Trying to mimic the procedure in this answer on a slightly more difficult problem: How to calculate the trajectory of a point on a cork screw.
I have built a matrix that can calculate it for me
$${\bf M}(\delta) = \left({\bf I}+\left[\begin{array}{cc}{\bf R}(\delta)&0\\0&{\bf S}(\delta)\end{array}\right]\right)^{}$$
With $${\bf R}(\delta)=\left[\begin{array}{cc}0&-\delta\\\delta&0\end{array}\right]$$
Being the principal matrix logarithm of the rotational group (as explained in the answer I linked to). And:
$${\bf S}(\delta)=\left[\begin{array}{cc}0&\delta\\0&0\end{array}\right]$$
If I do this, then just as claimed in answer I can verify that ${\bf M}(\delta)^{1/\delta}[v_x,v_y,0,1]^T$ does rotate one radian of $x$ and $y$ and simultaneously move 1 unit in the screwing dimension (third position in the vector: $z$).
But I am unsure how to motivate what $\bf S$ looks like. Obviously the derivative should be $1$ everywhere since the derivative of $f(t) = t$ which I want the screwing to follow (constant "speed" in z-direction) is 1 everywhere. But why does it end up in the upper right corner?
Here is the screwing motion I want to represent, constant angular speed and constant "forward" motion (in z dimension):
