How To Protect 120VAC PLC Input?

[kittydog42]

How do you wire an MOV or a resistor & capacitor combination to the input to protect it? Is it in series with the input to the terminal? I have an application where every now and then one of the inputs on a given unit will stop working. This particular input is connected to motorized gas valve which probably introduces a fairly high voltage spike on occasion, which causes the input failure. I am trying to figure this out instead of using a relay, because it would be cheaper and smaller than 8 relays. If I go the relay route, I will only add one for this input.

 

[bguinn]

AB and Phoenix Contact and I'm sure others sell relays for just such things. The ones I have used were solid state and only had one set of contacts on them. They were the size of a single terminal block and not very expensive. About 8 bucks for the relay and 8 bucks for the base.

 

[Stationmaster]

First let me say that I am actually a big fan of relays on INPUTS to PLCs as they result in a "dry contact" closure to the input. That said, I would also see merit in an appropriately sized FUSE (see PLC mfg recommendation) in series with the input, though it will require someone to change it every time it blows.

My REAL question is: How is this INPUT wired to a device in such a way that a
HIGH VOLTAGE SPIKE is generated? The DEVICE should have a "dry contact" that is fed on one side by your PLC common out---through the dry contact---and back to the PLC input. If this is NOT the arrangement, then a RELAY is DEFINITELY in order.

Your desciption ALMOST sounds like you could be talking about an OUTPUT. (?)

Stationmaster

 

[rsdoran]

How do you wire an MOV or a resistor & capacitor combination to the input to protect it? Is it in series with the input to the terminal? I have an application where every now and then one of the inputs on a given unit will stop working. This particular input is connected to motorized gas valve which probably introduces a fairly high voltage spike on occasion, which causes the input failure. I am trying to figure this out instead of using a relay, because it would be cheaper and smaller than 8 relays. If I go the relay route, I will only add one for this input.

Sounds like an output to me too, possibly analog. If analog have it ramp close, if digital you wire a MOV in parallel to the 120v connect i.e. one leg of the MOV will connect to neutral and the other to the output lead feeding the device.

If using terminals strips between the plc output's and devices look at Phoenix Contacts and others for din rail relays.

 

[kittydog42]

It is definitely an input.

I will probably add a relay as an immediate fix, but I will look into using MOV's in the future. I will try using a scope to monitor the voltage on the test bench, to see what I get with and without an MOV. That way, I can add them to the inputs inexpensively and not defeat the purpose of using a $119 PLC. Over 10 years, I have never used anything on the inputs for this application (probably over 200 PLC's), and I have only had this happen 3 times, all to Automation Direct DL05's. I never had it happen when I used Telemecanique Nano PLC's.

 

[keithkyll]

The protection device is connected across the input, not in series.
This is true for a MOV, TVS, or Snubber (47 ohm in series with .05 ceramic cap).
I would use a TVS across the input, and optionally a 2.2K fusable resistor in series. The resistor would absorb the transient, and reduce EMF radiation.
It's difficult to determine the solution without knowing the cause of failure. It would be nice to examine the failed unit.
Chapter 2 of the AD manual shows an equivalent schematic of the input circuit. It looks like caps are used for transient suppression. The maximum input voltage is 132 volts AC. I doubt that the input could handle the transient from a solenoid coil.
I also looked at the manual for the Nano. They don't show a schematic, or mention upper limits.

At minimum, you could probably use just a TVS across the input. Page 2-21 of the AD manual has a chart of devices for use on inductive loads. This is for protecting outputs, but the same info can be used to protect your input.

 

[Stationmaster]

I'm still curious about how you are getting a transient spike from a motorized valve on an input. I don't DOUBT you, I just can't quite picture it.

Another question: Is the loop out to this valve maybe more exposed to lightning or other external high voltage than the other loops on this PLC?

Stationmaster

 

[L D[AR2P#0.0]]

I sounds like the input is connected directly to the solenoid coil on the valve.

 

[Stationmaster]

All due respect....he's not talking about a "solenoid valve" he's talking about a "motorized valve" (many have limit switches built in). The OP surmises That the input failure is being caused by a spike. I'm not second guessing anyone....there have been some good responses aimed at protecting the input. I am just trying to picture a configuration where an input is connected to a portion of a motorized valve circuit that may generate a spike. Maybe I'll learn something new to be on the lookout for.

Stationmaster

 

[kittydog42]

I don't know for sure what is causing the failures, I am only guessing. The wire connected to the input is connected to the same terminal as the motorized valve. The other inputs connected to limit switches, etc., have never failed. Another possibility, though more unlikely, is that I have some units from a bad production batch. I went from one isolated failure over 10 years to two failures in two days on relatively recent units. One thing that is dynamic about this application is that the gas valve connected can be many different sizes. The gas valves on the two that failed were toward the larger end of the scale (and there are two in series on each), so that may be important as well. It isn't the first time that has ever been the case, however.

When I take it apart to repair it, I will know what failed and let you guys know, if you are interested.

 

[Stationmaster]

1. I have never seen two 120vac motorized valves wired in series.
2. I have never seen a motor lead wired directly to a PLC input (with no isolating relay).

Stationmaster

 

[kittydog42]

1. I have never seen two 120vac motorized valves wired in series.
2. I have never seen a motor lead wired directly to a PLC input (with no isolating relay).

Stationmaster

They are in parallel electrically, and they are in series on the plumbing side. I wasn't clear on that.

There is a terminal in the panel labelled "GV", which supplies power to both gas valves, and is also providing the 120VAC to the PLC input. There are three wires terminated on that terminal. The motorized gas valves are two pieces; a motorized actuator and a valve body.

The PLC application is remote I/O for a Modbus connection to a main panel, with minimal programmed logic (fail to run mode timer when not connected to Modbus network).

 

[Stationmaster]

Getting the picture now... So the input to the PLC is only gathering info on "whether the motors are getting power" instead of "whether the valves are open or closed"?

I highly recommend isolating this motor lead from the input with a relay (I know, I sound like a broken record). Motors generate harmonics. Two motors complicate things even more. If the harmonics cancel each other, no problem. But if the harmonics ADD.... well then you get your higher voltages.

Bottom line is: relays are much cheaper (and tougher) than PLCs even at $119.

Stationmaster

 

[kittydog42]

Yeah, it might be a good idea to just add a relay on the gas valve inputs from now on. The other inputs seem to be doing just fine. If I can get good results testing with MOV's or RC's, maybe I will change to that in the future. I will put that on the bottom of the to-do list (which means I will look into it around April, 2009).

 

[keithkyll]

You will find the optoisolator in the PLC is bad. The LED failed from overcurrent.
It's not clear if the EMF is only from the 2 gas solenoids in parallel, or if there is additional EMF from the motor.
The problem is the design, not the parts. Highly inductive components are connected to an input with no supression. The reason it worked for 10 years is that you 'got away with it', not a bad batch of parts. The input is getting 300V+ spikes when the valves turn off.

I'm not certain how the motor is connected in this circuit, so I can't say for sure if a supressor will be enough. If things worked before (10 years) then the surge is probably not severe and an MOV would do.
A relay is the first choice. TVS is second because it's fast. MOV third, and RC Snubber last.
RC Snubbers are best for arc suppresion on relay contacts, but not the first choice for voltage spikes.