Measuring the L-L and N-N resistance on a ring circuit is giving me 0.7 and 1.1 ohms respectively. Does this point towards a poor/loose neutral connection in the back of a socket somewhere? If so, is there a good way to locate where through testing as opposed to removing each socket for direct inspection?

On 16/07/2019 23:05, Mathew Newton wrote:

Measuring the L-L and N-N resistance on a ring circuit is giving me
0.7 and 1.1 ohms respectively. Does this point towards a poor/loose
neutral connection in the back of a socket somewhere?

Quite possibly, or something else not quite right.

If so, is there
a good way to locate where through testing as opposed to removing
each socket for direct inspection?

You could go for a "binary chop" type approach. Disconnect the ends of
the ring at the CU, and short L & N & E of one end together. Now using a
plug as a test point, plug it in somewhere on the circuit near to where
you think the middle might be. You can now check if the L to E and N to
E match. Whether they do or don't, you have just eliminated half the
circuit. Go half way again in the appropriate direction, and you can get
rid of half the remaining ones again. Should take no more than Log_2(n)
tests, where n is the number of sockets on the circuit.

Something like one of these makes the job easy:

https://www.amazon.co.uk/Kewtech-KEW.../dp/B0058HZLTK

Needless to say, make sure the socket you are testing at is on the
disconnected circuit, since test meters don't appreciate mains across
their probes when on a resistance range.



--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

On Wednesday, July 17, 2019 at 1:43:15 AM UTC+1, John Rumm wrote:
On 16/07/2019 23:05, Mathew Newton wrote:

Measuring the L-L and N-N resistance on a ring circuit is giving me
0.7 and 1.1 ohms respectively. Does this point towards a poor/loose
neutral connection in the back of a socket somewhere?

Quite possibly, or something else not quite right.

I should've said that this is a 2007-built house and so hopefully the issue is not a result of some bizarre circuit design/layout. Also, the E-E resistance is 0.9 ohms and all these readings are with a non-zeroed meter (lead resistance measures at ~0.5 ohms) so whilst they are all low it is the L-L/N-N inequality that raised my suspicions rather than the absolute values, and given the still-lowish N-N reading that's what led me towards it being a contact point issue rather than a break or similar.


If so, is there
a good way to locate where through testing as opposed to removing
each socket for direct inspection?

You could go for a "binary chop" type approach. Disconnect the ends of
the ring at the CU, and short L & N & E of one end together. Now using a
plug as a test point, plug it in somewhere on the circuit near to where
you think the middle might be. You can now check if the L to E and N to
E match. Whether they do or don't, you have just eliminated half the
circuit. Go half way again in the appropriate direction, and you can get
rid of half the remaining ones again. Should take no more than Log_2(n)
tests, where n is the number of sockets on the circuit.

Okay that sounds doable - thanks. To be fair I think there are only 13 sockets so it's not that big an installation, and one or two of those are spurs and thus wouldn't have contributed to the measured resistance (their point of attachment would of course be - and perhaps that's where it is quite likely to be?).

Something like one of these makes the job easy:

https://www.amazon.co.uk/Kewtech-KEW.../dp/B0058HZLTK

Hopefully this will be a one-off so a conventional plug with a removable back should equally suffice, particularly given the power will be off?

On 17/07/2019 09:31, Mathew Newton wrote:
On Wednesday, July 17, 2019 at 1:43:15 AM UTC+1, John Rumm wrote:

I should've said that this is a 2007-built house and so hopefully the
issue is not a result of some bizarre circuit design/layout.

In which case, yup, probably just a loose screw etc.

Also,
the E-E resistance is 0.9 ohms and all these readings are with a
non-zeroed meter (lead resistance measures at ~0.5 ohms) so whilst
they are all low it is the L-L/N-N inequality that raised my
suspicions rather than the absolute values, and given the
still-lowish N-N reading that's what led me towards it being a
contact point issue rather than a break or similar.

Yup, I would agree.

You could go for a "binary chop" type approach. Disconnect the
ends of the ring at the CU, and short L & N & E of one end
together. Now using a plug as a test point, plug it in somewhere on
the circuit near to where you think the middle might be. You can
now check if the L to E and N to E match. Whether they do or don't,
you have just eliminated half the circuit. Go half way again in
the appropriate direction, and you can get rid of half the
remaining ones again. Should take no more than Log_2(n) tests,
where n is the number of sockets on the circuit.

Okay that sounds doable - thanks. To be fair I think there are only
13 sockets so it's not that big an installation, and one or two of
those are spurs and thus wouldn't have contributed to the measured
resistance (their point of attachment would of course be - and
perhaps that's where it is quite likely to be?).

Something like one of these makes the job easy:

https://www.amazon.co.uk/Kewtech-KEW.../dp/B0058HZLTK



Hopefully this will be a one-off so a conventional plug with a
removable back should equally suffice, particularly given the power
will be off?

Yup the test sockets make it a bit easier (and safer), but working in a
controlled environment and with care, a normal plug and lead is fine. If
in doubt, you can always turn off everything at the main switch.


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

On Wednesday, 17 July 2019 09:31:40 UTC+1, Mathew Newton wrote:
On Wednesday, July 17, 2019 at 1:43:15 AM UTC+1, John Rumm wrote:
On 16/07/2019 23:05, Mathew Newton wrote:

Measuring the L-L and N-N resistance on a ring circuit is giving me
0.7 and 1.1 ohms respectively. Does this point towards a poor/loose
neutral connection in the back of a socket somewhere?

Quite possibly, or something else not quite right.

I should've said that this is a 2007-built house and so hopefully the issue is not a result of some bizarre circuit design/layout. Also, the E-E resistance is 0.9 ohms and all these readings are with a non-zeroed meter (lead resistance measures at ~0.5 ohms) so whilst they are all low it is the L-L/N-N inequality that raised my suspicions rather than the absolute values, and given the still-lowish N-N reading that's what led me towards it being a contact point issue rather than a break or similar.


If so, is there
a good way to locate where through testing as opposed to removing
each socket for direct inspection?

You could go for a "binary chop" type approach. Disconnect the ends of
the ring at the CU, and short L & N & E of one end together. Now using a
plug as a test point, plug it in somewhere on the circuit near to where
you think the middle might be. You can now check if the L to E and N to
E match. Whether they do or don't, you have just eliminated half the
circuit. Go half way again in the appropriate direction, and you can get
rid of half the remaining ones again. Should take no more than Log_2(n)
tests, where n is the number of sockets on the circuit.

Okay that sounds doable - thanks. To be fair I think there are only 13 sockets so it's not that big an installation, and one or two of those are spurs and thus wouldn't have contributed to the measured resistance (their point of attachment would of course be - and perhaps that's where it is quite likely to be?).

Something like one of these makes the job easy:

https://www.amazon.co.uk/Kewtech-KEW.../dp/B0058HZLTK

Hopefully this will be a one-off so a conventional plug with a removable back should equally suffice, particularly given the power will be off?

Sounds quicker to just unscrew the sockets and retighten the screws.


NT

After serious thinking Mathew Newton wrote :
Measuring the L-L and N-N resistance on a ring circuit is giving me 0.7 and
1.1 ohms respectively. Does this point towards a poor/loose neutral
connection in the back of a socket somewhere? If so, is there a good way to
locate where through testing as opposed to removing each socket for direct
inspection?

Visually check each socket's face, for signs of any over heating.

With the power off, you could make up a pair of (naughty) plugs with
bare wires coming out, then work your way around the ring, comparing L
to L and N to N resistances across sockets in the ring, until you
localise it, then delve into the socket for the cause.

Measuring such low resistances has always been a tricky problem.

On 17/07/2019 15:05, Harry Bloomfield wrote:
After serious thinking Mathew Newton wrote :
Measuring the L-L and N-N resistance on a ring circuit is giving me
0.7 and 1.1 ohms respectively. Does this point towards a poor/loose
neutral connection in the back of a socket somewhere? If so, is there
a good way to locate where through testing as opposed to removing each
socket for direct inspection?

Visually check each socket's face, for signs of any over heating.

With the power off, you could make up a pair of (naughty) plugs with
bare wires coming out, then work your way around the ring, comparing L
to L and N to N resistances across sockets in the ring, until you
localise it, then delve into the socket for the cause.

Measuring such low resistances has always been a tricky problem.

First of all measure the resistance of each of your test
leads too.

On Wednesday, 17 July 2019 01:43:15 UTC+1, John Rumm wrote:
On 16/07/2019 23:05, Mathew Newton wrote:

Measuring the L-L and N-N resistance on a ring circuit is giving me
0.7 and 1.1 ohms respectively. Does this point towards a poor/loose
neutral connection in the back of a socket somewhere?

Quite possibly, or something else not quite right.

If so, is there
a good way to locate where through testing as opposed to removing
each socket for direct inspection?

You could go for a "binary chop" type approach. Disconnect the ends of
the ring at the CU, and short L & N & E of one end together. Now using a
plug as a test point, plug it in somewhere on the circuit near to where
you think the middle might be. [...]

Well John, it looks like it really was 'something else not quite right'! .....

I was all set to tackle this at the weekend and disconnected the ends of the ring. Measuring the L-L and N-N resistances again I found them both to be 1.1 ohms with the E-E at 1.5 ohms. Factoring in the 0.4 ohms test lead resistance gives me 0.7 ohms and 1.1 ohms respectively.

I see that copper cable is 7.41 mohms/m for 2.5mm2 and 12.1 mohms/m for the 1.5mm" CPC so with my measurements that suggests my circuit is ~90-95m long which sounds feasible for a 45m2 house footprint, a dozen or so sockets over two floors and 2.4m ceiling drops.

My conclusion is that I'd screwed up the original measurements somehow - perhaps measuring the L-L resistance of another ring or already taking off the test lead resistance; either way it seems like user error and that I don't have a problem afterall. I can't help but feel a bit disappointed as I was looking forward to finding and fixing the 'fault'!

Thanks anyway John for the suggestion and I'll bear it in mind should I ever have to do this again.

On 30/07/2019 22:59, Mathew Newton wrote:
On Wednesday, 17 July 2019 01:43:15 UTC+1, John Rumm wrote:
On 16/07/2019 23:05, Mathew Newton wrote:

Measuring the L-L and N-N resistance on a ring circuit is giving
me 0.7 and 1.1 ohms respectively. Does this point towards a
poor/loose neutral connection in the back of a socket somewhere?

Quite possibly, or something else not quite right.

If so, is there a good way to locate where through testing as
opposed to removing each socket for direct inspection?

You could go for a "binary chop" type approach. Disconnect the ends
of the ring at the CU, and short L & N & E of one end together. Now
using a plug as a test point, plug it in somewhere on the circuit
near to where you think the middle might be. [...]

Well John, it looks like it really was 'something else not quite
right'! ....

I was all set to tackle this at the weekend and disconnected the ends
of the ring. Measuring the L-L and N-N resistances again I found them
both to be 1.1 ohms with the E-E at 1.5 ohms. Factoring in the 0.4
ohms test lead resistance gives me 0.7 ohms and 1.1 ohms
respectively.

I see that copper cable is 7.41 mohms/m for 2.5mm2 and 12.1 mohms/m
for the 1.5mm" CPC so with my measurements that suggests my circuit
is ~90-95m long which sounds feasible for a 45m2 house footprint, a
dozen or so sockets over two floors and 2.4m ceiling drops.

My conclusion is that I'd screwed up the original measurements
somehow - perhaps measuring the L-L resistance of another ring or
already taking off the test lead resistance; either way it seems like
user error and that I don't have a problem afterall. I can't help but
feel a bit disappointed as I was looking forward to finding and
fixing the 'fault'!

Thanks anyway John for the suggestion and I'll bear it in mind should
I ever have to do this again.

Oh well, that seems like a reasonable solution then... (either that, or
you have found a really tricky intermittent poor connection! :-))


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/