Hey there,
aberration of light should be corrected from on the results of astrometric reduction, as stated by @luc here (which does not require knowing the orbit of the target). However I’ve never fully understood why and it’s not easy to find good explanations for that particular context, especially because the word “aberration” is also used for other phenomenons like light-time delay or even in optics.
It’s hard to say if this is the effect you are seeing (I am not able to properly see the timespan of your experiments what is it?). Maybe try computing a gradient of the obtained range w.r.t. angular inputs (via finite differences or if you feel motivated the Gradient class) to see if the result is very sensitive to errors in the RA-Dec.
Best,
Romain.