# Why can a raclette grill with over 400kOhm resistance between phase and null against earth trigger a 30mA RCD in a 230VAC network?

I have a 1.2kW 230V~1AC raclette grill which triggers 30mA RCDs, probably when the bi-metal temperature controller turns it off.

Not to repair, but to get better in failure mode understanding and root-cause analysis, I started to take some measurements.

I am aware that there is a difference between applying a test voltage and measuring the current to just using a digital multimeter to measure the resistance, but I am not sure whether this is why I get this measurement, which does not make sense to me.

U=R*I gives me that a 30mA RCD should trigger at roughly 230VAC if I have a resistance from phase or null to earth of roughly 7.7kOhm or less.

However, I always measure values of over 400kOhm at room temperature.

When warming up the device manually, I see a decrease in the phase against null resistance (which is roughly 40Ohm at room temperature) and an increase in the isolation temperature, as I would expect assuming a safe design.

Taking both together, I do not see why 30mA RCDs should trigger in this case.

Please provide possible explanations/ suggestions how to root-cause this further.

• When it fails, the element probably isn't at room temperature. That 400K is not itself the problem, but it is a sign of degradation, which may be much worse at operating temp.
– user16324
Jan 1, 2022 at 18:32
• RCD is triggered if and only if current flows from line to earth, not when it's flowing from line to neutral. 30mA is too much to write down as leakage current, so something is probably broken inside your device. Line-to-neutral resistance is not really relevant, line-to-PE is the one you should measure. Jan 1, 2022 at 18:37
• Why assume it is caused by a resistance, as it could be due to capacitive coupling too? Jan 1, 2022 at 18:40
• Defects might be such that (substantial) current only leaks to earth at high voltage, which a plain multimeter in resistance measuring mode does not apply. Testing this requires a isolation tester. Jan 1, 2022 at 18:41
• If it trips the RCD only when the bimetallic strip switches the heater off, I would suspect a spark somewhere between live and earth at switchoff. Because the event only occurs when it switches, you can't measure anything useful with a multimeter when it is unpowered and not switching. Jan 1, 2022 at 20:41

A leakage to earth is not a linear resistance, it should be measured in same voltage that the grid. Probably a 500 Volts tester normally used for this purpose will give lower resistance than a standard tester. But you comment that the tripp happens when the grill gets temperature and bi-metal opens. That makes me think that bi-metals is connected to neutral of the grid, and the earth fault is on this point, or in the heating resistor close to the bi-metal terminal. When the grill is heating, the voltage from earth-leakage to neutral is low. Once the bi-metal switch off, the voltage in the earth-leakage will increase to phase voltage and current to earth will be higher. You can check this possibility swapping the plug/wires of the grill.

Why can a raclette grill with over 400kOhm resistance between phase and null against earth trigger a 30mA RCD in a 230VAC network?

• Phase to neutral resistance does not imply anything about phase to earth resistance.
• RCDs measure the difference current between phase and neutral.
• Thus, RCDs infer leakage current to earth.
• Phase to earth being 7.67 kΩ could imply a Y rated capacitor from phase to earth for reasons of EMI and meeting EMC standards.
• Because neutral and earth are fairly close in voltage, there would naturally be an imbalance current from phase/live to earth flowing through the upper Y capacitor: -

Image modified from here.

7.6 kΩ at 50 Hz implies 0.415 μF and isn't unheard of for some appliances but, realistically I have no idea whether your grill has got these components fitted internally.

I suspect your RCD is false tripping due to the defect in the RCD for this condition of an arc during thermal switch off due to heater inductance of the 1500W= 230V*6.5A heater element.

The RCD does not measure ground current rather it reacts to the differential in the line and return currents. If the coupled coils inside the RCD are not balance matched well-enough in parasitic capacitance and inductive impedance , it will force the breaker to trip.

The trip current sensitivity is inversely related to the time duration.

You have confirmed there is no resistive significant leakage to ground in the grill but not measured the sensitivity of the RCD to Townsend arc extinction or measured the duration of the arc, which is not easy to do unless you are a skilled Test Engineer.

## Recommendations

1. Call supplier and ask for a refund due to compatibility issue between cooker thermostat and RCD due to nuisance trips.

2. Test RCD for compliance. Only if you are a "diehard test engineer" Measure sensitivity to false triggers with arcs using various appliances of similar current and inductance with random phase turn off, like a portable hair drier using an intermittent plug. I estimate the inductance from the heater length, which I guess is about 1 meter and that is roughly less than 1 nH/mm so it is less than 1 uH. The arc suppression depends on the phase of the interruption and thus current as well as the gap and inductance, which is hard to test with AC, but easier with a steady DC voltage which can sustain an arc.

3. Add a 1 to 10 nF Metal Film X-rated Cap across the bi-metallic switch and/or mechanical switch, whichever triggers the false trip event. Making the capacitance large enough to quench the arc to a shorter duration, but not too large to cause excessive arcs on making contact. This also allows a small current of <1 mA thru the heat when off.

10 nF is about 300 kohms at 50 Hz But what impedance at 50 kHz or higher 5 MHz? That's the solution of suppression attenuation.

What is V=LdI/dt when dt is nearly 0? Even if L is a measly 1 uH? What is arc quench time? Longer than RCD sensitivity to resonant arcs.

• What Tony says. I have see inductive loads trip RCDs on turn off. Jan 2, 2022 at 0:34
• Very low inductive , but enough to trigger an arc, V=L di/dt.. what is dt? Much Less than the ionization time of ~1 us for a dry contact. What is the BDV.? Small gap in 1us after open.thus low BDV! Jan 2, 2022 at 16:56
• A 30mA RCD should trigger somewhat before the current is equal to 30 mA.

• A single RCD may be used for several rooms with some wall outlets.

• The same RCD may feed other appliances causing an error current of about 10 to 20 mA.

• The isolation resistance of the raclette grill between phase and earth may be much smaller when hot than the 400 kohm measured when cold.

If all these effects are combined it may look like the RCD triggers too early, but it is much better than too late after somebody was injured or even killed.