How to prove an MCB functions correctly?

Well, it would be unusual, but you could functionally test the breaker with an injection test (provided you can access a test set that can generate adequate short circuit/earth fault amps). It's reasonably easy to push several kiloamperes through it - it may be more difficult at higher ratings


Regards


OMS

It also depends what functions you wish to test - things like arc traps will be as good as the day they were fitted, likely age related failures are de-latching current due to worn mechanics or sticky lubrication, or overheating at steady load due to contact corosion.

If all you really need is some confidence it is not either stuck or too frisky, and will not overheat in normal use, then those factors can be tested with the right current but at a far lower voltage than the real event - an adjustable  variac plus transformer, to create a set-up  not dissimilar to an arc welder, capable of being set to various short circuit currents up to  a few hundred amps would be an approach.

A C50 should fire promptly at a current between 250 and 500A, and get warm but not dangerously hot after some tens of minutes at 50A.




To do it properly, i.e, break a real multi kA fault at full voltage is probably impractical.

Something I come across a lot. Whether it is a Crabtree C50 or some obsolete MCB from Ottermill, Federal Electric, Wylex or the like. Something of that vintage is usually better built than some of the modern stuff but I do wonder how, given its age, how it might perform when suddenly called in to action. If there was a test that was easily conducted how would you know that the device remained good after the vagaries of the test. C50s in a board in an old engineering shed might be a while world away from C50s in a board in a multi-million Pound building!

It has often struck me that we have complex test kit to measure RCDs and related devices, and yet the humble, but critical MCB is not considered testable.

I can only assume that this is I presume a hang over from the days of fuse wire, where a functional test would have been destructive, and so pointless.


Certainly the instant trip aspect of an MCB, although requiring a large current, only needs it to flow for a fraction of a second, and it should be possible, at least for currents up to a few hundred amps, to derive this in a portable test instrument (consider that a car battery will deliver 100a for several seconds, and 1000A long enough to weld the screwdriver to the terminals and raise the acid to boiling. Perhaps more sensibly, the super high discharge LiPo cells used in remote controlled models can be discharged at  up to perhaps 30 C rates, so some  very modest sized 2AH cells may be pushed for more than 60A, at least for some tens of seconds.

RCDs have either a rather delicate mechanical part or an electronic part.  The more rugged MCB seems to be considered more reliable, which I guess is the main part of why they're not tested.  For simply verifying that instantanous tripping happens at a suitable current, a variable autotransformer should do the job without much draw from the mains, and wouldn't have to be very big if only rated for brief currents of <<1s: this makes me suppose that difficulty and expense were not the main reasons for ignoring MCB testing. 

Modern batteries, capacitors and power electronics manage some impressive feats in recent secondary and primary injection testers for relays (as mentioned by OMS and mapj1) so probably one could now get a much lighter and smaller alternative to a transformer if there were a market for testing MCBs as a routine task.


In Paul's situation, if not having fancy things like a strong variac and DC clamp meter, oscilloscope or similar, I'd probably check for 'instantaneous' trip across a healthy car battery, and thermal trip by connecting a cooker and well-loaded ring to the MCB, attaching an AC clamp, and pushing up to 80 A or so.  Even if this doesn't check with good measurement of time at the minimum required currents, it gives confidence that it can trip both quickly and slowly. 

Testing this sort of kit is not trivial, largely because of the difficulty of measuring the current when it flows for a very short time.

If you ramp up the current slowly enough to read the meter, the thing will probably trip on thermal first. If you set up a circuit which carries a given high current then add the MCB, the circuit resistance will change and so will the current. Yes, given an oscilloscope or constant current circuits it can be done.

I had a lot of fun many years ago at Inverkip power station proving that an MCCB was being very erratic, the instantaneous trip would sometimes trip on one current and at others not trip at more that twice that current.

The client has gone with a sensible option, and that is to renew and feed via a modern 60898 mcb. So no need to test the old breaker.

However it is interesting reading the other opinions on here, and who knows, we may need to test mcbs in the future.