I
think you are describing a 3 phase 3 wire supply feeding a 3 phase isolating transformer?
Is it just 3 phases going to the transformer, and 3 phases and N coming out?
Key question - what is the earthing both sides of the transformer?
Does the load connected to this transformer have an earthed case?
(I guess another possibility is it's purpose is to avoid high frequency noise being back-fed into the supply.)
Knowing what it is there for and designed to do is rather essential before diving into the testing really.
I might be barking up wrong tree but one thought process is if the transformer is to provide an IT supply maybe the outer case should cause ADS if if became live, and the secondary earthing might be connected to a rectifier to achieve this, hence the different readings.
In the meantime obviously with supply disconnected I'd do dead R1+R2 tests for each phase on the primary side, and measure between phases, and to N, and to E on the secondary side to help build a picture of what is going on.
All that to say, this is the kind of territory where extra care is needed as it isn't 'typical'. More information about it will help.
(
@pc1966 ,
@Lucien Nunes this might be one for you guys )