Normally the OL isn't fault rated so the protective circuit breaker should open first to safely break the fault current. Opening high fault levels on a typical OL will usually destroy them (and the starter)
On a dead short, live to earth, the fuse or MCB will almost always operate before the motor overload.
If however an earth fault occurs part way along a motor winding, then the fault current will be much reduced by the impedance of the sound part of the winding, under these circumstances then the motor overload will probably operate first.
Another and rather dangerous possibility is that the earth loop impedance at the motor is very high, perhaps due to a missing or damaged earth connection. An excessive earth loop impedance might limit the earth fault current to say twice the full load current. That will eventually trip the overload, but probably not operate the MCB, which may be sized at more than twice the running current.
This is very dangerous since during any future earth fault, a dangerous voltage might persist for some time on the motor casing.
I would strongly advise an earth loop test at the motor.
You need to first find out what the motor is doing with regards to OL. I am assuming single phase so read the OL setting and understand the type of starter used.
Put clamp meter on the motor circuit and read current reading to see if it is showing a small increase or high load current.
Check out motor to see if you have staled or locked rotor.Also the OL block in the starter should be set at motor running current
When you have collated the info you can then use your insulation tester to check out the Motor.
I have found over the years always best to fit a new OL block in the starter but depends on the starter.
It was a 3 phase motor. Top bearing had gone and it was tripping the overload in a few seconds of running. When doing an IR test on the motor it was to earth. The resistance to earth was 1.8ohms.
Very glad to hear that you were able to find a convincing fault, - there is a general assumption that all faults to earth are a dead short from one line or another to ground. In his case presumably the fault was from part way along the windings to the case, so not taking a high current compared to the starting inrush, as would be needed to operate the instant protection, but it still adding enough extra current so the overload trip fires after the short period that is normally allowed for the thing to get up to speed.
