Voltage on CPC

[Midwest]

Went to a job the other day, where the client wanted a light fitting put up instead of the ceiling rose. In conversation they mentioned having problems with their electrics, and had had a new CU fitted. I decide to measure ZS & RCD at the said ceiling rose. However, my tester (Megger1730) aborted both tests due to 'high touch' voltage. Measuring voltage at the rose, 242v with light switched on & 184v with the light switched off. Light is controlled by 2 x two way switches, with a 3c & e at each, and tw & e at the rose itself, so obviously JB somewhere.

One of the switches is a metal plate switch, measuring voltage there (to neutral bar of adjacent CU); 242v on red strapper, 54v & 52v on blue & yellow strappers, and worringly 54v on cpc. Client said they thought they had received a shock from switch, but had dismissed it. Adjacent ceiling rose (on same circuit) & socket outlet, gave suitable ZS/RCD test results. I didn't do any other testing, as I ran out of time. So made light and switches safe and made an appointment to go back and further investigate.

Never come acorss this myself before, but have read of induced voltages. No flourescents or motors etc at the property (TN-S supply). I suspect an incorrectly wired JB, but I can't fathom these voltage readings?

 

[Plonker 3]

So if you didnt carry out any further testing how do you know you made it safe?

 

[Midwest]

I meant that I removed the ceiling rose and metal switch, placed cables in connectors etc. I left it in a safer state than I found it.

 

[SirKit Breaker]

You need to test this in the correct sequence, this should highlight any problems. So check your R2 or R1+R2, then do IR test, and see what you get. I would disconect the circuit fully at the CU before you test it, and then you should be able to confirm that it is this circuit that has the problem, or that the problem is another circuit, as you have common connections on CPC's and Neutrals in the CU the fault could be induced from elsewhere.

Cheers...............Howard

 

[Midwest]

Yep thats the best approach, just goes to show what was a simple job turns into something slightly more difficult. Just can't take anything for granted.

 

[SirKit Breaker]

If you don't test in a logical way, then you just end up chasing your harris around, and end up back where you started from and none the wiser for the experience.

Cheers...........Howard

 

[Midwest]

Yep point taken, but can anyone explain the voltages I'm getting?

 

[telectrix]

not without being there.

first test i would do is get a wander lead from the suspect cpc/s and measure ohms back to the MET.

 

[Midwest]

Are you saying I should carry out tests in order, continuity of protective conductors, IR etc?

 

[Engineer54]

Eh, ....that would be the logical approach wouldn't it. It'll be giving you a lot more information about what's happening than you have now, by just using a voltage tester!!

 

[NickD]

Yep point taken, but can anyone explain the voltages I'm getting?

Because it's a digital meter and places negligible load on the conductors under test, they could be induced voltages in conductors connected to bog all.

Anyone know if you can buy something that plugs into the back of your test probe leads and adds a small load, 100kohm or something? Obviously you could make one.

 

[Geoffsd]

Isn't it the other way round?

 

[Richard Burns]

A digital meter provides a near infinite resistance to the points being measured and therefore there is practically zero current flow, so any voltage present that has been induced from another circuit will not clear and so almost any voltage reading up to the supply level can be measured, however this measured voltage has little capability to do anything and if it were measured by an analogue meter that has a high but measurable resistance, the small current flow that occurs when measuring will drop the voltage to the actual voltage available to do work.
In simple terms.

 

[Engineer54]

A digital meter provides a near infinite resistance to the points being measured and therefore there is practically zero current flow, so any voltage present that has been induced from another circuit will not clear and so almost any voltage reading up to the supply level can be measured, however this measured voltage has little capability to do anything and if it were measured by an analogue meter that has a high but measurable resistance, the small current flow that occurs when measuring will drop the voltage to the actual voltage available to do work.
In simple terms.

In other words, always carry a cheap analogue voltmeter in your test kit, to check any suspected phantom voltage!!

 

[Marvo]

......Anyone know if you can buy something that plugs into the back of your test probe leads and adds a small load, 100kohm or something? Obviously you could make one.

I saw that Fluke do a 'low impedance' kit that piggy backs on one or more of their testers so I'd guess the other big manufacturers would have something similar.

Some newer testers have a low impedance option built-in;

The Fluke 117 has dual impedance capability: incorporating both regular high impedance test capabilities and low impedance functions for detecting ghost voltages, just switch to the Auto-V/LoZ (low impedance) test setting to whether measurement is a ghost voltage.

 

[Engineer54]

Cheaper all round, to just buy a £10-15 analogue meter!! lol!!

 

[Midwest]

In other words, always carry a cheap analogue voltmeter in your test kit, to check any suspected phantom voltage!!

Above my pay grade. But going back to my original post, this came to my attention because my Megger 1730 aborted ZS/RCD test because it sensed a 'high touch voltage' at that lighting point, but not at an adjacent ceiling rose (on the same circuit). So is this touch voltage, a phantom voltage, I can't see that?

 

[mobile]

Voltage could be caused by loose neutral but you need to carry out tests as previously mentioned

 

[Hawkmoon]

not without being there.

first test i would do is get a wander lead from the suspect cpc/s and measure ohms back to the MET.

Totally agree - This is almost certainly non-continuous cpc subject to induced voltages