Tracing Nuisance RCD Trips
The Problem:If you have a power board with a build in RCD, (Residual Current Device), and have experienced intermittent trips of the breaker which reset without any apparent cause then you will know how frustrating it can be to find the cause of the problem. I've recently had to deal with this, hence this article.
One Innovative SolutionThe RCD itself contains all the hardware that is needed to accurately measure the real-time individual earth leakage on every single circuit connected to it and the sum of all the leakages. The following describes a modification to a spare RCD that allows it to be used in place of the existing RCD as a diagnostic tool to measure the individual circuit leakages. The modification is fully isolated, completely safe and does not interfere with the normal operation of the RCD whatsoever. N.B. - although this minor modification is perfectly safe and cannot affect the operation of the RCD in any way, the modification must be carried out and installed by a qualified and competent person for temporary measurement/diagnostic purposes only & flash tested prior to use.
Firstly - a bit of background information:
How does an RCD work?
The RCD functions by comparing the current in the phase conductor with the current in the neutral conductor of any or all of the circuits. The phase and neutral conductors to be protected are passed through a magnetic core in such a fashion that the magnetic induction of the two conductors is in anti-phase and so if the circuit is in balance with equal phase and neutral currents, (no earth leakage), then the resultant magnetic flux in the core is zero. Any earth leakage diverts current from one of the two main conductors, (usually the phase conductor, but current can also be diverted from the neutral if it is not completely at earth potential), so the currents are no longer in balance and there is a resultant magnetic flux in the core. This resultant magnetic flux interacts with an isolated tertiary coil on the magnetic core to provide a current which releases a sensitive magnetically polarised relay to trip the RCD at a pre-determined main conductor imbalance current level, (typically 30mA).
What Causes Earth Leakage?
There are two forms of leakage, resistive leakage and reactive leakage.
Simple resistive leakages to earth from the main conductors of the sort that can be measured with an ohm-meter or a Megger, (a high-value, high voltage ohm-meter), have numerous causes. Obvious ones are the permanent ones such as a nail through a cable, or water ingress to a socket or an appliance fault, such as water ingress to a kettle element, immersion heater element or a cooker element. Insects, spiders and vermin getting into poorly sealed external power boxes can also cause problems.
Inductive Leakage: Lightning strikes and remote earth faults can cause induced currents that may trip an RCD, but unless you are unfortunate enough to be on the end of a long and vulnerable overhead power distribution system, (as I am), this is probably not a serious problem.
Capacitive Leakage: Leakage caused by stray capacitance is probably the most serious problem on an RCD protected system and it limits the number of circuits that can be reliably protected by an RCD. So, how does this stray capacitance occur and how does it mount up? Capacitance exists between every conductive medium and is proportional to the surface area of the conductors and inversely proportional to the distance between them. The alternating leakage current that flows through the reactive impedance presented by the stray capacitance is proportional to the voltage, its frequency and the value of the capacitance. The capacitive impedance is given by Xc=1/[2πfC], therefore the leakage current in amps is given by V/Xc=V[2πfC], where V is the voltage between the conductors in Volts RMS, f is the mains frequency in Hertz and C is the stray capacitance in Farads.
Where does it mainly occur?
In transformers & motors - wherever you have a coil wound on the metal armature and the armature is earthed then significant stray capacitance exists between the winding and earth.
In tightly packed earthed metal cased components e.g. between the case and the capacitor element in pfc capacitors.
In wiring - standard 2.5mm2 flat twin and earth has a measured capacitance between the phase conductor and the earth conductor of approximately 110pF per metre. [This is exacerbated by the design of the cable - see note 1]. 110pF per metre gives a capacitive leakage current of approximately 8µA per metre at 230V 50Hz in the power flat twin & earth cable.
In appliances - for example metal sheathed heating elements.
In mains input filters - It is not unknown for appliances to have a mains input filter consisting of series inductors and capacitors connected between phase and neutral and phase and earth. For example, three low energy floodlights from a reputable supplier were recently installed in one of my outbuildings fed from a spur off the RCD protected house power circuit. Any two could be switched on without any problem, but when the third was switched on, the RCD immediately tripped. On examining the floodlights the reason was obvious - there were 0.068uF class Y2 capacitors connected between phase and neutral and between phase and earth. The only solution is to remove any capacitor connected between phase and earth if you want to use the appliance on an RCD protected circuit.
The above are just a few examples - stray capacitance leakage occurs anywhere and everywhere, especially where live conductors are in close proximity to earthed conductors. All this stray capacitance adds up and the built-in reactive leakage that results reduces the margin to the RCD trip level, especially if the waveform is distorted as the reactance presents half the impedance at the second harmonic and so on and the RCD has no harmonic filtering.
Now to the nitty-gritty - how can you safely see exactly what is going on leakage-wise on every power board circuit using nothing more than a spare RCD and a multimeter that reads low AC voltages? (200mV AC). Read on....
My power board is an Ashley & Rock split board with RCD protection on the power side only using an A&R 80A 30mA Supasafe RCD type RC80/30G.These RCDs are apparently not made anymore, but the Wylex 80A 30mA WRS80/2 fits my board OK, so I obtained a second hand one via Ebay for the purpose of this article.
First of all I tested the replacement RCD by injecting a variable alternating current from a low voltage isolating transformer via a multimeter into one input leg of the RCD and noted that the trip occurred about 25mA. A bit light, possibly due to ageing of the trip relay polarising magnet - I've had one RCD where the magnet had lost virtually ALL of it's magnetism, (which prevented the RCD from being reset). It's conceivable that a severe phase-earth fault could put such a high peak current into the trip relay coil that the small polarising magnet may be partially or fully de-gaussed. That possibility is easily preventable - the design of these things leaves something to be desired.
Modifying a Wylex RCD to Read Actual Leakage
1) Using a suitable screwdriver, (Torx T9 or Pozidrive both noted), remove the two RCD front panel securing screws located between the input and output terminals and carefully lift the front panel off over the operating lever.
2) Carefully drill two small holes either side of the Wylex symbol on the front panel as indicated - these will take two pcb socket pins or two pins cut from a turned pin IC socket - whatever you have available to make two test points - drill the spacing and the diameter of the holes to suit. I drilled 2x1mm holes, 8mm apart to take a single section cut from a 14-pin turned pin IC socket:
3) Securely fix the PCB pins or turned pin IC sockets in the holes from the front so that they protrude through to the back of the front panel only just enough to make a low profile soldered connection. (Countersink the holes from the inside of the cover to enable connections to be made without intruding into the inside of the RCD).
4) Using very fine insulated multistranded wire, carefully make connections between the front panel pins and the RCD trip coil connections with a precision fine tip soldering iron. (I recommend very fine double insulated super-flexible Litz enamelled copper silk covered multi-stranded wire).
5) Carefully route the test point wires, replace the front panel onto the RCD and test the unit again with a multimeter set to 200mV AC clipped to the test points. Check by current injection as above that the test current to trip is still c. 30mA and notice the corresponding reading on the multimeter. In my case the modified unit still tripped at c. 25mA with c. 83mV AC on the meter. 30mA would correspond with 100mV AC on the test points. Perform a flash test between the test points and both main connections to ensure that the test points are fully isolated.
6) Install the modified RCD in the power board and check trip operation using the 'Test' button. You now have a real time method of measuring all of the leakage currents in the circuits as a whole or individually by isolating the relevant breakers and observing the measured leakage using a multimeter set to 200mV AC connected to the test points.
Why aren’t all RCDs fitted with this simple, cheap, safe and useful monitoring/diagnostic facility? All that is needed are a couple of isolated test points on the RCD. Every RCD protected power board should have one, along with a ‘power present’ LED indicator on the incoming tails from the meter.
In my case this useful diagnostic tool clearly showed that there was no particular appliance or wiring fault, just a general accumulation of leakage on all of the power circuits amounting to about 50% of the nominal 30mA trip value compounded by the fact that the original A&R RCD was way too sensitive when tested - it was right on the edge at 50%. Replaced with a tested 80A 30mA Wylex RCD and all OK so far - touch wood....
The phase and neutral conductors are heavily insulated where they pass through the RCD magnetic core. The tertiary winding also has its own insulation and is also physically separated from the phase conductor. This double insulation plus physical separation are assessed to produce a non-existent risk of any harmful potential appearing on the fitted test points, which should however be insulated to prevent accidental touching. Ensure that your RCD conforms to the above inherently safe construction and perform a flash test to confirm isolation of the test points from the main conductors after re-assembly and prior to use.
Disclaimer and WarningMains electricity is potentially dangerous. Do not perform this, albeit simple, modification unless you are fully qualified to do so and fully understand the principles involved in RCD design and construction. Any modifications are done at your own risk - no liability can be accepted for any event that occurs as a result of such work.
Note 1Single phase flat twin and earth cable phase-earth capacitance design problem:
I don't understand the logic behind this, (perhaps someone could enlighten me?), but single phase flat T&E cable is designed as above, (UK), so that an uninsulated earth conductor runs down the centre of the cable between the phase and neutral conductors. Unfortunately this layout is not only the worst possible layout from the point of view of phase/earth shorts, (there being only one layer of insulation separating the phase conductor from the bare earth wire), but also from the point of view of stray capacitance between the phase conductor and earth. An off the shelf batch of 2.5mm 2 flat T&E was measured to have a phase-earth self capacitance of approximately 110pF per metre. Had the cable been designed as shown below to have the uninsulated, or even better, insulated earth conductor positioned on the outside of the neutral conductor then the phase-earth stray capacitance, (and hence the reactive earth leakage), would be significantly reduced, along with the risk of phase/earth shorts, e.g:
If you find this article helpful, please click on the 'Vote' button below, thank you.
© bobbybear2 2014