
in 1994. These instruments are capable of providing all of the major functions necessary to construct
a fully automatic observatory.This project was undertaken to determine the feasibility of using these
instruments in the construction of a robotic observatory. We have endeavored to use very simple,
inexpensive methods to accomplish our goal.
2. Telescope
The 0.21-m Meade LX200 telescope is a fairly fast f/6.3 fork mounted Schmidt-Cassegrain
telescope which can be mounted in either Alt/Az or equatorial configuration. The equatorial mount is
an optional accessory. To simplify image rotation problems associated with ALT/AZ mounts, the
equatorial mount was chosen. There are no areas of the sky that cannot be observed with this
configuration. The telescope has motors with encoders on both axis and contains an integral
computer system which provides pointing and slewing functions. The computer also contains a ROM
with the coordinates of 250 bright stars, all Messier objects, 7,840 objects in the CNGC, 15,928
SAO stars brighter than 7th magnitude, 12,921 UGC galaxies, 5,386 IC objects, 21,815 GCVS
and 8 major planets. Most telescope functions are controllable though the integral RS-232 data port
which enables a personal computer(PC) to control the telescope. A very important feature of the
telescope is its ability to correct tracking errors caused by inaccuracies in the drive gears which
move the telescope. The integral computer can be trained to correct these errors, and stores the
corrections in its non-volatile memory. After the drive has been trained though the simple nine minute
manual procedure, the telescope can accurately track stars for over five minutes with no visible
trailing of star images. Polar alignment is achieved switching between two stars in the internal catalog
and by making corrective adjustments over a 60 minute period. The equatorial mount is constructed
in such a way that the telescope can be removed and re-attached to the mount without the need to
repeat polar alignment. A focus motor, available as an accessory, attaches to the telescope and is
actuated by commands sent to the integral computer.
A goal of any robotic telescope is to put the intended object on the detector imaging area when
slewing between objects, even on long slews. When accurately polar aligned, the telescope is
capable of 90 degree slews while still placing the object on the imaging detector. In practice, if a
long slew is necessary, an intermediate star is slewed to, centered, the telescope coordinates
synchronized, and then the slew is completed.
3. CCD Camera
The SBIG ST-7 CCD camera is based on the low noise Kodak KAF-400 CCD and uses a double
correlated sampling readout to achieve a readout noise of 15 electrons RMS. Dark current is 0.5
electrons/pixel/sec at 0 degrees C. The camera is offered either with or without anti-blooming gates.
The anti-blooming gates effectively halve the full well capacity and cause linearity problems as the
wells near saturation, therefore the camera was purchased without the anti-blooming gate option.
The inherently low noise of the KAF-400 minimizes cooling requirements so a single stage
thermoelectric cooler with temperature regulation is used. The CCD array of 765 X 510 pixels, each
9 um on a side, yields a field of view of 12 X 18 arc minutes. Each pixel subtends 1.4 arc seconds
which oversamples stellar images so the pixels are binned 2 X 2 with a resulting 18 X 18 um pixel
size and a better telescope/camera match of 2.8 arcseconds/pixel. The on chip binning also increases
the full well capacity to 370,000 electrons. The camera contains an electromechanical shutter that
provides exposure times from 110 ms to 1 hour with a 10ms resolution. SBIG also offers the same
camera with the KAF-1600 CCD which has an array of 1534 X 1020 pixels. Although increasing
the imaging area by a factor of four would be a nice feature, the added cost and additional increase
in the amount of data to process was prohibitive for this project. Communications with the camera is
done through the parallel printer port of the PC.
An additional feature of this camera is its ability to use its second CCD, a Texas Instruments TC211
to act as a tracking CCD. The chip is mounted off axis and par focal to the main imaging CCD. The
camera can send corrective signals directly to a port on the LX200 telescope designed specifically
for such use. These signals activate the guiding relays on the telescope. With the proper software to
determine the corrections, we have made 40 minute exposures with near perfect tracking. This
tracking feature is not implemented in our software because we use relatively short exposure times,
but is available using the software that is provided with the camera if imaging of faint objects is
necessarySimilar to many other inexpensive CCD detectors, the KAF-400 has poor blue light
response requiring increased exposure times to get useful data. Although we chose to use V and B
filters, a better choice would be to use the redder V, R and I filter set.
The CWF-8 filter wheel is an SBIG accessory which bolts directly to the ST-7 camera, making it an
integral part of the camera. The wheel has openings for five 1.25 inch diameter thread-in filters. The
CFW-8 is controlled by commands sent to the camera from the controlling PC. The wheel
positioning system uses a closed loop stepper motor with a stated positional accuracy of +-0.01
inches. No attempt was made to verify this claim.
An important feature of this camera, which makes this project possible, is the availability of a
software driver library from the manufacturer. We were able to link these driver functions into our
control program and gain total control over the camera. All camera functions including data readout,
temperature control, shutter control, exposure time, binning, and filter selection are programmable.
We wrote functions which effectively enable us to accomplish our goal without relying upon the
software supplied with the camera. Without the driver library, this project would not have been
possible.
4. Enclosure
A freestanding compact enclosure was constructed featuring a flip-top lid. The flip-top lid avoids
needing to deal with dome rotation to follow telescope motion. The lid was constructed to insure that
it cannot come into contact with the telescope regardless of telescope orientation. The lid is attached
to the enclosure with a large steel hinge and opened/closed with a sprocket and chain drive. The
motor and gearbox used to open the lid are mounted inside the enclosure. The telescope pier
consists of an 11.43 cm diameter steel pipe which is buried 1.6m in the ground. A circular flat steel
plate was welded to the top of the pipe to enable mounting of the equatorial wedge. The pier is
isolated from the enclosure. Power and communications cables are buried and terminate in a nearby
structure.
5. Control System
The observatory control system consists of a personal computer motherboard, custom controller
adapter board, modem, power supply, batteries, battery charger, hard disk drive, floppy disk drive
and relays mounted in a weatherproof steel electronics box. The electronics box is mounted on the
inside of the enclosure door. Opening the enclosure door causes the box to swing out providing
accessibility for equipment upgrades or repairs. The personal computer motherboard is a
386DX/25MHz type with 16MB of RAM and 250MB of disk storage. The custom controller card
plugs into a motherboard slot and contains circuitry to control the lid relay, instrument power relay,
cloud sensor amplifier, computer lock-up sensor, telescope position sensors and desiccator relay. If
the computer malfunctions, the controller circuitry will reboot the computer which causes the lid to
close and cuts power to the instruments. If the main power fails the lid will automatically close using
the batteries for power.The telescope must be synchronized with the sky each night. This is
accomplished using the time of day and two magnetic reed switch sensors mounted on the telescope.
The telescope is moved in both Right Ascension and Declination until the home position is sensed.
The altitude and azimuth of this point has been previously measured, so using the time of day, the RA
and DEC can be accurately computed. These coordinates are sent to the telescope and it is
synchronized to match these coordinates.
Remote operation of the observatory and data file transfer is accomplished using dial-up modems
and the DOS version of PC AnyWhere. A "Host" TSR runs on the observatory PC and the remote
PC executes the "Remote" module of the software.
6. Environmental Aids
To reduce heat build-up , the observatory enclosure is painted white to reflect a maximum amount of
sunlight and, in the warm summer months, the computer is turned off with a timer during the heat of
the day. Humidity is controlled using two thermoelectric coolers (TECs) acting as dehumidifiers. One
is mounted inside the electronics box and the other is located inside the telescope enclosure. The
coolers alternately turn on and off. During the on cycle, water vapor freezes on the cool side of the
device. When power is removed, the ice melts and the resulting liquid water is funnelled out of the
observatory.
7. Weather Station
The weather station presently consists of only a cloud sensor. A thermoelectric module is
sandwiched between two aluminum plates and mounted outside the observatory, one side pointing
toward the sky and the other toward the ground. The module will generate a small voltage directly
related to the temperature difference across the thermoelectric module. The voltage is amplified by a
high gain instrumentation amplifier whose output is fed to a voltage controlled oscillator. The output
of the oscillator is run to an input line of the computer printer port which counts the pulses. The
number of pulses can then be calibrated to indicate the cloud cover. A clear sky will "look colder" in
relation to the ground temperature than a cloudy sky. This technique gives fairly reliable results but
sometimes gives erroneous readings during extreme weather conditions such as rain, high winds, or
very rapid temperature changes. A rain sensor using a heated conductive grid (Tosti 1996) will be
added along with an anemometer to complete the weather station. We are investigating adding a
small window to the North side of the observatory to facilitate observing Polaris with the lid closed.
This could result in a near fail-safe method to determine cloud cover.
8. Software
The control software is written in C and operates under the Microsoft DOS version 6.2 operating
system. DOS was chosen because a large amount of control over system functions is available to the
programmer, the later versions of DOS are quite robust, and the purchase cost is very low. The
main control program controls all observatory, telescope, camera and communications functions and
avoids the complexity of multi-computer control systems. Functions available include:
Open/Close Lid
Instruments On/Off
Home telescope
Park telescope
Center brightest star
Sync coordinates
Slew to coordinates
Count number of stars in frame
Measure average FWHM of stars in image
Set CCD chip temperature
Select filter
Make exposure
Auto exposure sequence
Apply flat field
Subtract dark frame
Get sky background level
Auto focus
Build master dark frame
Build master bias frame
Build master flat frames
9. Data Reduction
Since HCRO can accumulate over 100 MB of image data in a single night, a method to quickly
reduce the images to photometric data was needed. The EzPhot program was developed to
automatically find and convert all stars to instrumental magnitudes and output coordinate and
magnitude data to simple text files. The operator then identifies and annotates the target stars in one
of the text files and the Register program will automatically align all of the image text files to this file.
As an aid to differential photometry, three stars may be chosen, a variable, comparison and check,
and the Register program will automatically output a text file with the time of observation and
difference data. This file can be input directly into a spread sheet program for graphing and data
analysis.
10. System Performance Characteristics
The system is currently being used to measure light curves of eclipsing binary stars. Figure 1 shows a
representative sample of the photometry this system can achieve using one minute exposure times.
The eclipsing binary star BX Draconis has a V magnitude of 10.6.
11. Conclusions
We believe this feasibility study has proven the concept of using the Meade LX200 and SBIG ST-7
CCD camera in a robotic observatory. Many questions still remain concerning the reliability of such
inexpensive equipment. Even if components such as telescope motors and gears break or wear out,
replacing them is quite inexpensive. The whole telescope can be easily removed from the enclosure
and shipped to Meade for a complete overhaul for under $300 including shipping. Given the low
purchase price of the camera and scope, spares could be kept on hand should problems occur.
The telescope and imaging camera, usually the most complex and troublesome instruments of robotic
observatories, are now off-the-shelf items. Given that fact, the most difficult parts of this project
were software development and the design and construction of the enclosure. The lid opening
mechanism was designed and then redesigned until a working, reliable method was found. Sealing
the lid against rain and snow was also a challenge which was not met without some reworking. If a
dome or enclosure already exists, building a similar system could become quite straightforward.
It would seem that remotely operated robotic observatories are now within the reach of small
educational institutions and dedicated amateur astronomers. To promote further development of
similar systems, the software source code, electrical schematics and enclosure plans will be made
available to interested institutions or individuals.
This research was partially supported by a grant from NASA administered by the American
Astronomical Society. We are grateful to Janet Mattei, Director of the AAVSO, and Dirk Terrell of
the University of Florida for endorsing the project thereby enabling funding to amateur astronomers.
We also thank Barry Redenbo and Brian Hakes of the Peoria Astronomical Society for many helpful
conversations and advice along the way. Special thanks go to Dan Kaiser for assisting in the
development and testing of the image reduction software programs.
References
Richmond, M. W., Treffers, R. R., and Filippenko, A. V. 1993, PASP, 105, 1164 Tosti, G.,
Pascolini, S., and Fiorucci,M. 1996, PASP,108,706 Trueblood, M., and Genet, R. 1987,
Microcomputer Control of Telescopes (Richmond, Wilmann-Bell)