I guess various ways are used to calculate the modified motor angles.
With a single axis rotation (as has a polar mount), you cannot follow the Clarke Belt exactly/perfectly. At max you can have three points at the visual clarke belt where you fit perfectly, and at the other points you are a little bit off.
The methods differ, as to at which points (visual, and/or below horizon) you have the perfect fit.
The 42.65 is the 0-180 degrees fit (180 degrees as fictional satellite aiming, at the other side of the earth; two touching points).
Also the 0-90 degrees fit is often used for tables/charts; that would give 42.69 if I'm not mistaken (three touching points, but two still below horizon).
You could also use a 0-horizon angle fit, or a 0 to half or 2/3 horizon angle, I guess. They might be a little bit better, but I've never done the calculations on that.
Will maybe do that someday, combined with the calculation test for the best fitting rotation angle settings for USALS for a modified angles setup (as I want to check the exactness of my hypothesis on that!).
Apart from the method used, some satellite calculators use radius of the equator for their calculations, instead of 'mean earth radius' (the latter would be better, I think). These could affect the outcomes also a little bit (but max about 0.01).
I've seen a russian chart/programm somewhere where you could enter your modified angles, and check at what satellitepositions you have the perfect fit.
But all these methods result in differences in the range of less than 0.1 degrees, IIRC. So if you use modified angles instead of the traditional angles, you're far better off, regardless of which modified angles you use.
And the satellites themselves can also move about that much degrees in position, I believe.
So, which chart is the best for you.... I don't know!
Greetz,
A33
