I have always had the issue of the secondary shadow being off to one side a little but the Center dot on the primary and the focuser tube reflection were dead center. In my chesire ep as well.
I also cut a template on the 4" diagonal and marked the dead center.
When attaching the diagonal, I used my laser collimator and positioned the diagonal so the laser hit my center mark on my template.
Your above setup ensures the optical (primary mirror) intercepts the geometric center of the secondary mirror. This is NOT an optimal alignment, why? Because if you cross section a code then the geometric center of the cross-section ellipse will not coincide with the cone axis, try it on a piece of paper.
See attachment. What you have is similar to the left pic but what you need to strive for is what is shown in the right pic. Note how the secondary shadow (silhouette) is centered in the left pic but the secondary mirror itself is not centered under the focuser. Bottom line, you canâ€™t center both the secondary mirror and its silhouette simultaneously. One has to be off-center and that should be the silhouette â€“ not the secondary mirror itself.
One more data point: Squaring the focuser and centering the spider vanes is desirable to certain extent but doing so is NOT a requirement for accurate collimation.