Response time of residual current device and sampling times

[blue_diode]

The table below is from this document: Schneider RCD

When the maximum response of the device is 40 mS, does it measure current doing RMS or peak detection? The document does not explicitly say it. However, in general, RCDs should trip quick (below 50 ms) when protecting humans. My question is: Is a time below 50 ms enough to perform RMS and trip the device? Let's say that we have a digital RCD that uses a microcontroller and ADC. I think the only way to trip that fast is comparing instantaneous values to a threshold. But what if the signal is noisy and deciding based on inst values cause nuisance tripping? In this case RMS would be better a measure at the expense of being slower.
My second question: In 40 ms I have 2 cycles of 50 Hz and like 2.4 cycles of 60Hz. In this case the RMS measurement for the 60 Hz would not be accurate because is not a integer number of cycles, right?

 

[KlausST]

Hi,

Although I did a 2 Ch, 1600 smpls/s, 800 times/s true RMS calculations (V, I) plus P, Z, I_DC, V_DC, with 16 (15) bit output precision with an 8 bit AVR... I don't think that RCDs do it this way. I also did fast_ response RMS tools that are stable to 16 bits output RMS within 30ms
I think RCDs do it without semiconductors. But I'm not sure.
They don't need to care for waveform distortion, nor for high precision. It doesn't make much difference whether it is 29mA or 31mA (2mA/30mA means 6% or 4bit precision) ...
It just needs to be fast enough and reliable. I guess the tests are made with fairly pure sinewave.
Maybe 30ms magnetics reaction time and 10ms mechanical delay.

I guess there are informations in the internet how RCDs work.
I remember (thirty years ago) there was a (common mode) coil where all live wires and the neutral wires are wound around, plus an output winding that carries the difference current to make the switch to trip.

Klaus

 

[FvM]

It just needs to be fast enough and reliable. I guess the tests are made with fairly pure sinewave.

Thats's what the standards, e.g. IEC 61008-1 require, except for DC sensitive devices that are tested with pulsating DC. Measurement methods (e.g. peak or rms) to be used in the RCCBs aren't specified.