1

I am wondering if it is possible to trip an RCD safely with a custom circuit.

I know the RCD detects current leaking to ground, so could a relay and a resistor be used to short the active to ground? I want to make a form of emergency stop that can be triggered remotely or by an MCU.

If this is indeed possible, is it unsafe and or bad practice to do this?

ocrdu
  • 8,705
  • 21
  • 30
  • 42
jdm
  • 63
  • 4
  • 4
    That's exactly what the "TEST" button does. – Dave Tweed Nov 24 '19 at 23:49
  • I measured the resistor in my RCD and found it had a resistance of about 3.9k. This means it could draw 60mA. So would that mean I need a 15W Resistor? I found a forum where someone said there was only a 2W resistor in the RCD. How could a 2W handle even 30mA at 240V? – jdm Nov 25 '19 at 00:31
  • 1
    The duty cycle is very low. It only has to handle the current for as long as it takes the RCD to trip -- a few tens of milliseconds at most. – Dave Tweed Nov 25 '19 at 01:01
  • related discussion about an RCD (GFCI) test circuit: [Automatic Self-Test Circuit in GFCI outlet](https://electronics.stackexchange.com/q/402520/7036) – Nick Alexeev Nov 25 '19 at 04:02

4 Answers4

1

I am wondering if it is possible to trip an RCD safely with a custom circuit.

There is!
Short upstream Live to downstream Neutral on the RCD (with a resistor). It's what the test button does.

Never are you allowed to intentionally run current trough ground. When the RCD is faulty ground will be live!


If you must mechanically open any miniature circuit breaker type device you can buy the shunt trip unit. It's an add-on module that trips it mechanically with a small solenoid operated from 12/24/110/220/400 Vac/dc. Check the catalog of your RCD vendor.

Example: Eaton FAZ-XAA

Jeroen3
  • 21,976
  • 36
  • 73
  • 1
    I think the "with a resistor" bit deserves to be more than a side note in brackets. Do it without the resistor, and the results could be unpleasant. – Simon B Apr 15 '22 at 20:36
0

Shunt trip breaker

This is the easiest (and most official) way to do this. A shunt trip breaker sits in your panel (or on your DIN rail) and hooks up the normal way. It also has two "low voltage" wires on a pigtail. Apply 24 volts to these for 50ms and the breaker will trip.

The breaker already has a (very thick, few or even 1 turn) solenoid which operates the magnetic trip mode (intended for shock overloads such as a dead short). The shunt is adding another coil to that same solenoid, which causes the breaker to trip in magnetic mode.

You would need to find a shunt-trip breaker that also supports RCD, or just use a separate RCD device.

There are also remote control breakers which allow a shot of 24V to either trip or reset it. However your chance of finding that in an RCD device is probably ... remote.

Or... do it properly

You can't use random electronic components (e.g. RU listing) in mains wiring. You must use proper equipment/devices (e.g. with a UL listing), and you must use it according to instructions.

One of the doctrines of mains power is that you must keep mains and low voltage separated. Either the mains power goes inside the junction box and the low-voltage outside, or with a divider within the junction box.

So you need a relay that is listed as equipment (e.g. UL) and ideally has the low voltage and mains wires separate. This will likely have a 24V coil. Some relays even contain their own 24VAC transformer with just enough power rating to run the coil - simply shunt two wires and that operates it.

For instance, Aube makes relay-transformers that mount in a junction box knockout, providing mains inside the box and the low voltage terminals outside the box. Others have a relay & transformer mounted on a junction box cover, again with mains on one side and LV on the other.

As far as a resistor to induce the ground fault, again the trick is finding one listed as equipment. Fortunately, the intersection of dimmers and LEDs has created a commercial market for "dummy load" equipment -- such as Lutron's LUT-MLC - and they are implemented as capacitors.

And as far as inducing a ground fault, do not wire between load (protected) hot and ground. Ground isn't for that. Wire between load (protected) hot and line neutral.

0

Notes:

IANAL.
All care, no responsibility.
YMMV.
People may die if you get this wrong.

As various people say, you can buy equipment that does this and, yes it's possible to do it yourself.
The back of one's hand also works :-)*.

You want to use a trip current that is usefully above the "must trip" RCD current and that will withstand the worst case thermal surge - which by definition doesn't last long.

The resistors used MUST be rated for the voltage concerned.
This may involve using 2 resistors in series (especially with 230 VAC or if US "two phase" is used) unless you have resistors rated for mains voltage directly.
As this is a "fail safe" application you may wish to use two series strings, each capable of tripping the breaker - so if a resistor fails open (as Murphy can arrange on a bad day) the other still 'does the job'.

You could consider using the "test" resistor from an existing properly certified RCD. These "are likely" to be acceptably certified with respect to tripping ability, voltage rating and power rating.

Given the application and the implications of failure you should look VERY carefully at the legal, ethical and plain common sense implications of building one of these yourself. While building one should be extremely easy, getting it wrong could prove to be a very very major mistake.


*Don't try this at home:
Long long ago I decided that it would be useful to know what an RCD-tripping shock felt like.
With suitable cavalry standing by, I arranged an RCD protected mains shock across the back of my hand.
It hurt about as much as I expect mains shocks to hurt, but only very briefly.

Russell McMahon
  • 147,325
  • 18
  • 210
  • 386
0

I want to make a form of emergency stop that can be triggered remotely or by an MCU.

What you are describing would not be described as an emergency stop by any safety assessor. It has too many potential failure modes. Failure of the trip button contacts to close, a wire or connection break, a failure of the resistor or a failure of the RCD itself would all render the circuit inoperable. Adding an MCU and software is a whole other selection of potential failure modes.

True emergency stop circuits use normally closed contacts providing power to an always on circuit. The e-stop buttons have a forced contact disconnect when the button is pressed and this cuts power to the circuit directly or through an adequately rated relay or contactor. Industrial safety circuits typically use dual-channel circuits to provide redundancy and monitoring prevents reset if one of the circuits fails to switch off correctly.

See my answer to a question on safety relays for more details.

Transistor
  • 168,990
  • 12
  • 186
  • 385