Some ( heavy ) bed time reading ...
https://globaljournals.org/GJRE_Volume13/5-RCD-Nuisance-Tripping.pdf ... Although specifically written about nuisance tripping and not country specific it does include some information about various types of RCDs ( active electronic, passive electronic and purely electromechanical ) and how they react to faults in the system
Thanks. I've only just glanced at it so far, but will read it properly later. It's interesting to have a Russian perspective! One thing which initially confused me is that 'they' (apparently 'the standards') define an electronic RCD as one requiring an "auxillary ('dependant') power supply". I presume that includes the RCDs which seem pretty ubiquitous in thye UK, in which that "auxillary power supply" is an intergral part of the RCD.
As I see it, there are several approaches one can take to the 'low voltage' problem, including .... :
1...As you have often suggested, the problem can be completely removed by reverting to purely electromechanical RCDs. However, as I've said, I'm not convinced that one could produce sufficiently precise and (long-term) reliable devices using that technology - at least, without their becoming prohibitively expensive (and probably with mandated limited in-service lifespans). The great advantage of 'electronic' ones is that the signal operating the trip mechanism is dichotomised (on or off), which means that it can have 'overkill' and thus cause the trip mechanism to operate reliably. The 'analogue' system of the electromechanical RCD is, in one way or another, totally dependent on very precisely overcoming some sort of spring, and friction. To make a mechanism which did not operate at, say 25mA imbalance, but which reliably did at 30mA (and would continue to be reliable for years, despite probable changes in friction etc.) would seem to be far from a trivial challenge.
2...Use 'active' RCDs. However, the article suggests that this is not allowed in a number of countries. The loss of power (until manually reset) as a result of a power cut or brief fall in voltage certainly is not without its problems/inconveniences.
3...Store enough electricity (to operate the electronics and trip mechanism) within the RCD, in a capacitor (or possibly battery, but that would limit device life) for long enough (only tens of millisecs) to enable to device to function satisfactorily in the event that the 'loss of L-N voltage' is due to an L-E fault close to the RCD (particularly with TN-C-S). It would seem incredibly improbable that at L-E fault would arise in an installation at the very time that there was a supply-side problem (e.g. 'loss of neutral'), so one probably does not really need to cater for that.
4... I'm not sure how feasible this would be, but one might consider supplementing the usual supply for the electronics with a supply derived from the RCD's sensor coil (separate additional winding), which would provide at least some power (enough?) in the event of a serious fall in the L-N pd at the RCD at the same time that there was an appreciable L-N current imbalance (i.e. when RCD operation was required).
People talk about providing RCDs with an earth, and deriving power for the electronics from L-E, in case of the infamous 'lost neutral'. However, in the case of TN-C-S (which is when people seem to worry most) that would presumably have to be a 'local earth' (earth rod), since the 'lost neutral' would also mean loss of the installation's normal 'earth'.
Probably more, once I have read the article properly.
Kind Regards, John