Zs wanders around

[JohnW2]

It affected by the magnetic core - higher current increases the magnetic field in the core and as it approaches magnetic saturation the impedance decreases.
An inductor with a magnetically saturated core isn't an inductor any more.

Thanks. That much I understood, but I was surprised by the implication that the core could become saturated with a current something between 15mA and 100 mA

In any event, the figures they presented would seem to indicate that, in the absence of any additional residual current due to connected equipment, the meter they were talking about would always give a reading which was about 0.5 Ω too high. Is that what they are saying/implying?

If nothing else is connected to the installation then the additional impedance should be consistent, however in real installations there is likely to be some residual current from connected equipment, and as that varies the additional impedance seen by the test equipment will also vary.

Indeed, but if I understand correctly, the OP undertook his testing, with varying Zs results, when there were "no loads on the circuit".

 

[plugwash]

Poor contacts can definitely exhibit a "resistance" that depends on the test current. That is one of the reasons we do tests at relatively high currents in the first place but for a Zs test behind a RCD you are obviously limited by the RCDs trip current.

I'd guess the OP's tester always does a high current test for the P-N loop.

 

[JohnW2]

Poor contacts can definitely exhibit a "resistance" that depends on the test current.

Indeed - and not only "poor contacts" but also "poor contact". That's why I've always been amazed at how credible to 'very low resistance' measurements that electrocutions (or their MFTs etc.) manage to do. - if I use a meter, be it an MFT, multimeter or whatever) to measure a very small resistance, I can usually get variation of at least an Ohm or so just by moving the probes around, or altering the pressure I apply to the probe, yet people seem fairly confident that their measurements are correct to within a pretty small fraction of an Ohm.

That is one of the reasons we do tests at relatively high currents in the first place but for a Zs test behind a RCD you are obviously limited by the RCDs trip current.

Agreed.

I'd guess the OP's tester always does a high current test for the P-N loop.

Yes, and whatever current it uses, one presumes that it always uses essentially the same current when measuring resistance./impedance in a certain ballpark,

 

[ebee]

Some good answers here.
My immediate thought when reading the thread was the "high" Ze reading. In the area where I live there is a lot of old housing stock and the incomers were always traditionally TNS, it was/is very normal to see the Ze on those homes to be around 0.23 to 0.24, other areas may vary of course.
In my areas any reading higher than say 0.25 would start to make one ponder and do a repeat test. So it did cross my mind if there could be a problem with the earth connection at the incomer, it does happen around here, found it a few times meeself and the DNO have ended up digging front/back gardens and sometimes in the street.road to make a repair. Fractures can vary in pressure between a good or poor reading and sometimes luck helps you to see that.
I remember on a I & T course the tutor mentioned the test rigs - mock up of rings and radial circuits for the pupils to make tests with there meters and they found some variances, on plugging/unplugging a few times the results benefited overall, oxides and dirt cleaning off a bit with every insertion.
The point Flameport made if testing an RCD protected circuit certainly can make a difference, the old circuits with rewireable fuseways did not suffer that problem.
Then loose connections can happen even in what appears to be the best wiring.
All can play a part, so it is wise to get to the bottom of it.

 

[martygturner]

Again many thanks to all, have found the cause and I admit its a simple fault.

This morning we took out the RCB, whilst we had checked the tightness of the connection we had not removed any of its connections..

Inside the live and to some degree the neutral connector was some blobs of fire seal which had coated the cable. Gave it a clean, now rock steady Zs.

Sorry to have prompted a debate.

 

[pete01]

Told ya

 

[plugwash]

As I've been discussing elsewhere, I think the test button is a relative crude test, and will only 'fail' if the device is way out of spec (or doesn't work at all). A MFT ought to give pretty good figures - but, of course, if there is an 'inconsistent/varying problem' (e.g. due to iffy contacts), then any measuring device will give results which reflect that (actual) inconsistency.

A RCD test with the test button tells you that the RCD trips at whatever current the manufacturer chose for the test button.

RCD tests with a MFT can tell you either how fast (if at all) the RCD trips at a given test current (regular test) or how much current is required to trip the RCD (ramp test).

Neither tells you anything about how good the contacts in the RCD are (unless they are so bad that the test cannot function)

 

[JohnW2]

A RCD test with the test button tells you that the RCD trips at whatever current the manufacturer chose for the test button.

Indeed, and it seems that, for whatever reason, they generally choose 2 - 2,5 time IΔn.

RCD tests with a MFT can tell you either how fast (if at all) the RCD trips at a given test current (regular test) or how much current is required to trip the RCD (ramp test).

All true.

Neither tells you anything about how good the contacts in the RCD are (unless they are so bad that the test cannot function)

If you mean the RCD's 'switch' contacts, then I agree. However, I wasn't totally sure what 'contact(s)' you were talking about when you wrote ...

Poor contacts can definitely exhibit a "resistance" that depends on the test current. ......

The previous discussion had suggested that the varying Zs figures the OP was getting might be due to a 'poor contact' somewhere in the fault loop path - connections to the sockets having been suggested as one possibility. If something like that were going on (it wasn't me who suggested it ), then the MFT would see that as a varying Zs, wouldn't it?

However, in relation to all the above, how on earth did these comments about testing of an RCD with an MFT get into a discussion about varying Zs measurement results?

 

[ebee]

A RCD test with the test button tells you that the RCD trips at whatever current the manufacturer chose for the test button.

Indeed, and it seems that, for whatever reason, they generally choose 2 - 2,5 time IΔn.

Agreed - in my experience both measured and data that tallies - makes sense it is a rough test if it works at all (and wipes the mechanism so reducing "stiction") - so it is a rough sort of go/no go test, it works at some figure or it does not work at some figure is probably the best thing to have for a test button if you take all scenarios into account I suppose.

A decent MFT will, hopefully, give a reasonable Indication of the "parish" it works within at the time of the test - I suspect that the test current will always be a percentage above the stated threshold for the test to be meaningful . So again I am OK with that in principal , gut feeling a 5 to 10& above might be reasonable but I have no idea what they aim for.

Switch contacts? Residual Capacitance/Inductance certainly appear to play a part in altering the MFT test if the outgoing circuit is left connected during the test procedure -I remember an Inspector doing the test on two connected circuits in a lunch break of other trades and the results of one of those circuits were unexpected but still within so I suggested a quick swapping over of RCD to circuits confirmed the odd result followed the circuit not the RCD and the circuits were back in service just as lunch break ended.
By coincidence that very evening I had interesting discussions in night school class and the class was divided into those who did the test with no circuit attached at all and those who did with circuit connected but not appliances and those who did with both circuit and appliances connected.
views varied but yes switch contacts I would be inclined to accept could play a part wherever they are.