Strange measured Zs Readings!?

[eveares]

So those who have seen my previous post know that I now own a Megger MFT 1730 what has recently been calibrated and that I was also experiencing higher than permitted Zs for my kitchen circuit what is protected by a 32A Type B MCB.

With my MFT1730, this is what I get from the same socket in the kitchen!:

3 Wire No-Trip setting = 1.08Ω

2 Wire No-Trip setting = 0.06Ω

Ze + (R1 + R2) = 0.74Ω (L-E link with R2 Socket Adapter)

Ze is 0.16Ω (Measured at CU using 2-Wire high current setting with main pipe bonding disconnected.)

Socket tested is about 25 Meters in cable length away from CU and is wired in 4.0mm² T&E. (Estimate). 2 and 3 Wire Np-Trip tests were done while the circuit was live.

So now for my questions:

1) Are No-Trip/low current loop readings normally that dramatically different from Ze + (R1 + R2)?

2) Why was the 2 Wire No-Trip reading so stupidly low?

3) Did doing the No-Trip tests on a live loaded circuit affect the readings?

Regards: Elliott

Edit: http://www.theiet.org/forums/forum/messageview.cfm?catid=205&threadid=51169 & Yes I do have the MK 7860s RCD's.

 

[EFLImpudence]

1) Shouldn't be - but I have read threads which conclude the Meggers aren't very good at no-trip tests.
Are you positive it wasn't 1.06 - just asking, just in case.

2) I don't know.

3) You have to do it live. It wouldn't be a circuit otherwise.

Can you bypass the RCD and do high current test?

 

[eveares]

Opps, I meant to write loaded circuit and not live circuit, I blame end of day tiredness for that mistake. Yes the two wire test did read 0.06Ω (2.66 x lower than Ze)

Ze + (R1 + R2) was definitely around 0.74Ω as Ze from the other day was 0.16Ω and R1+R2 from today using socket R2 breakout was 0.58Ω.

Another thread about MK Sentry RCD's massively screwing up the Zs readings: http://www.theiet.org/forums/forum/messageview.cfm?catid=205&threadid=53482

At least I now know that it was my RCD's likley screwing with all the No-Trip Zs readings that I have been doing.

 

[flameport]

1 - They should not be massively different. Measured Zs being lower is likely, certainly not the other way round.
2 - Interference from other equipment on the circuit, loads varied while the test was being done, solar panels are installed, low current tests are inevitably flawed, test machine is busted, all of the above.
3 - Very likely.

(L-E link with R2 Socket Adapter)

I hope you didn't use a socket adaptor with a link between L&E and then test the resistance at the consumer unit end?

 

[JohnW2]

Another thread about MK Sentry RCD's massively screwing up the Zs readings: http://www.theiet.org/forums/forum/messageview.cfm?catid=205&threadid=53482

I don't really understand what is being suggested in that thread ...

... the L path surely goes 'straight through' an RCD, other than for the switch contacts - so I can't see that anything other than very dodgy contacts could result in a difference. The difference in loop impedance measurements 'before' and 'after' the RCD in that OP was 0.76. If the switch contacts were introducing that amount of impedance, when the (63A) device were fully loaded, it would result in a VD of about 48V, and more than 3kW would be dissipated in the contacts (that would be fun!).

Am I missing something?

Kind Regards, John

 

[Iggifer]

What Ze reading do you get with bonding connected? The fact that Zs is lower than Ze (Technically Zdb) sounds like there's some parallel paths helping out a little.

And, I'm assuming you did the Zs reading L-E and not L-N?

 

[Iggifer]

I don't really understand what is being suggested in that thread ...

... the L path surely goes 'straight through' an RCD, other than for the switch contacts - so I can't see that anything other than very dodgy contacts could result in a difference. The difference in loop impedance measurements 'before' and 'after' the RCD in that OP was 0.76. If the switch contacts were introducing that amount of impedance, when the (63A) device were fully loaded, it would result in a VD of about 48V, and more than 3kW would be dissipated in the contacts (that would be fun!).

Am I missing something?

Kind Regards, John

All you're missing is that MK shouldn't even be allowed to make sockets anymore, never mind circuit protection

 

[EFLImpudence]

I make 25m of 4mm² 0.42Ω so if a bit longer then 0.74Ω could be the correct Zs.

 

[JohnW2]

All you're missing is that MK shouldn't even be allowed to make sockets anymore, never mind circuit protection

That may be true, but the story reported in that other thread still makes no sense to me - unless there were flames coming out of the RCD

Kind Regards, John

 

[Iggifer]

That may be true, but the story reported in that other thread still makes no sense to me - unless there were flames coming out of the RCD

Kind Regards, John

I know RCDs can be strange beasties, but it makes no sense to me either.

Nor did the old rule 'always test RCDs with the loads disconnected' until one failed on 1x and 5x, disconnected a refrigeration load and it was well within spec (was a brand new Schneider RCBO so it should bloody well pass!)

 

[JohnW2]

I know RCDs can be strange beasties, but it makes no sense to me either.

They may be strange beasties, but as far as the measurement of loop impedance is concerned, they are surely no more than a DP switch (with contacts closed)?

Nor did the old rule 'always test RCDs with the loads disconnected' until one failed on 1x and 5x, disconnected a refrigeration load and it was well within spec (was a brand new Schneider RCBO so it should bloody well pass!)

Although I struggle to think of a mechanism, I must say that I find that far less difficult to 'believe'!

Kind Regards, John

 

[flameport]

the L path surely goes 'straight through' an RCD, other than for the switch contacts - so I can't see that anything other than very dodgy contacts could result in a difference.

It's the electronics inside that cause problems, just as electronics inside certain appliances can.

Switch mode type power supplies designed to operate over wide input voltage ranges in particular.

 

[JohnW2]

It's the electronics inside that cause problems, just as electronics inside certain appliances can. Switch mode type power supplies designed to operate over wide input voltage ranges in particular.

Do you mean that they result in noise in the circuit, which somehow confuses the measuring device?

Kind Regards, John

 

[flameport]

Either noise, or the characteristics of the device itself.

A power supply which works over a wide range of voltages will by design change it's apparent input impedance as the voltage changes.
Add to that a testing device on the same circuit which either draws current from the supply or injects a current and expects to see a voltage change, and all kinds of strange behaviour could result.

 

[JohnW2]

A power supply which works over a wide range of voltages will by design change it's apparent input impedance as the voltage changes. Add to that a testing device on the same circuit which either draws current from the supply or injects a current and expects to see a voltage change, and all kinds of strange behaviour could result.

Interesting. I must say that I would have thought that the 'apparent input impedance' of the 'power supply' (which I suspect may often be very crude) within an RCD would, even at its lowest, be so high in comparison with the loop impedance being measured that it would not have any appreciable effect on the measurement - but it sounds as if I may be wrong in thinking that.

Kind Regards, John