unwanted input signal on limit switch release. please help

[ArmstrongPLC]

i connected a 2.1k resistor between 2c and 2m input. this has reduced the problem from 1 in 5 cycles to 1 in 30 cycles.

ATU; when you say directly on the input, i could be wrong, but to me this is parallel between 2c and 2m. Series would be between 2c and the limit switch.
are you saying between the limit switch and 2c?

[ATU]

Yes, put the resistance in series. You want to increase the resistance in that circuit. I was using a Photo eye with an NPN output which dropped another 2-3 Volts. Since you have a contact in the circuit, you may be able to have more resistance. For me, I took a 5k pot connected the wiper to the sensor and one of the ends to the input and adjusted it until the sensor just turned on the input reliably. Then measured the pot resistance to get an idea for a fixed resistor.

[Greg]

Guys, the only thing I would caution you on, is be careful about increasing the resistor value because of the voltage drop. Tim (the engineer of the CTRIO module) said, "The connections you mention will form a low pass filter. The 1k resistor could drop as much as 12V at the CTRIO maximum input current of 12mA, so be sure your power supply is at least 21VDC.  The value of the cap will set the time constant, so 1uF would set a 1ms time constant. That seems like a good place to start. If you still have the issue, try increasing the cap size."

Increasing cap size has the same effect on time constant as increasing the resistor value with the added advantage of not increasing the voltage drop.

[ATU]

I'm just relating what worked for me. I sweated for 3 days over an injection molding press with handfuls of capacitors and a digital scope trying to clean up that signal.
After the conventional approach, I was trying everything.  Sometime you get noise and you just can't isolate where its coming from.

[ArmstrongPLC]

resistor and cap still didn't work (but there was a small improvement) so i grounded out the switch by placing a copper wire against the switch lever so it rubs on the lever, constant contact. The switch case was already grounded. This corrected the problem. i suspect it is a static discharge when the roller lever leaves the edge of the melamine board. i think this is a good application for a non contact capacitive proximity sensor.

[Controls Guy]

I don't know if you'd be able to tolerate the extra delay, but I once had a servo with an input that was turning on from noise, and I placed a mechanical relay right next to the servo with very short wires and energized it from the signal that was supposed to turn on that input.  Voila, problem solved!  I had tried all the resistors and caps (though not ATU's series resistor idea) with little to no improvement, and the relay did it.  Your application sounds like it might be harmed more than helped if there were any bounce in the relay, though, so it might not work for you.

[ATU]

Good to hear that you solved your problem.  Perhaps you may want to look at the source of the static build up. Static discharges in a dusty environment may not be a good situation. I would look at how your material handling is grounded and it might even be as simple as changing the material on your roller to something more conductive.