High earth loop impeance at Transformer

Certainly I agree about testing properly first. That's probably the main thing. Measure voltage independent of test-current, four-point as Mike said.

Presumably it's a 400 V secondary, not e.g. 690 V (for motors) which would go some little way to explaining the difference.

If the primary supply is among the weakest in the world it could add the extra 0.1 ohm seen at LV.  Like 3.3 kV primary with 10 miles of tiny overhead conductor such as 40 mmsq ACSR.  It would not be normal or reasonable to connect a 800 kVA tr to that.  At the more likely 11 kV you'd be super hard pressed to add much impedance.

0.024 ohm sounds high, perhaps an  ESI thumb-rule with an assumed source impedance and LV conductors.

Certainly I agree about testing properly first. That's probably the main thing. Measure voltage independent of test-current, four-point as Mike said.

Presumably it's a 400 V secondary, not e.g. 690 V (for motors) which would go some little way to explaining the difference.

If the primary supply is among the weakest in the world it could add the extra 0.1 ohm seen at LV.  Like 3.3 kV primary with 10 miles of tiny overhead conductor such as 40 mmsq ACSR.  It would not be normal or reasonable to connect a 800 kVA tr to that.  At the more likely 11 kV you'd be super hard pressed to add much impedance.

0.024 ohm sounds high, perhaps an  ESI thumb-rule with an assumed source impedance and LV conductors.

This is a strange thing to attempt to test. Your private transformer I presume? Where is the N-E link? Your test instrument is useless for this job close to the transformer, and conceptually wrong. Essentially you are attempting to measure the output impedance of the transformer, which at 800 kVA is very small indeed. A loop tester has neither the resolution nor accuracy to give any kind of result. UKPN has given the only sensible result available, using the transformer data. The PSSC at the terminals will be around 24 kA N-Phase or 35 kA P-P with any kind of useful HV supply.

Ignoring upstream impedance, 24 kA for 3-phase at LV terminals; less for P-P; similar or more for P-N.   Agree 0.024 sounds about 2.5x too high.   (But all this isn't very important ... the essential point is already well made, that more than a simple two-wire tester is needed for sensible result at low levels.)