Kinematic model for a 5axis mill with universal (nutating) head

From ChatGPT an explaination of teansposed matrix

Please see attached picture for the rotary head configuration.After removing the C axis from the table side and replacing B axis with C axis on the spindle side, C axis looking good in simulation, but only when Rotary A is at home position.After changing the angle of Rotary A, TCP start to create narrow oval.Do I have to change (Rotary A) transformation matrix as well.     
 

I can't read the numbers in the data sheet but if your nutation axis is not 45° then you would need to change the rotation matrix for A using the formula in my documentation.
I would suggest to start with the A axis (C = 0°) and work backwards. You will also need to think about which geometrical offsets are going to be relevant in your particular assembly.

[edit]
This section is where you will need to adjust for angles other than 45°.
 
Note that in your case you may not be able to use a simple algebraic expression like 1 over the square root of two.

He provided PDF that is readable at the end of the post, and the B axis is indeed at 50° 

Do I have to use decimal numbers or is there any other solutions?

I would first try using the exact algebraic expression (ie sin(50°), cos(50°)) and see what the rotation matrix looks like.