Electric shock through live apparently earthed metal face plate.

how did you verify the earthing of the back box, and to what ? Connection verified to MET may or may not mean connection to the supply neutral.

What did an earth fault loop impedance test of the metal box reveal?

Z.

First guess would be lack of an Earth loop - perhaps on the supplier's side rather than in the installation. How back 'back' did you check the continuity?

It's not unknown for people to connect the Earthing conductor to the Earth terminal on the cut-out without checking that it's actually connected internally (and it's quite possible it isn't if the supply was intended to be TT - for a farmhouse that is plausible).

Other question was what else were they touching at the same time (there must be something to create the voltage difference to give a shock)... the other possibility is that's live and the switch is Earthed.

   - Andy.

Then there are some really nasty but hopefully rare possibles, like L-N reversal. The absence of RCD means that a crummy earth, say 10 ohms or more, perhaps from plumbing only,  may not actually operate any ADS, but it should show up in the electricity bill, and tingles from anything else earthed on the same side of the high Z.

The meters will tell all. There may be a reason to replace the CU as well, but for this problem it would not be my first action. A high current L-E and L-N loop or PSSC test probably would. 

I must admit 'sold by an electrician' does not cut much ice with me -some of the worst wiring I have seen is on home jobs by the trade -the old joke about the cobblers' children running barefoot is  sadly often true, in that some folk are too busy out earning to get round to doing the jobs at home so it becomes a pile of temporary patches of bits left over from that other job.. Do not get me wrong, not always all by any means, some are works of art,  but some you do wonder.(and if I'm brutally honest some of my home stuff is in that permanent temporary category too, especially the workshop.)

Mike.

Hi Gary, this may be of no help but I once dealt with reports of shocks from a light switch on the way into a  golf course workshop  with a metal door. Metal clad light switch with metal conduit from metal trunking, first things first checked all earth continuity of everything, Zdb, RCD test etc(the db was not far away and had an upfront 30mA RCD) so then I gave it a bang test walked in the door switched the lights on and got quite a snap between door and switch, more like an electric fence jolt than 50hz direct contact, which may have been due to the lights being fluorescents?, the cause on closer inspection with a DMM there was a thin greasy/metallic contamination around the plastic switch inserts with 240v between the switch module and earth, I replaced the switch with an outdoor type plastic switch, job jobbed :)

After i had removed the face plate i noticed a socket on the wall pretty close and did a simple continuity check over to the lug.  I did a loop test at the socket to confirm a decent earth connection , this was around 0.6Ω. 

I will be doing an EICR in the new year and they have asked me to quote to upgrade the CU's as well.

Old farm house, extension, dedicated PME supply? First I would measure Zs at the fault point. 

And confirm that it was low enough to be PME, not just some cable running off of a possible TT supply from the old part of the building.

Having no connection back to neutral via the MET or being TT would explain the reason for the circuit breaker not tripping even though there seemed to be a reasonably good connection to earth. If the MET doesn't actually have a good connection to earth then the resistance could easily prevent the circuit breaker from ever tripping.

Sold by an electrician and no RCD's? If that was not a very long time ago, then it's time to expect more problems. A test in 2021 and no records? that should be cause for concern. The owners should have been given paper copies at the very least.