mechanical failure or lack of voltage suppression on inductive load?

Lost an output on a Micrologix 1500 the other day. Had several spares so I just re-assigned it. When I went to update the schematic, there was a note indicating that the same thing had been done about 3 years ago. The machine was roughly 4 years old at that point.

It's a relay output on a 1764-24BWA base driving a 24vdc, 5.4 watt, solenoid valve.

My first thought was that I should put a diode in to protect the contacts against the reverse voltage spike when the coil's magnetic field is collapsing, but I'm beginning to feel like this might be a plain ol' mechanical failure.

The valve can be cycled as many as 40 times a minute with a 450msec pulse. A rough estimate is 15 million cycles over the last 3 years.

What do you cats think?

You probably wore the poor PLC out. We use these (solenoids) all the time, at a much lesser rate, and haven't had a PLC falure.

An alternative is to add in a relay; it will relieve the PLC of most of the voltage, and make it last longer.

Had a similar situation but using relay outputs to drive solenoid coils(220V). These kept burning the relay contacts. Had a Scneider rep out and he advised using a ZBZ-VM supression unit directly across the solenoid coil. Viola problem gone. Maybe you can get a Sneider rep to source a low voltage unit. I did open one and it is just a MOV inside ( an expensive one..............)

Hope this helps

Rockwell recommends surge suppression for inductive loads, which you have there. A 1N4004 diode across the solenoid coil will do it. See this link for some relay reliability info: http://www.leachintl2.com/english/english2/vol6/properties/faq.html
See items 10 thru 15.

For that high of a cycle rate I would look for a solid state solution, a relay will always fail eventually with a high cycle rate.

Yep, she just plain wore out I would think. A solid state output would be better with that duty cycle.

monkeyhead said:

My first thought was that I should put a diode in to protect the contacts against the reverse voltage spike when the coil's magnetic field is collapsing,

Click to expand...

monkeyhead,

If you do decide to to this, be prepared for a difference in how the solenoid behaves. It will almost assuredly be of longer duration, even without changing any logic.

monkeyhead said:

I'm beginning to feel like this might be a plain ol' mechanical failure.

The valve can be cycled as many as 40 times a minute with a 450msec pulse. A rough estimate is 15 million cycles over the last 3 years.

What do you cats think?

Click to expand...

15 million is between mechanical only cycles and electrical resistance load only (no inductive load) cycles of your typical relay. Acording to this Idec PDF.
http://www.idec.com/language/english/brochure/Relay_20Selection_20guide.pdf
Cycle wise you are a lot closer to the mechanical lifetime than the electical lifetime.

RDS said:

A 1N4004 diode across the solenoid coil will do it.

Click to expand...

If you don't mind me asking, how do you calculate the size? It looks like this guy is rated for 1A/400V.

milldrone said:

If you do decide to to this, be prepared for a difference in how the solenoid behaves. It will almost assuredly be of longer duration, even without changing any logic.

Click to expand...

because the induced voltage from the collapsing field is feeding the coil through the diode?

milldrone said:

Cycle wise you are a lot closer to the mechanical lifetime than the electical lifetime.

Click to expand...

I suppose I could change to the base that has 6 FET outputs and 6 relays. Drive the solenoid with one of the FET and protect it with voltage supression.

I wouldnt drive a coil of this size direct from the PLC, there will always be suppression issues. Much better to have an interposing relay/SSR, which can be 'sacrificed' if anything goes wrong.

If you suppress the coil with a reverse connected diode then the 'off' time will be increased, probably by enough to upset your application if you are working at 40 times a minute.

Two ways round this:

Put a zener diode in series with the suppression diode, anything from 12 - 24v, this will allow the magnetic field to collapse quicker.

Overrate the interposing relay, and suppress the coil with a VDR, which will limit the voltage to a reasonable level

Since you're switching a relatively small current into an inductive load at a fairly rapid rate I'm with RDS and would also put in an interposing solid-state relay for maximum reliability. These normally have a flyback diode internally so all you would need is the relay+base. In addition, the switching time will be a little bit more repeatable (sticky or overloaded relays have caused me problems in the past) and consistent, although I have no idea whether it matters for your process