Temperature compensation in astroimaging setup

Precise focusing in astroimaging setup is crucial. There is a parameter called Critical Focus Zone (CFZ) that tells you how accurate you need to set focus plane to achieve best possible result. For example with CMOS camera that has 3.9um pixels and f/8 focal ratio instrument CFZ for green light is 159um (http://www.wilmslowastro.com/software/formulae.htm#CFZ). But the same camera attached to fast f/2 telephoto lens requires focus plane to be set with accuracy of 10 microns only! 

Typical Crayford focuser with attached stepper motor to control focusing remotely.

Once you set your focus point, then actually our story begins 🙂  If you are very lucky owner of telescope that is fully temperature compensated, then you can skip this blog entry and wait for another one. But that kind of instruments are equally rare and expensive. Usually the instrument is either not compensated at all or partially compensated. From end user point of view it is actually the same situation – you need to adjust focus point when temperature changes. There are few elements that affect it:

  • Telescope tube. It can be made of material that changes its size significantly when temperature changes (like steel or aluminium). It can be also made of low temperature expansion material (like carbon fiber), then tube itself will not change its size with temperature change
  • Optics. Each apochromatic or dublet lens changes its focal point with temperature, because the glass they are made of has non zero thermal expansion coefficient. The same is valid for all telescope mirrors made of BK7 glass or Pyrex. Some mirrors are made of pure quartz, or SCHOTT ZeroDur glass, and that materials have really low thermal expansion. 
  • Focuser.  This element is usually made of metal, but it is usually small device compared to telescope and it means, that it does not affect much total focus point changes with temperature change. 

If you are still reading this entry, it probably means, that focus point temperature drift is something you are familiar with. The good news it, that you can fight it if you have a focuser controlled from the PC and temperature sensor. The temperature sensor should measure instrument temperature. It does not need to be placed next to the focuser. It is better to stick it somewhere to the telescope tube, or back side of the mirror. And you should not place it close to any element that radiates heat, like dewcap heater or cooled imaging camera case. 

First thing you need to do is to determine how focus point changes with temperature for your imaging setup. You can do it during imaging session – you can note down temperature and focuser position each time you refocus the telescope. The temperature should change over the session at least 3-4 degrees, so the coefficient you calculate will be accurate enough. You need also to remember, that too fast temperature changes will cause, that telescope temperature will not be in balance with ambient, and the focus point may not be accurately correlated with measured temperature. It is especially the case for large and heavy instruments and when the temperature will drop by 15-20 degrees over night. Remember, that you need to collect these data points using one filter only! Even filters defined as parfocal have some tolerance of their thickness, and that differences will affect the data you collect. Once you have your session session data points noted down, it is time to put them on the plot. You can use your favourite software, or any online tool like Google Docs. If you own AstroLink 4 mini device, then you will find there dedicated Compensation calculator. On the horizontal axis you need to put temperature, and on the vertical axis focuser position. Then you need to calculate slope of this and the slope value will be the compensation coefficient. But let’s take a look to data points first, because they may be distributed in different way. 

First possibility is, that your data points fits straight line pretty good. It means your instrument reacts to the temperature changes in a linear way. It also means that compensation coefficient can be calculated with good accuracy, and it will work well over wide temperature range. If you see that one or two points are clearly out of order and do not fit to the line, you can remove them from calculation.

Next possible scenario is that your system reacts to temperature changes in non linear way. Then the plot you created could fit a curve of some shape. The compensation coefficient calculated in linear way will not approximate the data with good accuracy. The solution here can be to split the temperature range to smaller sets and calculate different coefficients for different time spans and use them at corresponding temperatures. In the example below you can see, that temperature ranges 10-15 and 15-20 degrees could be approximated to line with good accuracy. 

The most unfortunate possibility (but also highly unprobable) is that data points neither fit straight line, nor any curve in any reasonable way. When you calculate compensation coefficient basing on these points, it will probably not work well. You can still try to collect new set of points during next session. If the new points still indicates this behaviour, then your setup probably cannot be set to work with temperature compensation due to complex behaviour under changing temperature.

The compensation coefficient that you calculated now needs to be entered in the proper field of your focuser driver (assuming the driver supports temperature compensation). Once it is entered you can switch compensation on and from now on the focuser position will be adjusted each time when temperature changes. It is up to the driver how often it will be adjusted and what temperature change will trigger the compensation. If there is available another parameter to control this, you should put there value about 1/3 to 1/2 of the CFZ of your setup. This compensation threshold will be expressed in steps or microns.

There is one little bit inconvenient aspect of this implementation of temperature compensation, that you cannot decide when it will be made. So if you run imaging session, focuser driver decides itself when to compensate its position, and most probably it will happen during exposition. Depending on focuser quality it may lead to image degradation.  If the compensation may be triggered with external command, then one workaround for this may be a script that triggers compensation after each subframe exposition. Such script can be run after each frame (for example in MaxIm DL or Sequence Generator Pro) and do focuser compensation if it is required. Such example was described in AstroLink 4 mini compensation tutorial. This way if calculated compensation exceed the threshold it will not happen immediately, but only between subsequent subframes, and no image degradation will occur.

Clear skies!

Leave a Reply

Your email address will not be published. Required fields are marked *