Does your servo have a 0-10 volt input, and is that where the noise is manifesting itself?
Add a .1uF monolythic ceramic cap across the 10 volt input screws at the servo.
Make sure control wire shield is grounded at the servo, and floating at the source (PLC Analog output card). Noise is RF. With one end of the cox floating, you have maximum RF signal at the float end, and zero RF at the grounded end.
Try running the control cable through a ferrite core (donut). At low frequencies, you will need 2 or 3 loops.
Return lines have resistance. During a solenoid or stepper motor event, the voltage can be a few volts above ground for several milliseconds on the return line. If this is the case, increase the diameter of the wire. Make sure nothing else is 'sharing' this 24VDC return wire on it's way back to the 24DC supply.
For grounds, establish one point as the 'zero-voltage' point. From here, understand that all metal has resistance. Noise will take the shortest path to the 'zero-point'. Is the shield or return of your control wires in this path?