Sam Ayre said:

when carrying out an EICR if you come across an unidentified circuit with no documentation then this would be classed as verification and therefor not part of the EICR process

But if the the circuit exists and is in use, then surely this is within the scope of an EICR according to the first sentence of Regulation 651.1 ?

I believe you should always test every accessible socket outlet on the circuit being tested, using R1+R2 tests, which will verify the connection and allow you to inspect them visually. The sampling rate for the inspection over five years is 20% of the socket outlet circuits of the whole installation. So you are not liable if you have not tested or inspected that circuit yet. This only applies to more complex installations. For a domestic property, you need to test everything every time.

I sense a degree of self-contradiction here: "every accessible socket" ... "For a domestic property, you need to test everything every time". What about the socket behind the Victorian triple wardrobe or bookcase?

If a socket is hidden behind a Victorian triple wardrobe or bookcase and cannot be reached for inspection, you should state in part 6 of the EICR that the socket is not accessible.

I appreciate your feedback, but I don’t think there is any need to be so meticulous about my wording or sentences (not the first time) I clearly stated at the beginning of my statement that I was referring to accessible sockets only. Do you have a particular reason for questioning my choice of words? Did you study engineering or English? 

Ouch!

Agreed!

AMK said:

part 6 of the EICR that the socket is not accessible.

Exactly 'LIM' in the same way that there's a circuit you don't appear to be able to trace/locate ('LIM') - either might be dangerous, the risk of one that's inaccessible is different to one that is accessible, but if you couldn't trace/find a circuit, one might be really pedantic and argue it's not accessible (to the best of your knowledge and belief at the time)?

I agree with you AMK.  In your average home/dwelling the 20 % sample rate is a bad recommendation and should be changed to something higher or made clear to treat in commercial/industrial is 20% and in a dwelling it should be a min of 80%.

At the end of the day if you can do 100% in a dwelling then that is what should be done.  For example in an average 3 bed house it built 100 years ago it will of proabably had

CU change 2 to 4 times

Single pendant in all rooms changed to maybe downlights in certain rooms like bathroon and kitchen

Orignial cabling done in singles in a metal conduit system with probably no CPC replaced at a later date by VIR cable and then maybe PVC/PVC cable

Home improvements by the home owner/occupier which include electrical work.  Out door flood lights is common.

This brings up many things

Backboxes for lights and sockets may have been made from wood and now should be replaced for metal or plastic if in plasted board

Downlights may or may not be class 1 or 2, they may be fire rated they may not(however for the price difference I would always recommend fire rated downlights even if there is not habitable space above)  Another common issue with downlights is single insulated cables outside of enclosure and no CPC.  Downlights can be a pain as to try and remove them can sometime cause cometic damage to ceiling.

Loose cables in Accessories

Plasticiser migrationn

The list goes on.  Thus I would say where possible in an average EICR for a dwelling try and increase the sample rate from 20% to 100%.  There are also ways to further improve on this

Check fire/smoke/heat/carbon monoxide detector for functionality and date of install and report on EICR

PAT (Portable Appliance Test)

While EICR for an average home are not very common at the moment, EICR are now common well mandatory for rentals.  The point of EICR is to see if the installation is safe or requires improvements. 

Sergio Fernandez said:

Thus I would say where possible in an average EICR for a dwelling try and increase the sample rate from 20% to 100%.  There are also ways to further improve on this

Sergio Fernandez said:

In your average home/dwelling the 20 % sample rate is a bad recommendation and should be changed to something higher or made clear to treat in commercial/industrial is 20% and in a dwelling it should be a min of 80%.

I don't think that one should be prescriptive about this - GN3 is not.

The science of sampling is anything but simple. The sample must be large enough to have a reasonable chance not only of finding a defective socket if there are any, but also not wasting time sampling everything if there are none. That would require data for different types of buildings and different ages.

A small house might have only two socket circuits: upstairs and downstairs. That means that any sample could be only 50% which would not be appropriate.

Then of course, you must not sample samples, so you would have to inspect all of the sockets in the chosen circuit.

Rather better would be to sample each circuit, but even then be careful. I have 5 socket outlets in my kitchen. I wouldn't rely on a sample of 1. If the first 3 were all perfectly sound and there were no visible sign of damage on the remaining 2, it would seem perfectly reasonable not to peer behind them.

Samples must be chosen at random. If you pick only the most accessible sockets every time, then nothing can be said about the others.

On the whole, I would say that all but the largest houses (say over 500 m²) are too small for sampling to be appropriate.

Part of the problem is not knowing what was inspected last time, or even when or perhaps if there was a last time. There is a very real risk that the same good bits that are easy to reach without ladders or moving the cooker are inspected each time, and no-one looks at the rest. In a commercial setting one may have the benefit of the last set of results, but not always.

On the other hand, I suspect that with a bit of practice the most likely to fail items can be inspected first and or tell-tale signs like woodscrews holding the switch to the wall and things that are badly corroded and an RCD that looks like it gets a greasy hand print far too often may well give an early warning that somethings need a deeper dive. 

And what is the risk - if the down-lighters have been masticed into the decor, but behind is an un-boxed choc block, how bad is it ?

I was amused, (and we turned it down and got a friend I trust to mark my homework), when we rented out a properly and the rental agent reckoned he could find someone to get it the right electrical paperwork at the drop of a hat and at price that meant it could surely take less than an hour...

At the shallow end not a lot actually happens.

Mike.

Part of the problem is not knowing what was inspected last time, or even when or perhaps if there was a last time. There is a very real risk that the same good bits that are easy to reach without ladders or moving the cooker are inspected each time, and no-one looks at the rest. In a commercial setting one may have the benefit of the last set of results, but not always.

On the other hand, I suspect that with a bit of practice the most likely to fail items can be inspected first and or tell-tale signs like woodscrews holding the switch to the wall and things that are badly corroded and an RCD that looks like it gets a greasy hand print far too often may well give an early warning that somethings need a deeper dive. 

And what is the risk - if the down-lighters have been masticed into the decor, but behind is an un-boxed choc block, how bad is it ?

I was amused, (and we turned it down and got a friend I trust to mark my homework), when we rented out a properly and the rental agent reckoned he could find someone to get it the right electrical paperwork at the drop of a hat and at price that meant it could surely take less than an hour...

At the shallow end not a lot actually happens.

Mike.

If the local council held a database of EICR and installs then people could look up the last test details.  After all, notifiable works are listed by the local councils.  This in the long run would give some kind of audit log.