From the CFD of Xray1 (90deg) we get multiple trigger signals every ~5sec for unknown reasons. 
No hint from the preamp signals.

The solution now is to have an extended blocking gate active in the trlo-config of node 2.
There is now a trigger_stretch of 2500ns for all Xray triggers:

trig_stretch(1)  = 2500ns;
trig_stretch(2)  = 2500ns;
trig_stretch(3)  = 2500ns;

This prevents the multiple triggers to reach the lmu (before_dt scalers)
However, this will also lead to a very small loss of good signal triggers that overlay within 2.5 us.

This has to be taken into account for deadtime determination.