You are here

Pointing Model Update

I have been spending some time recently testing and debugging the new 3-point star alignment code following the 1.3.1 release. I've got the code to the point where it will compute a correction based on 3 nearby alignment points and it will correctly slew the telescope to the corrected position with a counterweights-down slew. So that is relatively good news. The not so good news is that the corrected position isn't good enough and typically doesn't fall into the FOV of typical eyepieces. A 1-point slew produces a much more accurate result. This could be because there are still numerical issues to iron out and I will try to do this through more testing. However, I have started to wonder whether this approach is the best way to go for giving EQMac a pointing model.

In light of this I started doing a bit of research into how other applications such as MaxPoint and TPoint approach this. These applications work by building a mathematical model of the mechanical errors in the mount and telescope setup and then use alignment points to produce a statistical fit of the model. From the examples I have seen so far, this seems to produce very accurate pointing across the whole sky (at least that is what the examples try to show and what the marketing claims). This statistical approach is also used by many large telescope installations around the world. One of the advantages of the approach is that once you have a fit to the model, it is possible to quantify specific issues in the setup such as polar misalignment, cone error, offsets in the home position, etc. This can be used to correct those faults if desired, although the pointing model serves to correct them automatically.

I am at a bit of a cross roads with what to do with EQMac's pointing model. I am not getting along very well with the current approach, although I do not believe I have implemented it completely correctly yet. However, I do like the mechanical error model approach, more so the more I research it. So I think I will take some time over the summer to implement a prototype of the mechanical model approach and see if that gets better results than the 3-point mechanism I have been pursuing. I should then be able to release something of quality that works properly, which is evidently not the case at present. Pointing models do appear to be somewhat challenging to get right, so please bear with me on this and continue to use the 1-point model that EQMac has always had.