Volt drop problem

[SUNRAY]

I have been out to look at a problem in a commercial property with time limited access due to the nature of the site.
Very simple circuit: single 30A Wylex fused switch containing yellow 20A fuse feeding:
A DSSO in adjacent building with 4mm² T&E guess 15m.
A DSSO in adjacent room with 2.5mm² T&E guess 4m.


While decorating the room the steam wallpaper stripper didn't work but a LED table light does (no surprise).
Extension lead from same circuit from the other building runs the steamer.

My initial suspicion being the socket or plug, tried with their kettle (most common load in other building) and confirm report so fitted a different socket - no change.

Access to fusebox is difficult due to security and a long treck round the building and grounds (300-400m) but can get fusebox switched on/off with ease via phone call.

Switched off I can see open circuit between L & N & E (3 tests) at both sockets, shorting L & N at either socket shows a short circuit at the other. but this is only using a multimeter.

A colleague has been in tonight and confirms symptoms, even swapped sockets over. Couldn't get access to fusebox and repeated tests between sockets using a KEW MFT: L-N loop and L or N - E loop. He didn't send the results other than 'spot on' (second image)

I think this is all it is:

Now waiting for a time for two of us to go in.
Any ideas?

 

[JohnW2]

Access to fusebox is difficult due to security .... but can get fusebox switched on/off with ease via phone call.

I'll read the rest of your post shortly, but .... I hope that Mr 'Elf and Safety didn't read the above

Kind Regards, John

 

[JohnW2]

A colleague has been in tonight and confirms symptoms, even swapped sockets over. Couldn't get access to fusebox and

How did he 'swap the sockets' without being to 'get access to fusebox' - is this something else that Mr 'Elf & Safety should not be told about?

repeated tests between sockets using a KEW MFT: L-N loop and L or N - E loop. He didn't send the results other than 'spot on'

If the L-N loop impedance measured at the socket is 'spot on' (which I assume means 'very low'), then it surely makes no sense that the socket cannot satisfactorily power a sizeable load, does it?

Has someone measured the voltage at this socket when a largish load is connected?

(second image)

I don't really understand that image - it appears to be showing L & N (of a live circuit?) connected together!

Kind Regards, John

 

[SUNRAY]

How did he 'swap the sockets' without being to 'get access to fusebox' - is this something else that Mr 'Elf & Safety should not be told about?

The fuse box is in an area easily accessible by their security staff and a call to them works well... yes one does have to trust them but they have been fine so far.

If the L-N loop impedance measured at the socket is 'spot on' (which I assume means 'very low'), then it surely makes no sense that the socket cannot satisfactorily power a sizeable load, does it?

Hence the reason for my post.
I only had a multimeter with me and measured loop resistance socket to socket, as diagram 2. my colleague has done the same test with a KEW, he may very well have done a loop impedance test.

Has someone measured the voltage at this socket when a largish load is connected?

Yes with the kettle plugged in my multimeter shows it dropping to virtually zero.

I don't really understand that image - it appears to be showing L & N (of a live circuit?) connected together!

with the fusebox switched off we have both measured the loop resistance socket to socket by placing a short at one position.
My multimeter showed something like 0-½Ω loop resistance

 

[jj4091]

Surely it can only be a bad connection at the switch or a cable fault. Testing with meters won't apply any load, will they?

 

[JohnW2]

The fuse box is in an area easily accessible by their security staff and a call to them works well... yes one does have to trust them but they have been fine so far.

In view of your earlier comment, that what I assumed - but, even if "they have been fine so far", you probably should not let Mr H&S know about that

Hence the reason for my post. I only had a multimeter with me and measured loop resistance socket to socket, as diagram 2. my colleague has done the same test with a KEW, he may very well have done a loop impedance test. Yes with the kettle plugged in my multimeter shows it dropping to virtually zero.

I still don't really understand what is going on in your second diagram.

However, what you appear to be describing sounds crazy (which I imagine is what you also think). IF a ('live') measurement of L-N loop impedance shows the expected 'low' reading at the affected socket, then it makes no sense that the voltage at that socket drops to near zero when a load was connected - Mr Ohm could not understand that!

The only 'logical' conclusion one could draw from that is presumably that when you plug in the load that somehow introduces a high resistance (upstream of wherever you are measuring the near-zero voltage) that doesn't happen when you plug in your loop impedance meter - and I can't thing how that could happen!

Kind Regards, John

 

[bernardgreen]

Is there an auto resetting circuit breaker that trips when the load is applied and then resets before the next test is performed

 

[SUNRAY]

Surely it can only be a bad connection at the switch or a cable fault. Testing with meters won't apply any load, will they?

My multimeter deffinitely not more than a μA or so but hopefully my colleagues KEW is running at its 200mA.

 

[JohnW2]

Surely it can only be a bad connection at the switch or a cable fault. Testing with meters won't apply any load, will they?

My multimeter deffinitely not more than a μA or so but hopefully my colleagues KEW is running at its 200mA.

All this talk about cable/connection impedances is all very well but, as I've said, if a live L-N loop impedance test from the socket gives the expected answer (i.e. less than a couple of Ohms or so), then there is no rational explanation for how connecting a load to the socket can result in voltage at the socket dropping to near zero, is there? (other than bernard's suggestion that some device temporarily 'moves the goalposts' whenever a load is connected, but you'd surely know if any such device were involved, wouldn't you?)

Kind Regards, John

 

[SUNRAY]

Is there an auto resetting circuit breaker that trips when the load is applied and then resets before the next test is performed

Sadly not, just as the sketch straight off a Henley block and Wylex switched fuse... or is it a fused switch. Either way one socket works and the other doesn't.

In view of your earlier comment, that what I assumed - but, even if "they have been fine so far", you probably should not let Mr H&S know about that

I still don't really understand what is going on in your second diagram.

With power switched off effectively all there is two sockets connected together so I've done R1+R2 test end to end.

However, what you appear to be describing sounds crazy (which I imagine is what you also think). IF a ('live') measurement of L-N loop impedance shows the expected 'low' reading at the affected socket, then it makes no sense that the voltage at that socket drops to near zero when a load was connected - Mr Ohm could not understand that!

I haven't done a loop impedance test, only R1+R2 (between the 2 sockets with the power off) with a multimeter. My colleague may have done a loop impedance test but all he's actually stated is:
"I see what you mean it's a strange one, there was some sort of issue l just after I got in there and could not get passed the security door only round the back, it sounds like I done all the same tests as you done and there spot on as I had the sockets off for the tests I swapped them over but its still the same problem with there kettle. I used an extention lead and found it neutrel whats faulty. They went to full lockdown before I could do anything else. Tels going to try and get us both in together, what are you doing next week?"

The only 'logical' conclusion one could draw from that is presumably that when you plug in the load that somehow introduces a high resistance (upstream of wherever you are measuring the near-zero voltage) that doesn't happen when you plug in your loop impedance meter - and I can't thing how that could happen!

Kind Regards, John

I'm with you and all I can think is something is breaking down under heavy current but not under tester current.

 

[SUNRAY]

Reply from colleague:
"didnt do loop impedance as PFC was only 2A in the hut its 900A"

 

[JohnW2]

With power switched off effectively all there is two sockets connected together so I've done R1+R2 test end to end. .... I haven't done a loop impedance test, only R1+R2 (between the 2 sockets with the power off) with a multimeter.

Oh, I see - I obviously misunderstood what you meant by "L-N loop"

My colleague may have done a loop impedance test ....

That (which I thought was what had been done) is surely what was needed and, in a second message, you go on to say ...

Reply from colleague: "didnt do loop impedance as PFC was only 2A in the hut its 900A"

Which makes one 'wonder' about your colleague He says he "didn't do loop impedance" yet he quotes the PFCs (which indicate that he has done loop impedance measurement) !! I obviously don't know what the voltage actually was, but if one assumes 230V, then a PFC of 2A translates to a loop impedance of 115Ω and a PFC of 900A translates to a loop impedance of about 0.26Ω.

So there is no 'crazy anomaly' as I had thought and, in fact, you/he have essentially localised the problem. Since both sockets are fed from the same switch fuse, if the loop impedance is 115Ω at one socket and 0.26Ω at the other, then this obviously means that there is something very high resistance going on in either the cable supplying the first socket or the connections to one or both ends of it.

The next step is surely to measure the end-to-end resistance (separately for L and N) of that cable (your multimeter should be more than adequate). Given that, as I understand correctly, that cable is only 4m in length, that shouldn't be too difficult.

Kind Regards, John

 

[SUNRAY]

Oh, I see - I obviously misunderstood what you meant by "L-N loop"

That (which I thought was what had been done) is surely what was needed and, in a second message, you go on to say ...

Which makes one 'wonder' about your colleague He says he "didn't do loop impedance" yet he quotes the PFCs (which indicate that he has done loop impedance measurement) !! I obviously don't know what the voltage actually was, but if one assumes 230V, then a PFC of 2A translates to a loop impedance of 115Ω and a PFC of 900A translates to a loop impedance of about 0.26Ω.

So there is no 'crazy anomaly' as I had thought and, in fact, you/he have essentially localised the problem. Since both sockets are fed from the same switch fuse, if the loop impedance is 115Ω at one socket and 0.26Ω at the other, then this obviously means that there is something very high resistance going on in either the cable supplying the first socket or the connections to one or both ends of it.

The next step is surely to measure the end-to-end resistance (separately for L and N) of that cable (your multimeter should be more than adequate). Given that, as I understand correctly, that cable is only 4m in length, that shouldn't be too difficult.

Kind Regards, John

By default yes the tester has done a LI test however with my Robin set I can do either test with direct reading, I assume the same with the KEW, on that basis I see no reason to 'wonder' about him.


Again I concur it does appear from the latest info there is a high resistance in the shorter cable, except that my multimeter didn't find it doing this test, redrawn to simplify and show what I believe is in place with power off:

My meter shows about 0.4Ω with just probes shorted, Doing that test I measured <1Ω but didn't note the exact reading due to the inaccuracy. To get an idea of performance I've just measured 2x 2.2Ω resistors in parallel as 1.6Ω, my Robin shows that as 1.12Ω so I don't think it's embarrassing itself.
The LED table lamp shows no dip in mains voltage (hardly a surprise at a watt or three) but the kettle pulled it down virtually zero, assuming 3KW = ~18Ω then I'd have expected ~1/8th of 230V or ~25V and not the near zero I saw. From the kettle test I reckon the H.R. is more like 3.6KΩ.

There is something more than a simple high resistance. But of course it's going to be something stupidly simple when found.

I have nothing further to add until I get back there or hear from others.

 

[JohnW2]

By default yes the tester has done a LI test however with my Robin set I can do either test with direct reading, I assume the same with the KEW, on that basis I see no reason to 'wonder' about him.

I 'wondered', because he said that he didn't do loop impedance measurements, whereas he obviously did! He quoted PFCs at both sockets, and PFC is nothing more than voltage divided by loop impedance. One therefore cannot possibly have a PFC without having measured loop impedance (and voltage).

Again I concur it does appear from the latest info there is a high resistance in the shorter cable, except that my multimeter didn't find it doing this test, redrawn to simplify and show what I believe is in place with power off:

If (per your colleague's PFCs) the loop impedances is about 115Ω at the affected socket (but 0.26Ω at the OK one),there is clearly a high resistance (nearly 115Ω) somewhere between the downstream end of the cable and the fuse. That resistance does not necessarily have to be in the cable itself - it could be in its connection to the switch-fuse.

Can you not get at this switch fuse and disconnect the cable from it? If you did that, you could simply measure its end-to-end resistances (with your multimeter - which could easily distinguish between ~115Ω and 'near zero'). If that showed the cable to be OK, that would surely only leave one possible culprit - namely its connection to the switch fuse.

There is something more than a simple high resistance. But of course it's going to be something stupidly simple when found.

Yes, I'm sure it will end up as being something "stupidly simple" but, in conceptual terms, I can't see that it can be anything other than a high resistance between the fuse and the downstream end of the cable.

Good luck!

Kind Regards, John

 

[SUNRAY]

I 'wondered', because he said that he didn't do loop impedance measurements, whereas he obviously did! He quoted PFCs at both sockets, and PFC is nothing more than voltage divided by loop impedance. One therefore cannot possibly have a PFC without having measured loop impedance (and voltage).

I've already agreed with this.

If (per your colleague's PFCs) the loop impedances is about 115Ω at the affected socket (but 0.26Ω at the OK one),there is clearly a high resistance (nearly 115Ω) somewhere between the downstream end of the cable and the fuse. That resistance does not necessarily have to be in the cable itself - it could be in its connection to the switch-fuse.

I agree with all of that...
Except that unless I have got it wrong, ignoring the Wylex which is effectively a junction box (shown as orange circles) with the power turned off with all of the 2.5mm² & 4mm² simply connecting the 2 sockets together

and testing like this

with a short at one socket and meter at the other must, by definition, find that high resistance? That's the test I've done, in both directions.

Can you not get at this switch fuse and disconnect the cable from it?

I have been there once, however the property has an exceptional security requirement and suffers regular lockdowns/evacuations of visitors, as colleague found out.

If you did that, you could simply measure its end-to-end resistances (with your multimeter - which could easily distinguish between ~115Ω and 'near zero'). If that showed the cable to be OK, that would surely only leave one possible culprit - namely its connection to the switch fuse.

Yes, I'm sure it will end up as being something "stupidly simple" but, in conceptual terms, I can't see that it can be anything other than a high resistance between the fuse and the downstream end of the cable.

Good luck!

Kind Regards, John