Stability measurement of an astrophotography setup

This article discusses stability measurements of an astrophotography setup using astro-images. By taking images in a fixed direction (in altitude and azimuth) with a telescope fixed to the base of the set-up, and by plate-solving and converting them to altitude and azimuth on JNOW, the stability of a astro-imaging set-up can be analysed at an arc-second level.

Auto-collimation setup on InFINNity Deck. The Leica measures on a first-surface mirror.
Figure 1: Auto-collimation setup on InFINNity Deck. The Leica measures on a first-surface mirror.
Since I image unguided, I was curious how stable my observatory is. After all, correct tracking of the sky depends on the stability of the setup. If the mount tilts during an imaging session, it will not notice and correct for this and the stars will get smeared during long exposures.
To get an idea of ​​the required tracking accuracy and thus maximum deformation, a simple calculation can be done. Suppose a FWHM of 4 arc-seconds and the wish to keep the eccentricity below 0.5, then the drift may be a maximum of 0.6 arc-seconds. The longest subs I take are 420 seconds. This means that in 420 seconds 0.6 arc seconds of drift may occur, which means that per hour (3600 seconds) drift may amount to a maximum of 3600 / 420 x 0.6 = 5 arc seconds.
Last year I had already performed a so-called auto-collimation measurement with a high-end total station (a surveying instrument, see figure 1), and it followed that no significant deformation occurred over the half hour that I had made the measurement (less than 2 arc seconds). That measurement required that I was constantly present (for the observations) and was not allowed to move. Since this is not inviting for a night-long session, I was looking for another way to execute such a deformation measurement.

Inspired by the work of 10Micron owner Massimiliano Chersich, I started programming myself. While Chersich used his own recording software to generate the FITS, I wanted to use standard (preferably free) software for recording and my own software for processing. Two free programs were chosen for the recording: FireCapture and NINA, but since FireCapture cannot collect and store the necessary weather data, the final choice fell on NINA. The processing required proprietary software, as only Chersich had a solution for this. That is why I recently wrote FITSalize, a command-line tool that uses ASTAP to analyse the images. FITSalize extracts the following data from the photos:
  • Lat/lon from the observatory (straight from the header)
  • The time and exposure time of the image
  • RA/DEC in J2000 (after plate solve with ASTAP)
  • Focuser position (straight from the header)
  • Temperature and air pressure (straight from the header)
  • SQM values ​​(after plate solve with ASTAP)

The Bresser Messier 90/500 achromat with ZWO ASI174MM at the base of the mount.
Figure 2: The Bresser Messier 90/500 achromat with ZWO ASI174MM at the base of the mount.
The lat/lon and time are required for altitude and azimuth calculations. Depending on whether start, middle or end of exposure is recorded, the time can automatically be corrected with half the exposure-time used (configurable via the accompanying FITSalize.ini file). The RA/DEC position of the FITS are converted to JNOW by calculating precession, nutation and aberration. The resulting JNOW coordinates are then converted to alt/azi, where the air pressure and temperature are used to determine the refraction to further refine the altitude.1 The focuser position and SQM values ​​are not important for the deformation measurements, but can be used to determine the temperature coefficient of the telescope and the light pollution of the environment.
As equipment I chose a Bresser Messier 90/500 f/5.5 achromatic refractor that I had lying around (see figure 2), but in principle any guide-scope would be suitable. The telescope is attached to the base of the mount with a Vixen dovetail clamp. The telescope was very consciously chosen not to be on the mount, but below it (i.e. on the column), since the axes of the mount can change orientation when stationary due to temperature changes and are therefore not stable enough for such a measurement. Behind the telescope I mounted a ZWO ASI174MM, which I normally use for planetary imaging. NINA was used for the recordings, with only the camera, a MGBox weather station and a simulated mount (see below) connected. As an alternative to the MGBox, the OpenWeatherMap ASCOM driver can be used (this comes standard with the ASCOM 6.5 installation). It is also important that the computer-clock is synchronised. This can be done via Windows Time Synch (w32tm) or via the Meinberg NTP Package.

Deformatiegrafiek van InFINNity Deck, gemaakt met FITSalize.
Figure 3: Deformatiegrafiek van InFINNity Deck, gemaakt met FITSalize.
To ensure that lat/lon from the observatory actually ends up in the FITS header, a mount must be connected. For the deformation measurements I chose to use the standard ASCOM telescope simulator, so that the actual mount could remain switched off. Simply set the latitude and longitude (lat/lon) and use the virtual keypad to position the simulator in approximately the same direction as the measurement is being taken and the measurement can begin (within approximately 10° to 20° is good enough).
After the required FITS are collected, FITSalize can get started, and will produce an output file (for explanation and instructions see the FITSalize-page). This file can easily be converted into some graphs in Excel with the templates supplied with the package. Figure 3 shows the result of the first deformation measurement. Both altitude and azimuth are stable within about 4 arc seconds, the steepest part is about 3 arc seconds per hour, which is still too low to seriously interfere with deep-sky imaging.
I have had frequent consultations with the author of ASTAP Han Kleijn. Han is working on the implementation of this routine into ASTAP, so that a somewhat friendlier user interface will soon be available for this type of measurement [in the meanwhile it has been fully implemented]. In addition, it offers the possibility to compare the accuracy of the programs with each other as FITSalize is entirely based on the work Meeus, J., Astronomical Algorithms, Second Edition, (Richmond, 2005), while the deformation tool in ASTAP is based on P.T. Wallace's SLALIB positional astronomy library of the Rutherford Appleton Laboratory.


[1] If a set-up is stable, its altitude and azimuth should remain fixed over time, at the most influenced by random noise in the imaging and calculations (the latter may also have some bias). So, any drift seen in altitude and/or azimuth is an indication that the set-up is not stable.

If you have any questions and/or remarks please let me know.

Home Geodesy Navigation Astronomy Literature
InFINNity Deck... Astrophotography... Astro-Software... Astro Reach-out... Equipment... White papers...
Hardware... Imaging...
Balancing system Camera centring errors Collimator construction Dome Azimuth Calculation Filter focus-offsets RC Collimation Removing lens artefacts Removing Newton-rings Solar Seeing Monitor (DIY) Stability measurements