Nice gif. What software are you using to generate the individual frames?

I have 3 elements:

-process the individual 60 second frames - I use ASTAP. This can also give plate scale and alignment of north

-Shift them to follow the predicted motion - I use IRIS as I can shift each frame an individually specified amount 

-beforehand, work out how to shift the files to follow an 'invisible' object.

 Here I use a home-grown excel spreadsheet with VB macros to: 

-import predictions from JPL Horizons or other, covering my observation window

- create an analytical expression to map the target position as a function of time. This is in the same style as Besselian Elements for Total Eclipses and can follow a curved trajectory.

- generate the script file for IRIS to open each file, shift it the specified amount, tart it up, save it. This can be thousands of lines long.

Once all aligned, I stack them as I want. Here it is Sum 30 step 5 using IRIS again, followed by movie creation in VirtualDub

HTH

Thanks for that full account. Really interesting. Had kind of hoped someone made a tool that made it easy! 

Looks a fun project to get code of my own going and look at some old C3 data.

Am hoping to image Juice tonight if visible.

One final question: Were the source mages autoguided?

No, the mount is an iOptron CEM120EC with an optical encoder. My previous mount was an HEQ5Pro with a separate guide scope. When it worked well it was great but...

Thats interesting. Thanks for letting me know. Was trying to think of sneaky ways the dim background pattern could be reduced.... using my own code rather than something like pixinsight. Food for thought.