I am hoping I have understood your question. If not, please come back and correct me. 

This is what I think is the situation.
The loop impedance is > 0.53 ohms, so (assuming it is a 230V system in your place ?) the maximum fault current would be something rather less than ~ 500A . A 63 A breaker will trip on this, but not always instantly. Depending in the actual fault current it it may take several seconds.
But the  400A MCCB has some sort of earth fault detection.

What is the tripping current of that earth fault detector ?( not the 400A but the earth current)

When it is tripped is it instant or is there a delay ? -  Sometimes earth fault relays have a delay so that faults on smaller circuits have time to operate their own trips without cutting out the whole installation.

If I understand it right, then yes  the MCCB will trip for earth faults above its tripping limit, that do not cause something else to trip first.
This may be an example of a good economic design, or it may be very unwise, that all depends on what else loses power when the big trip operates and how important it is that the rest of the installation has power.

Do come back with more info if you can.

Mike

Thanks for the quick reply, the system is running 220V (Hong Kong). And that 400A MCCB is currently equipped with a delay earth fault ( the tripping current should be around 40A (if i remember it correctly) , CT 400/5A, 0.5 setting , 0.1S trip second)

Your reply actually helps me a lot :) Thanks mate!

Ok, now the more difficult bit, which Mike mentioned. Are you happy with the likely disconnection times? You may be but a better design would be to use a 63A MCCB also with an Earth trip, and adjustable overload and Earth trip times. Disconnecting the whole installation because of a 40A Earth fault is surely very inconvenient, and the potential disconnection of a L-N fault is rather uncertain? What features do the other MCCBs have, because the design overall is looking a bit dodgy? Why is the 400A Earth loop so high? If it is a measured value I suggest you try another meter more suited to high current tests. This needs more work!

Or simply provide supplementary bonding to the 63A circuit and check energy let-through.  GF protection might be difficult to add to the sub-circuit if it exists as you will also want selectivity with the 400A GF protection. Other than that change from T/M MCCB to electronic and set parameters to suit.

I understand that earth-loop is bad, this is an existing design for one of the old buildings which my coworker is taking part in replacing the 63A MCCB work, and that 400A actually wasn't included in the work. And yes, we might redo a high current test again and tinkle a bit of the earth pit grounding, hoping to get the best result .

I will do some more research about it, and thank you for your comment:)

"tinkle a bit of the earth pit grounding, hoping to get the best result ."

Is tinkling on the earth rods common practice? I can see how it might work.

" earth pit grounding"

Aha! is this installation TT ?

That is to say there is no N-E bond at the consumer premises and  no earth wire back to the transformer ?. If so you are doing well to get a half ohm earth resistance at all. In that case a cascaded RCD approach with progressively tighter levels of earth leakage protection as the you get nearer the final loads is in order.
Mike

yep it is, especially with military generator trucks, as is laying the rod in a puddle and then parking the trailer or the land-rover back wheel on it. Depends on how much of a hurry you are in .

I suspect that there is a pithy  phrase for it, but right now it eludes me ;-)

Mike

Interesting!!! I didn't know that generator tucks need to laying rod for earthing.

Normally, we clean off the earth pit and will replace the rod if required.

For old building earth pits usually have a lot of junk, and sometimes people might steal the earth tape/cpc for money, so we do redo it often. Normally, after this work, the earth-loop will improve a bit.