Yes Eric,
these are helpful rules. I followed all of them, except for those which lets the costs explode.
I am sure the disturbances come by radio interference from the relay contact spark.
I can see them on _every_ signal with the scope.
I have no problems when switching the relay without connected contacts. Only when I apply 230V to the contacts, even without load, I can see the spikes. You can say that I switch 3cm of pcb tracks on and off.
There are no problems when using an SSR instead of the mechanical relay.
This is my solution today.
I cannot afford a fully shielded metal case.
These is really not my first project, but the first one with a Cygnal, and the first one with this problem.
I hoped that someone could explane, why those strange things happen. I switch _no_ load! I could believe it, if I were switching a motor or something. But nothing is connected to the relay.
Maybe some day I will know the reason...

Thank you,
Carsten