Relay to control mains

[bc250]

I'm thinking to use a double pole relay to control devices powered from mains (as a remote controlled plug socket), one contact of the relay for the live wire and the other for neutral.
If the relay fails, is it possible for one contact to remain stuck (live) and the other to disengage (neutral). Is this type of failure possible? What can I do to eliminate this risk?

 

[AnalogKid]

Yes it is possible. One way to reduce the risk is to oversize the relay, such as 20 A rated contacts for a 5 A load current.

ak

 

[Diver300]

If the mains load that you are controlling is made to the normal safety standard, it shouldn't matter that if the live remains connected.

It would only matter if the load had exposed parts that someone could touch that were connected to neutral.

All mains equipment should be made with live and neutral both insulated to protect against the full mains voltage, as all it takes is a broken neutral connection somewhere for live and neutral to be at mains voltage.

When working on mains powered equipment, after turning off you should always confirm that it is isolated, either by measuring the voltage to earth or by unplugging.

 

[Tony Stewart]

Residential protection for single line (L1 or L2) loads is only done with single pole breakers because Neutral voltage rise is designed to less than 5% at shared wiring and breaker loads. Although there is no problem using DPDT, it does not increase reliability of the switch. I speculate it may decrease safety to the user who thinks the power is off if/when the line contact is welded but the neutral contact opens.

If you are asking about the risks of a live neutral, I wonder if you are aware of how to protect and de-rate relay contacts with reactive loads that may have 700% rated surge currents and arc energy that reduces switching life rapidly by orders of magnitude. For example snubbers may be internal or external resistance to plastic film capacitors such as 100 ohms + 100 nF for a 10 ~15 A relay.

Reliability also improves with more conservative ratings as suggested by ak. Using the full rating of a relay is never a good idea for long life expectancy. A well designed relay with rated reactive loads may last 10k cycles.

 

[bc250]

AnalogKid, I'm using the highest rated relay, but I have no control on the user side and the user might decide to connect an oversized load.
Diver300 yes, I assume the powered device is designed to meet the safety standards (both live and neutral insulated), but if only one pole of the relay disconnects the user might think there is no power as long as the device is not running and he might decide to disassemble the device. For example, the user might decide to control a bulb and he sees that the bulb is not emitting any light so he believes there is no power and decides to change the bulb. Although the bulb is insulated and meets the safety standards, once he decides to unscrew it, he creates a hazardous situation. Usually, domestic light commutators are installed on the live wire and once you put the commutator in the off position the bulb is not powered any more, but in my case I don't know which wire will be live or neutral (the European plug can be connected in any position), so the only solution I came with is a double pole relay which is still hazardous if one pole gets stuck.
Tony Stewart, I am using a 40A relay, but if the user decides to connect an inductive load, the relay can be damaged no matter how big it is. Just a fuse is not a solution because it should be installed on the live wire, but I have no control over wires, live can be swapped with neutral at any time.

Of course, the user would be blamed for changing a bulb without being sure there is no power or plugging an oversized load, but I think the best approach would be to design the device with some sort of protection and not blame the user.
If this situation can be anticipated, I think the best approach would be to design some sort of protection.

 

[MaxHeadRoom78]

What is the type of relay?
You should use a motor-rated contactor for this application.
The other option is to set up a Safety Relay, they are more prevalent in Europe, but slowly N.A. is making them mandatory.

 

[augustinetez]

Of course, the user would be blamed for changing a bulb without being sure there is no power or plugging an oversized load, but I think the best approach would be to design the device with some sort of protection and not blame the user.
If this situation can be anticipated, I think the best approach would be to design some sort of protection.

What you're basically talking about is designing something that is 'idiot' proof - believe me, you will be fighting a losing battle. The ingenuity of modern idiots is absolutely amazing.

 

[bc250]

MaxHeadRoom78 the project is not designed to operate motors, but only low power loads. I only want to protect it in case someone is trying to use it for motors or in case one of the contacts of the relay fails. Unfortunately for my project, a contactor would be to bulky and to expensive. A safety relay might be better, thank you.
augustinetez, this is exactly what I'm trying to do. Making the device idiot proof.

 

[rjenkinsgb]

Safety relays are generally control devices, not not power switches - typically 10A or less. And as expensive (or more so) that contactors. A basic Omron guided relay with a socket is around £60

I am curious what 40A relay you have found that is much smaller than a small contactor? The contacts and connections have to physically be a certain size to carry the current!

Be wary of advertisers info; often they are non-technical and may pick something like the short term overload current or voltage on things. Confirm any ratings from manufacturers data sheets.


Also note you should be going on the AC3 rating if the load is unknown, rather than the AC1 (thermal) rating.
Things with eg. power transformers are inductive loads, not just motors.

With an R-C snubber across the output to minimise arcing you could probably use a lower AC3 rated device, eg. 50% over rated rather than double.

Most power contactors are "positively guided" like safety relays - so you can use a normally closed contact to detect a fault if a power contact welds on; the NC does not close when the contactor releases, if the power contacts have not opened. (But verify the guided contacts with the makers data sheet).

You can use that to indicate a problem and prevent the unit being switched back on, with a bit of extra circuitry.


And, you could add indicator lights to the output? Green L-N and L-E, Red N-E (same idea as in a simple "Socket tester").

Label it to warn the user that if a green light is ever on while the unit is off, not on when the unit is on, or the red light is ever on, it should be disconnected and not used.

 

[ahsrabrifat]

A simple voltage monitoring circuit or feedback from the controlled device can detect if power is still present after switching off and trigger an alarm or cut power entirely.

 

[rjenkinsgb]

A simple voltage monitoring circuit or feedback from the controlled device can detect if power is still present after switching off and trigger an alarm or cut power entirely.

That would not detect one pole welded and the other open? The device would be unpowered but in a possibly "live" state.

 

[danadak]

Detect heat change on motor (latency a problem) , use optical interrupter on
contact arms of relay (if open frame), current transformer in each leg (range would
have to handle one contact leakage detection), detect LED on load side (newer
high brightness even produce light with leakage levels).....