40A MCB will trip, but 63A won't: Is this logic even logical?

[Wrichik Basu]

@Wrichik Basu it's interesting your electrician said that there a lot of nuisance complaints. The basic technology behind RCD and GFCI equipment is the same.
-
In the USA GFCI devices are available in a number of configurations. The most common is a GFCI receptacle. They are used in bathrooms, unfinished basements, garages, outdoors, kitchens within a certain distance of the sink, etc. They are used any place a shock is more likely to be more harmful. Also, GFCI function can be put into a circuit breaker and installed in the breaker panel. Less convenient but effective. It means a trip to the panel if it trips.

Based on the stories you told before I can imagine that. Proper working of sensitive devices are hanging on the code supporting them, and without that code they are - well: sensitive.
The question is, whether you chose to adopt the code or the mess.
I think you are trying to go for (a) code.

I believe we presently do have something similar to a GFCI receptacle: a PRCD plug top. I am seeing this very recently though, and from only one manufacturer, so pretty sure they have not been in the market for too long. We can buy a few of these over time and attach the costly appliances to these.

I have seen some discussions elsewhere on the internet that attaching an RCD downstream of an MCB will not trip the MCB in case of a short or overload; so an RCD should be connected upstream of the MCB. Is this true?

 

[rbelli1]

I have seen some discussions elsewhere on the internet that attaching an RCD downstream of an MCB will not trip the MCB in case of a short or overload; so an RCD should be connected upstream of the MCB.

This doesn't make sense. It is the standard connection for any RCD socket. The two are in series so the current will be the same no matter what order they are in. Can you be more specific other than "elsewhere"?

BoB

 

[Wrichik Basu]

This doesn't make sense. It is the standard connection for any RCD socket. The two are in series so the current will be the same no matter what order they are in. Can you be more specific other than "elsewhere"?

BoB

Was probably some website called "electronicsforums" or something like that, came up on Google search. Or I think the electricians were telling us something like that yesterday. But yeah, that's what I thought too; being in series, they shouldn't be affecting each other as they work differently. Thanks for the clarification.

 

[Tom.G]

Funny story.

About a month ago, my computer did an emergency shutdown. I spent an hour or so hunting down the problem.

The configuration is a single wall outlet feeding some outlet strips feeding printers, scanner, display, a couple lamps, router, telephone modem, external disk-drive dock..., most of them supplied by a UPS (Uninterruptable Power Supply , aka battery backup).

A pop-up immediately showed up on the screen saying "The computer will now shut down", which it did by going into Hibernation (saving everything to disk).

This is expected behaviour when the power fails.... Except that the UPS normally has 10 to 20 minutes of runtime, and lamps in the room, on the same Circuit Breaker, stayed on.

So I crawled under the desk to 'fix the problem.'

Jumping to the solution (assumed root cause):
As mentioned above, GFCI feeding outlet strips and UPS.
I surmise there was a powerline surge that tripped the surge suppression in the outlet strips. These correctly shunted the surge to the third-wire protective ground. The GFCI, noting an unbalance in current 'in' and current 'out', considered that a Ground Fault and shut off the incoming power.

Now the immediate computer shutdown:
The the UPS battery lifetime is around 3 years, and I occassionally check the estimated run-time; when it drops I reorder. I was planning that reorder in the next couple days.

After plugging everything back in and resetting the GFCI, the UPS showed a runtime of 1 minute.

New battery installed and everything back to 'normal' with UPS reporting 20 minutes of runtime.

Ain't komputing FUN?

Cheers,
Tom

 

[DaveE]

... similar to a GFCI receptacle: a PRCD plug top... We can buy a few of these over time and attach the costly appliances to these.

These devices are used to detect current leakage to ground, which might include you. It makes a lot more sense to connect them to appliances that are most likely to electrocute someone (maybe the cheap ones), or devices in a more dangerous environment (outdoors, bathrooms, etc.), than to use them on expensive things. They aren't intended to protect the things, they are meant to protect you.

 

[Tom.G]

It makes a lot more sense to connect them to appliances that are most likely to electrocute someone, ...than to use them on expensive things. They aren't intended to protect the things, they are meant to protect you.

Well, even hammers have more than one use.

The combination appeared to disconnect everything from an over-voltage condition.
That puts it in the SUCCESS bucket for me.

Remember that the surge protectors often have a fuse in series so that when/if the snubber fails shorted they don't go up in smoke or start a fire.

If you have any protected outlet strips with a pilot light on the surge protector indicating they are functional, see how many of them are still on. (often a Red LED)

Of the four here that are connected directly to the line, one has failed; that's not a particular problem because it is a 'convenience use'; all it is used for is a couple of fans and occassionally a vacuum cleaner.

Cheers,
Tom

 

[Guineafowl]

Based on the stories you told before I can imagine that. Proper working of sensitive devices are hanging on the code supporting them, and without that code they are - well: sensitive.
The question is, whether you chose to adopt the code or the mess.
I think you are trying to go for (a) code.

Pretty much what I was going to say - we’ve had whole-circuit earth fault protection for decades in Europe. An RCD is bound to ‘nuisance’ trip if the wrong type is slapped into a rat’s nest of wires. The arcing light switch tripping is not something I’ve seen. You’d need some badly arc-ed terminals to get that. But then, where’s the fault - RCD or switch?

Getting back to the OP - the max earth fault loop impedance (Zs) for a 40A ‘C’ type MCB is 0.49 ohm at 25 degC. That’s pretty low, and will need some very beefy wires to achieve on a TN earth system, pretty much impossible on a TT system. Please be careful.

To protect the installation (not user), you could spec a 100 mA trip level, and/or use a time-delay version, for example, to get around the problem above. A 100 mA trip level will bump your maximum Zs to 500 ohm, assuming a max touch voltage of 50 V.

I appreciate you’re in tricky circumstances, but I’d say the closer you can get to proper, the better.