It might not be the DNO

Two situations come to mind immediately where it isn't/

1) your customer's electrode is carrying a steady fault current to earth, and the ground around it has risen with the CPC of that building. (I fault * Ze =83V)

1a) the current to earth  is small < 30mA but the house electrode connection has failed and is very high resistance.

2) if say their network has a sub 20 ohm electrode at the transformer, but on another property has a fault live to something buried at say 50 ohms, then current in to their electrode and out at the substation will leave the substation neutral at some volts off the true earth voltage.

It could of course be a DNO fault. And even if it isn't they have the power to knock on doors and see which house is pulling the fault current and whose RCD is sticking..

Its this sort of thing that makes me very nervous of metal CUs and enclosures on RCDs etc on TT supplies any fault from Pre-RCD live to the earth trips nothing ever, just puts the bill up a bit.

EDIT To confirm which electrode is dropping the 83 volts,  you need a temporary third one that you know is carrying negligable current.
So put a non insulted screw driver or garden fork or similar into the ground well away from the building or the transformer.

Run a wander lead from there to the CPC of the house.  If the voltage is low, its not that house's electrode with the fault current. If the voltage is a significant fraction of the 83, then that property has a leak to earth, and or a poor electrode.

If the tx is a pole pig and the transformer electrode wire is accessible, and on a modern one it won't be but older poles had bare wire down the side, then the same voltage test between it and an electrode of opportunity will show where the volt drop is occurring.

Or each of the neighbours earth rods in turn.

This is a gloves on sort of test, as on a significant fault, the step voltage around the electrodes, and the  voltages impressed on the 'earthed' metalwork in the house, are potentially injurious.

Mike.

Definitely not the installation earth electrode. I suspect an uncleared P/E fault in a neighbouring property, some of which are very old. The reading P to E may be the voltage drop across the DNO electrode and the electrode in the the installation with the fault. The N to E voltage may be the voltage drop across the DNO electrode. Same issue on two former occasions both of which were sorted by serendipity! 

I suspect an uncleared P/E fault in a neighbouring property,

That is where my money would be too, a fault on the wiring pre RCD, or perhaps a faulty or missing RCD, connecting the phase solidly to something stuck in the ground - which does not have to be the house electrode or wiring CPC, it could as easily be a metal fence or a cable chafed where it enters a metal framed building or something.. 
If the fault was in pre-meter cabling , there would be no obvious clue in the form of a large bill to alert the users either.

"sorted by serendipity! "

Need to call in Serendipity again. How many folk share the transformer, and how chatty are the neighbours?

Mike

I suspect an uncleared P/E fault in a neighbouring property,

That is where my money would be too, a fault on the wiring pre RCD, or perhaps a faulty or missing RCD, connecting the phase solidly to something stuck in the ground - which does not have to be the house electrode or wiring CPC, it could as easily be a metal fence or a cable chafed where it enters a metal framed building or something.. 
If the fault was in pre-meter cabling , there would be no obvious clue in the form of a large bill to alert the users either.

"sorted by serendipity! "

Need to call in Serendipity again. How many folk share the transformer, and how chatty are the neighbours?

Mike