2000 and 2010 Shoemaker Grant Recipient
The 60-cm Cichocki Sky Survey Telescope
Robotically observing since 2004
The 60-cm Cichocki telescope (named after Bruno Chichocki, an important benefactor to our program) is an internet robotic facility, designed and built during the period 2000-2002 and put in regular operation in March 2003. It is a world-class robotic telescope designed for wide range of astronomical observations, including search for minor planets and comets, explosions of supernovae in distant galaxies and detection of optical afterglows of Gamma Ray Bursts - probably the most energetic events in the universe. It may image the sky either in consecutive single shot or drift-scan along the great declination circles mode. The latter technique is much more effective and preferred for scanning the sky for asteroid and comet search under the framework of PIKA program. Very few existing modern telescope of this class are able to use this advanced technique.

We developed dedicated hardware and software solutions that enable us real time remote control of the observatory and the telescope. Since 2004, all observations are performed in an unattended robotic mode. Custom made internet interface to a observation scheduler was introduced in 2005 for more effective and user friendly telescope targeting. Using the weather station and all-sky camera, the overnight weather situation is continuously monitored by a watchdog program. In case of sudden cloud cover, the telescope is automatically parked and the observatory roof closed.

Observing Statistics
Cichocki Telescope Discoveries

System Overview
Telescope Design
Optics
Mount and Tube Assembly
Electronics
Operation and Software
Calibrations (pointing and tracking, optics collimation)
Calibrations - Photometry
Photometry Statistics and Deep Imaging
Scanning Along the Great Declination Circles
Gallery
Donators

Robotic Telescopes of the World
System Overview
The 60-cm Cichocki Sky Survey Telescope was built to succeed the PIKA asteroid and comet search program with a larger and more effective imaging system. Extensive sky coverage by professional robotic surveys are producing asteroid discoveries at ever-deeper magnitudes. Existing telescopes at Crni Vrh Observatory were not sufficient anymore to be competitive both for follow-up and asteroid discovery. To comply with increasing technical requirements, we built a 60-cm, f/3.3 Deltagraph (Deltagraph is a commercial name for the wide field optical system, designed by Astrooptik Company).
Figure 1: 60-cm Cichocki Sky Survey Telescope at Crni Vrh Observatory
The 60-cm Deltagraph is custom made, advanced technology, wide-field imaging system, designed for sky survey applications. All functions (pointing, imaging, focusing, filters exchange) are controlled by a single computer program and the system may work unattended from dusk to dawn. It may be operated either locally or remotely over the internet.
Primary mirror 600 mm, f/3 paraboloid
Coma corrector Wynne 3 lens with 1.1x optical power
System focal length 1989 mm
Diameter of corrected field 100 mm
System resolution 0.93 arc sec with 9um pixels
Field of view on 4k x 4k CCD 1.0o x 1.0o
Telescope mount fork type, friction drive + servo motors
Limiting magnitude in 1 min. exposure 20.0m with V filter

Table 1: Basic system specifications

Optical components were made and tested at LOMO and delivered by Astrooptik Company. All other components, the telescope mount, optical tube, filter wheel, flaps and servo motor electronics were designed by observatory group, based on experience of our previous projects. The project funding has been covered in large by group members, The Planetary Society Shoemaker Grant 2000, and by sponsors. The telescope was assembled and tested in a local workshop before the final installation at Crni Vrh Observatory.
Telescope Design
The telescope projects were made with help of advanced computer programs that support 3-D design and structural analysis as well as various simulations. These enable carefull design of all parts with telescope final properties being well known in advance. In that way, we avoided possible construction errors and reduced the manufacturing time and overall costs.

3-D computer model of a 60-cm telescope mount.
3-D computer model of primary focus assembly.
Figure 2: Computer models of a 60-cm Deltagraph.
Telescope design is compact and of low weight. We used sheet metal to obtain the structure of a sufficient strength, free of flexure and vibrations, while keeping the overall weight to minimum. Stainless steel and aluminum were used on optical tube assembly and many other delicate parts for minimum maintenance and lower mass. The overall weight of a working telescope is thus only 500 kg. Carbon tubes were introduced for temperature stability of tube assembly structure. This will help to keep the distance between the mirror and coma corrector as stable as possible, to avoid overnight focus changes. Motorized focusing device enables automatic refocusing of the system at any time during the observing session.
Optics
Deltagraph is a wide field optical system and consists of a 600 mm, f/3 parabolic mirror and 3 lens Wynne corrector which produce images free of coma and other aberrations over the wavelength range 400nm to 700 nm. According to spot diagrams, we obtain stellar images of less than 5 m diameter inside the field radius of 20 mm and less than 20 m diameter at the field radius of 50 mm. Stellar spots are thus much smaller than CCD pixels across the majority of 100 mm field diameter. Back focus is rather small (63.8 mm) which implies the usage of very thin filter wheel.
Figure 3a: Spot diagram for 60-cm, f/3.3 optical system. Box sizes are 20 um x 20 um. Our current CCD pixels are 24 um x 24 um which imply that the optical resulution of a 60-cm system is not fully exploited. Diagram courtesy P. Keller, Astrooptik Co. Figure 3b: True color image of M33 spiral galaxy, taken on October 28, 2002 is composed of BVR (180, 120, 120 sec) images using Gimp image processing software.
Beside the CCD imaging, the system may also be used for wide field classical photography. The field of view on a 6x7 film format is 1.6o x 2.0o. Due to the placement of field corrector inside the optical tube, the telescope can not be used for visual observations. The system vignetting is low, with light drop at the edge of a 50 mm field radius being less than 20%.
Telescope Mount and Tube Assembly
Imaging of asteroids, comets, NEOs and other objects across the sky requires a mount design that is able to access any part of the visible sky with minimum time delay. Another requirement was that telescope tracking system should be able to follow the objects with the accuracy of +/- 1 arc sec over the period of 10 minutes. Otherwise, we need to make another system for precise telescope guiding which would increase the costs and make the system less easy to use.

To meet these demands, we made a telescope mount with fork-type friction drive system and with servo motors on both axes, enabling precise tracking as well as fast slewing with up to 5o/second. The polar and equatorial disks are driven by 1:24 and 1:32 rollers, respectively, through 1:180 worm gears and servo motors. Unguided exposures of up to 10 minutes duration are possible. Also we added motorized flaps driven by Belimo actuators which prevent the mirror and corrector getting covered by dust and moisture when the telescope is not in use. The mount base, forks and lower part of OTA were made of thin sheet steel and shaped on CNC machine for maximum precision.

Figure 4: On site assembling of the telescope mount. Mounting the RA disk. Figure 5: Assembling of the telescope mount. Mounting the OTA main frame.
Extra care has been devoted to perpendicularity of polar and equatorial axes which have been carefully adjusted to +/- 0.5 arc min, using laser technique. The system has two level limit switches for secure unattended operation.
Figure 6: Final inspection of primary mirror assembly before installation on the telescope. Figure 7: Performing faucault test before the first light.
The primary mirror support was designed with Plop computer program. We used an 18 point classical flotation and radial support system of our own design that we applied on this telescope for the first time. A fan has been added at the bottom of a mirror cell, which helps to speed up the mirror cooling, down to ambient temperature.
The system primary focus assembly contains optical corrector housing, motorized focusing device, four position filter wheel and CCD. We use standard Bessell BVR (W) 50 mm diameter filter set from Omega Optical. Focusing and filter wheel control is fully motorized and computer controlled. The 1k x 1k Finger Lake Instruments CCD camera has a thinned, back side illuminated sensor. Camera specifications are given in table 2.
Figure 8: 3-D computer model and picture of finished 60-cm telescope primary focus assembly (right). Figure 9: primary focus assembly with field corrector. Filter wheel and CCD are already mounted on the carriage ring.
Because the corrector lens distances are critical, to meet the desired degree of image correction, the housing has been milled to high tolerances. Special focus carriage device is driven by stepper motor and carries filter wheel and CCD camera.
CCD Sensor TK 1024x1024 pixels, 24 um x 24 um pixel size, thinned, back illuminated
Quantum efficiency 80% @400nm; 93% @550nm;
89% @700nm; 49% @900nm
Cooling two stage Peltier at -50oC below ambient
Dark current 2.0 e-/pixel/sec dark current @240K (-33oC)
Read noise 16 e @ -30oC
ADU Conversion 4.5e- per ADU
Field of view 42 x 42 arc min
Data format FITS (16-bit)
Limiting detection about 20.0m in one minute exposure with V filter

Table 2: CCD camera specifications

The optical system has been laser collimated. Final tests on stars show that the optical components are precisely aligned. Further tests also show that the focus is very stable and independent of temperature variations or optical tube position (flexure).
Telescope Electronics
Overview
Figure 10: Telescope control electronics
The main purpose of the electronics is to accept commands from the main computer and to control the servo motors. It consists of several units. The power stages of the servos are two Epsilon Eb Digital Servo Drives from Emerson Motion Control. The main feature from the constructor point of view is that they are controlled in the same way as simple stepping motor controllers - by a direction signal and a step pulse. So the rest of electronics does not need to worry about details of servos and can be used for steppers without any change. The main processing unit is built around the Rabbit 2000 microcontroller, which has a command set based on the old Z80 processor. The program for it written in the Dynamic C language which is an extension of C which allows execution of seemingly parallel programs. Rabbit processor is the one which communicates with the main computer using either RS-232 or RS-485 interface. On a separate board, on which the main unit is piggy-backed, there are further three Atmel AT90S2313 microcontrollers which are used as intelligent precision pulse givers - one for the sidereal clock and the other two for the right ascension and declination axis. There is another board using AT90S2313 which only takes care of switching the telescope electronics on and off through a solid state relay. The telescope can be switched on and off either using push buttons on the front panel or remotely by computer. This board also has input for the Dallas DS18S20 temperature sensor to control the possible overheating of the electronics. Finally, there is a power supply for 5 V and 24 V.

In a separate box close to the telescope there are another two Atmel microcontrollers which are responsible for the filter wheel, the focuser motor, the mirror flaps and for the mirror ventilator. They communicate directly with the main computer using the RS-485 interface.

All in all, there are 7 microcontrollers in the system, which control all together 6 motors and keep track of their state so that the control program running on the main computer actually does not need to keep track of any current parameters of the system.

Main processor
Figure 11: Rabbit2000 microcontroller
The main processor serves as a connecting point between all other parts. It is connected to a 4-line LCD which constantly shows the telescope position both in sky and alt/azimuth coordinates, as well as the current sidereal and UT times. The processor is autonomous. Once the telescope coordinates are adjusted, they are not lost even after the telescope is switched off, because of the backup battery. The telescope can follow the sky at an virtually arbitrary rate on both axis and can be slewed either using commands from the main computer or from a hand paddle which is connected to the main processor through optocouplers.. Movement by the paddle can be either at 1x sidereal rate, at 16x sidereal rate or slewing at up to 2 degree / second on both axis. Slewing by the computer is set to a maximum of 5 degrees / second, which is achieved after a 5 second linear ramp. There are several security elements in the program to help protect the telescope from hitting the stop switches at the full speed. For example, the telescope can not be slewed to below the horizon using a command from the computer - only with the hand paddle. In a case when the coordinates would be lost for some reason there is a mercury switch which gives a signal when the telescope is below the horizon, so both the slewing and guiding are switched off.

There is a constant communication between the main processor on one side and the pulse-giving processors and the switch-on processor on the other side. So the system is constantly checked for possible errors and can be stopped if such a need occurs.

The communication with the main computer uses a protocol of 9 bytes, last of them being a CRC code. From the computer point of view the commands and the answers are in the form of ordinary ASCII strings, but before the transmission they are cut into shorter pieces and then assembled on the receiving side. This protocol was adopted because on the same RS-485 network there are other processors, too. The program written in C is about 4300 lines long.

Pulse generators
Figure 12: Atmel AT90S2313 microcontrollers
One difficulty of the telescope control is the wide range of frequencies with which the pulses for the motors are given. They can range from an order of 1 pulse / second to follow an asteroid or a comet to about 100000 pulses / second for the fastest slewing rate. It is also desirable that the both axis are moved at the same time. The easiest solution to this was to dedicate a separate simple, but fast microcontroller for each telescope axis. They do not just give pulses at the specified frequency but also keep the coordinates and take care to stop when the stopping switch for a given direction is in contact. Another pulse generator serves as a precise sidereal clock which interrupts both the main processor and the right ascension pulse generator. This in turn keeps the coordinates in right-ascension format meaning that it automatically compensates for the Earth rotation while slewing the telescope.
On/Off switch
With the ambition to have a remotely operated telescope a simple electromechanical switch becomes inadequate. Therefore we constructed special electronics which controls a solid state relay to which the power stage is connected. The motors can be switched on or off either by pressing the buttons on the front panel of the electronics box or by a computer command. The main processor monitors the state of the switch and immediately responds to special states such as motors off or temperature overflow indicated by the temperature sensor which is connected to the switch. So, when the motors are off, no pulses are given and consequently the coordinates are not lost.
Small motor controller
Figure 13: Electronics for the filter wheel, focuser motor, mirror flaps and mirror ventilator.
The filter wheel and the focuser are moved by small stepping motors which are driven by Motorola MC3479 IC. The book-keeping job of the coordinates and the limit switches is taken care by the AT90S2313 microcontrollers. Additional free pins are used to control the mirror flaps and ventilator.
Operation and Software
Figure 14: System operation diagram
The telescope server accept commands from the TCP/IP network and execute them. When we want to move the telescope to a desired position, we simply send "point" command with RA and Dec coordinates and telescope slew to the desired position. On the same way, we send command to a CCD camera and obtain images. Each image has a header, containing various data such as date-time of acquisition, duration of exposure, filters, WCS field coordinates and other information about the observation.

These commands can be used in a program script to perform specific tasks, such as automatic image calibration acquisition, (flat, dark, bias), focusing, filter exchange, field scanning for searching asteroids and comets. Operation of a imaging system is controlled by "watchdog" program which alert the operator in case of any hardware and software malfunctions, or sudden cloud cover.

Calibrations (pointing and tracking, optics collimation)
In order to evaluate the final imaging system performance, we measured the telescope pointing and tracking accuracy, optics collimation by measuring the star profiles as well as tracking ability on fast moving NEOs.
Figure 15: All sky pointing accuracy measured on a sample of 25 fields across the entire sky. Raw pointing accuracy is beetter than 120 arc sec RMS (All Sky). With software correction, we obtain 30 arc sec RMS (All Sky). By using Image Feedback Correction (IFC), we achieve pointing accuracy better than 1 arc sec RMS. Figure 16: RA drive periodic error (without corrections) during the 16 minute time span is +/- 1 arc sec.
We collimated the telescope by using laser technique. Star profiles were analyzed with IRAF software accross the whole 42' x 42' image plane. The star images are very uniform in quality, so we may conclude that the optics collimation was well done.
Figure 17: Contour analysis show the star images are fully symmetrical. Figure 18: Radial profile of the same star.
The telescope is able to follow any fast moving object (asteroid or comet). We simply enter the rate and direction of motion and telescope start to follow the object.
Figure 19: Demonstration of telescope tracking accuracy on fast moving NEO asteroid 2003 JD11. The NEO magnitude was about 18.5.
Calibrations - Photometry
As to perform precise CCD BVR color photometry, we calibrated our photometric setup with Johnson-Cousins photometric system. For this purpose we obtained color transformations, using standard stars in M67 open cluster. Transformations were done for the combination of Apogee Alta U9000-HC CCD camera and Astrodon 50x50 mm BVRc filter set. We obtained several BVR image sets of M67 open cluster on 2012 Mar. 15 with 60-cm Cichocki telescope under good conditions. Images were then combined and processed with IRAF CCD package. After obtaining instrumental magnitudes we calibrated them with various photometric catalogues, available at WEBDA Database Service

The following color correction equations were derived: 


V = v + 0.017*(B-V)
B = b + 0.063*(B-V)
V = v - 0.001*(V-R)
(B-V) = (b-v) + 0.046*(B-V)
(V-R) = (v-r) - 0.035*(V-R)
V magnitude calibration, depending on (B-V) color index of stars in cluster V magnitude calibration, depending on (V-R) color index of stars in cluster
(B-V) color calibration, depending on (B-V) color index of stars in cluster (V-R) color calibration, depending on (V-R) color index of stars in cluster
True color image of M67 open cluster, obtained on 2012 Mar. 15 with 60-cm, f/3.3 Deltagraph telescope and Astrodon BVRc photometric filters. Color image was composed from 3x30s V, 3x30s R and 3x60s B filter exposures using IRAF and MaximDL image processing software. Copyright © 2013 Èrni Vrh Observatory.
This research has made use of the WEBDA database, operated at the Institute for Astronomy of the University of Vienna.
Photometry Statistics and Deep Imaging
Figure 20: Stellar field centered at RA=02:21:16 Dec=+45:56:46 was taken on November 10, 2002 with 1k x 1k CCD and V filter. Exposure time was 60 seconds. Camera cooling was set to -20oC. After the calibration with Tycho-2 catalogue, we obtained the frame V limiting magnitude is 20m. See also the photometry statistics graph (Figure 21). Figure 21: CCD photometry statistics for stellar field at RA=02:21:16 Dec=+45:56:46. Standard deviation magnitude error is plotted against the V Johnson standard magnitude. Theoretical predicted curves for photon, sky and camera noises are plotted over the measured data points. The red curve represent all theoretical noises summed together. Measured data well conform to theoretical curve.
Using a 4k x 4k CCD camera device with 9 micron pixels we obtained a series of 12 five minute unfiltered/unguided exposures of a sky field in Perseus, centered at RA=18:10:00 DEC=+50:00:00. All images were then dark subtracted and added into a single 1-hour exposure composite image. Close inspection show hundreds of faint distant galaxies accross the whole 62' x 62' field.
Figure 22: One-hour composite deep image taken with 60-cm telescope and CCD. Clik for a full resolution 4k x 4k picture (2.5 Mb). Stars down to magnitude 23 are visible together with hundreds of distant galaxies. Note also uniformly sharp star images up to field corners. They demonstrate very good telescope optics quality.
Scanning Along the Great Declination Circles
Since most of the activity at the Crni Vrh observatory is dedicated to a wide field search for new NEOs and comets, it is of utmost importance to maximize the area searched. With the existing CCD camera and no filter we actually reach under ideal conditions the limiting magnitude of 20 in 15 seconds, however, with the camera readout time of about 25 seconds it is clear that having so short exposure times leads to very unconvenient ratio of on-sky and dead times. We came to conclusion that the best solution to this problem is to keep the camera shutter constantly open and use so-called drift scan mode where lines of CCD pixels are read out at the same rate as the sky image travels over the CCD. The drift scan mode is usually used on configuration when the telescope is static and the CCD sensor is oriented so that the natural sky rotation moves the image in the direction of the readout register. Although such a method is convenient in a sense that the telescope does not need to be moved and thus does not need to be very accurately built, there are at least two serious disadvantages:

  • The readout time is strictly dictated by the telescope focal length, size of the CCD pixels and the declination at which the telescope is pointing. For our configuration the exposure times would be too long, so that we would not gain anything in our sky coverage.
  • This method is only useful in a rather narrow region around the celestial equator. Outside this region the differential drift between the imager sides towards the equator and towards the pole becomes too large and the images become distorted. In our case even at declination of only 10 degress the differential drift already exceeds 2 pixels and and higher declinations the problem only becomes worse.
Figure 23: sample image with M97 planetary nebula, obtained during the telescope scanning along the great declination circles. Three scans, each of 25 second duration were stacked with Fitsblink software to obtain this 75 second composite image. Bright horizontal tracks, traversing bright stars are due to camera readout artefacts in drift scane mode.
Exceptional mechanical quality with which the Cichocki telescope is built and the flexible electronics with which it is driven allowed us to rotate the problem by 90 degrees and to use the method of scanning along the great declination circles. Of course, in this mode the telescope must track precisely in both axis. However, all previously detailed problems suddenly disappear: the exposure time can be arbitrary (down to the readout time of the camera) and since the scanning is along the great circles, it can be performed anywhere in the sky.
Using the new method we were able to increase the sky coverage by a factor of 2, covering about 22 square degrees per hour using a strategy of three exposures per field of view. The results were more than satisfying: in only one year of using this mode we found 4 NEOs, in spite of exceptional sky coverage by the professional survey programs, which operate under better skies and with wider fields of view and are therefore likely to find new NEOs before us.
Home | Copyrights © 2002-2010