Periodic error . . . what am I missing?

[David Jardine]

I am a very happy PHD2 user, who is getting good results (~ 0.6 arc seconds total error). I have arrived at this point by gradually learning more about PHD2 and with cautious tweaking of the settings. Mostly for my own curiosity, I have a question about a periodic error spike that is occurring at the 200 second mark. My mount is a CEM40, which has a PE of 400 sec. (page 30 of manual). Surprisingly, the largest PE spike on the FFT is at 200 sec. and there is no error at 400 sec.. When I look at less processed waveforms, I see the same error periodicity and no obvious error with a 400 sec. frequency. This makes me think that I am misunderstanding something. If it is not too much trouble, can someone explain this to me (or point me to an explanation)? This is not critical to my image quality, but I would like to understand this.
Thanks.
David

[Bruce Waddington]

Hi David.  The 200-sec period error is coming from some component lying upstream of the worm gear.  Whatever that component is - maybe some kind of spur gear that is driving the belt system - it is rotating at 2x the speed of the worm gear.  In some cases, a belt-drive system can also have periodic errors like this if the belt moves laterally on whatever spindle system it uses.  In order to measure periodic error, you should run the your guiding session near Dec = 0.  If you do that on your system, you will probably find that the error is about 2x larger than what you've shown here.  I think there is also an error contribution on your system at about 400 seconds, it is just smaller, about 0.7 arc-sec for this pointing position.  Since the LogViewer is estimating the frequency spectrum by "backing out" the guiding corrections from normal guiding operations, there is some uncertainty in period estimation. 

Since the largest error has a frequency that is an integer harmonic of the worm frequency, a high-quality periodic error correction should address it.  But that would probably need to be done by an app like PemPro, one that evaluates performance over multiple worm periods and rejects contributions from seeing noise and outliers.  But it's also possible that the PHD2 PPEC algorithm is already handling this well so you may not want to fool with it.  We usually recommend that people use their imaging results as the standard of performance.

Hope this addresses your question,

Bruce

[David Jardine]

Thanks, that is very informative. I wondered why I was seeing a periodic error that was half the period of the worm gear. I do not think that the error is coming from the belt, as the belts in this mount have a period of approximately 700 seconds. I discovered that earlier this year, when the RA belt had a problem and gave a large error with a (approximately) 700 sec. periodicity. Fixing the belt corrected that error.

The next time I open my mount, I will look carefully to see if I can identify a source for the 200 sec. error, just to satisfy my curiosity. I am quite happy with the tracking as it is currently set, so I am inclined to leave well enough alone. I am relatively new to astrophotography, so learning more about post processing probably offers the biggest opportunity for improvement.

Thanks for making such a great program. The more I learn about it, the more I like it.

David