3
\$\begingroup\$

Is there a standard method of detecting a stuck relay in PLC?

Say, I want to build a H-bridge using relays:

schematic

simulate this circuit – Schematic created using CircuitLab

In case one of the contacts in RLY2 gets stuck, I'd cause a very bad short-circuit, therefore before switching direction I'd want first to disengage "Enable", VERIFY that at least one contact of "Enable" has disengaged, toggle "Dir" and then VERIFY that RLY2 has switched both contacts correctly. Only then I'd switch "Enable" on.

What kind of sensors (commonly used with PLC controllers) would I use to achieve that?

Optional extra problem: PLC input contacts not rated for the PSU voltage.

EDIT: eh, I saw the short circuit problem with my imagination's eyes, then failed to reproduce the circuit that would cause it, making one immune to this problem instead... Here's my former worry (obviously most easily mitigated by turning it into the above version...) - and yes, in this case one of contacts of RLY2 stuck would mean a short circuit.

schematic

simulate this circuit

\$\endgroup\$
8
\$\begingroup\$

If you have to be absolutely sure that the relay contacts have transferred, the best bet is to use a "Force-Guided Relay", or "Safety Relay". They are made by various manufacturers. On the FG Relay, use an extra contact as feedback to make sure that the relay has transitioned.

By design, FG Relays cannot only partially transition, even if one contact in the stack is welded together, the others cannot transition.

Even using FG relays, depending on the size of the motor, as you stated in your question, I would make sure that your direction switching relay only transitions when the power is disconnected to the motor. That will eliminate some contact arcing. Put some snubbers around the actual power switching contacts.

For what you are doing, that seems about the best approach with an electromechanical solution. Depending on the size of the motor, you might do better with a small DC regenerative drive.

When all else fails (and from industry, I know it will), fusing is always a wise idea.

\$\endgroup\$
  • \$\begingroup\$ +1 for this. On the subject of safety relays - the good ones will include self-monitoring of all contacts. The self-monitoring circuit can be connected to your PLC as an alarm signal. The PNOZpower range by Pilz is an example of such a type. \$\endgroup\$ – Li-aung Yip Apr 22 '15 at 6:24
2
\$\begingroup\$

You could add an Auxiliary Contact block on your relays, most Industrial Relays have this option. Then wire Control Voltage to the Aux Contact Block, monitor it with a PLC in and a PLC Timer block (in software).

\$\endgroup\$
  • 2
    \$\begingroup\$ That still doesn't show that a contact has properly transitioned. If a contact welds, the aux contact will almost always still follow the commanded state of the relay, as the armature still operates the aux contact properly. \$\endgroup\$ – R Drast Apr 21 '15 at 23:32
  • 1
    \$\begingroup\$ @RDrast is correct. Auxiliary contacts don't necessarily follow the state of the main contacts. I have heard of accidents caused by people thinking a circuit breaker was open, because the auxiliary contact was open, when the circuit breaker contacts were still closed. \$\endgroup\$ – Li-aung Yip Apr 22 '15 at 6:26
2
\$\begingroup\$

I don't see the potential problem in this cse.

Relays are generally made with the contacts mechanically coupled so that one pole cannot stick without the other sticking. In this case, even if the direction DPDT was implemented with two separate relays and one failed, the load would simply see no voltage.

Shorting from N.O. to N.C. contacts (what, I think, you are worried about) is not a normal failure mode of a relay. The contacts are attached to the same bit of metal that swings back and forth, so even if the relay is abused by a surge current or the contacts eventually wear out they will never short N.O. to N.C.

Now, we can imagine a pathological situation where such a short does happen- maybe the 3rd shift operator fails to crimp the contact into the spring adequately and it falls off, rolls around and makes the short we're worried about. Okay, there's a short. It's not a particularly "bad" short, it's just a short- same as if some idiot poked a screwdriver across the wires. The other relay closes (probably damaging it by the surge) the (properly sized) wires experience a brief surge in current, and the properly rated fuse or circuit breaker (not shown) opens, safely interrupting the fault current. Service personnel diagnose the fault, replace the bad parts, and all is well.

You must have the proper fusing (rated to carry the load current, to interrupt the maximum fault current, and for the maximum source voltage) and proper wire sizes (surge capability is built into the recommended minimum wire sizes) for safety even with no relays in there, so that is a non-negotiable requirement.

I have not considered any potential safety situation with the motor sticking "on" but you may wish to add some redundancy to prevent that from occurring (perhaps another SPST relay or replace the DPDT with two SPDT). In that case, you might have to add a feedback input to the PLC to detect potential sticking because otherwise you would not know service was required until the second failure (which could be an unsafe condition). A better solution would be a big red mushroom-top emergency stop button rather than anything involving the PLC!

\$\endgroup\$
  • 1
    \$\begingroup\$ Normal relays (not safety or other specialty) in my experience CAN have one contact stick the the other move completely. They move a short distance in parallel and have some independent spring action to take up whatever mechanical slop there is because they're cheap to make. If one sticks, there's enough spring (again because they're cheap) to allow the other to fully transition. And (once again because they're cheap), they don't have aux contacts that can be used to detect this. I would look at a more fault-tolerant/evident circuit, even if it's more difficult to drive. \$\endgroup\$ – AaronD Apr 21 '15 at 16:33
  • 1
    \$\begingroup\$ Agree with AaronD... One contact out of a standard multi-pole can absolutely stick, generally from being welded to its mate. Sometimes other contacts on the same armature will work normally, and sometimes one or more may fail to transition properly. \$\endgroup\$ – R Drast Apr 21 '15 at 23:30
  • \$\begingroup\$ @RDrast And that will not cause a short, of course, so it's irrelevant to the question at hand. \$\endgroup\$ – Spehro Pefhany Apr 21 '15 at 23:50
1
\$\begingroup\$

Not a direct answer to the question as stated, but it solves both that problem and a bigger one:

What you have will control on/off and direction, but it's not really an H-bridge. This is an H-bridge:

schematic

simulate this circuit – Schematic created using CircuitLab

(don't energize both A's or both B's at the same time!)

My off-road winch simplifies it to this:

schematic

simulate this circuit

It's the same idea, but it combines the two A's and the two B's into a single SPDT each. This pretty much guarantees that you can't short the supply, even if a relay fails (automotive starter batteries are scary when shorted), in addition to making the controls easy. The side-effect is that the motor becomes a drag brake when off, which is useful in a winch. You can evaluate whether that's useful to you as well.

If you want to detect a failure and respond automatically somehow, you can do some electrical conversion from the motor terminals, or you can watch the motor shaft or the machinery that it drives.

\$\endgroup\$
  • \$\begingroup\$ Actually, my circuit is equivalent of your second one, except with one coil driving A and (NOT B). \$\endgroup\$ – SF. Apr 22 '15 at 0:33
  • \$\begingroup\$ @SF.: Okay, you're thinking of splitting your RLY2 into my A and B, exactly one of which is always energized (conceptually). Then you control the whole thing with another relay. Now that I see that, could you please tell me again where your supply short is? Or do you not want to short (drag brake) the motor? \$\endgroup\$ – AaronD Apr 22 '15 at 0:58
  • \$\begingroup\$ As R Drast correctly noticed... nowhere. I misinterpreted the schematics. \$\endgroup\$ – SF. Apr 22 '15 at 1:12
  • \$\begingroup\$ I believe that was Spehro in reply to R Drast, but yes. Please upvote his comment if it helped. \$\endgroup\$ – AaronD Apr 22 '15 at 1:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.