Insulation test of a long cable

[Guineafowl]

In order to do some welding a long way from a power socket, I have made up a 200m long ‘extension lead’ from some 6mm2 armour cable. The cable is needed for another job later on.

I have an test unit that:
1. Checks earth/ground bond with 8Vac and 25A.
2. Then measures insulation resistance between L+N and Earth at 1250Vac.

The Earth test passed, but the IR test failed (trip at 4mA).

However, when I repeated the IR test using a 1000Vdc (another tester - I like testers) the test passed (IR >10Gohm).

Why? I see two possibilities:

1. 1250Vac has a much higher peak voltage than the DC 1000V.
2. The ac test was affected by wire-wire capacitance.

 

[.Scott]

It definitely sounds like option 2 (capacitance) to me.

 

[Rive]

1250Vac - I like testers

Just curious - does that tester works at 50/60Hz?

 

[Guineafowl]

Just curious - does that tester works at 50/60Hz?

I guess it steps up line voltage (240V) to 1250V at line frequency (50Hz). It’s a big buzzy transformer, anyway.

 

[anorlunda]

When distances get long, it is not just capacitance between conductors, but capacitance to ground that is significant. I expect that your results might be different if the cable was lifted high off the ground.

 

[Guineafowl]

When distances get long, it is not just capacitance between conductors, but capacitance to ground that is significant. I expect that your results might be different if the cable was lifted high off the ground.

That’s another difference between the two testers - the AC one is mains-powered, although I’m not sure if the Earth reference is carried through to the output. The DC one is battery-powered, so floating.

 

[jim mcnamara]

By high do you mean as in an electric service drop? ~8m. A priori that sounds expensive for 200m. Probably 3-4 power poles. The cost would be competitive with a Lincoln welder/engine that generates it own power.

https://www.lincolnelectric.com/en-us/equipment/engine-drives/Pages/engine-drives.aspx
US site, look at the Bulldog model down lower. Circa 2000US$.

 

[anorlunda]

Another all-important difference. With DC, if you put a capacitor to ground it charges up, and then no more current flows. With AC, a capacitance to ground is like a high impedance short-circuit.

 

[Guineafowl]

By high do you mean as in an electric service drop? ~8m. A priori that sounds expensive for 200m. Probably 3-4 power poles. The cost would be competitive with a Lincoln welder/engine that generates it own power.

https://www.lincolnelectric.com/en-us/equipment/engine-drives/Pages/engine-drives.aspx
US site, look at the Bulldog model down lower. Circa 2000US$.

This length of cable will be buried eventually to provide power to outhouses. But before that, it makes a convenient, beefy extension lead.

 

[anorlunda]

This length of cable will be buried eventually to provide power to outhouses. But before that, it makes a convenient, beefy extension lead.

That makes the capacitance effect worse. When it rains and the ground gets saturated, it will be even more severe.

When we bury high voltage AC power transmission cables, we put them in conduits. The conduit keeps it dry, but by putting the conductor on the center line, it also provides some distance between the conductor and earth. Even with all that, underground and underwater AC power cables are very limited in the distances because of the capacitance problem.

Also, restive heating of the conductors is a bigger problem with directly buried cables because the ground conducts heat away less effectively than air. That's a second reason for a conduit. The air space in the conduit and the bigger surface area of the conduit, help keep the cable cool.

I didn't think of it before, but your 200m extension cord might also violate the electric codes of your locality and thus be illegal. That could result in your insurance being rendered invalid and make your house impossible to sell in the future. You should check the local codes before going further on this project. In all cases that I know, compliance with electrical codes is mandatory.

Because of the code problem, I am going to close this thread for now. @Guineafowl , click on my name and send me a PM if you find that it is OK with your local code, and I'll re-open the thread.

 

[anorlunda]

Based on PMs from @Guineafowl and @.Scott , I am convinced that we can reopen this thread.

Don't be afraid to challenge mentor decisions in PMs, or in reports. We are as fallible as anyone else.

Thread open.

 

[jim hardy]

I have an test unit that:
1. Checks earth/ground bond with 8Vac and 25A.
2. Then measures insulation resistance between L+N and Earth at 1250Vac.

The Earth test passed, but the IR test failed (trip at 4mA).

I doubt it's a low capacitance type of cable
How much capacitance would it take to get a "FAIL" ?

1250 volts / 4ma = 312.5 kohms
at 50 hz that's 10,186 picofarads
which would be just 51 picofarads per meter
or 15.5 picofarads per foot
which is in the realm of low capacitance audio cable.

I'd say your tester presumes it's testing short extension cords..

old jim

 

[Guineafowl]

Thanks, @anorlunda - I see the confusion between electrical installation work, and pluggable ‘appliances’. The former is more tightly regulated, at least in the UK.

My contention was, for those interested, that the extension cord would be classed as an appliance, hence not subject to installation codes. The only regs applicable would be for portable appliance testing (PAT), and then only for a business premises. This involves continuity, polarity, Earth bond and insulation resistance checks, and a general physical inspection (which, in fact, picks up almost all faults - damaged cable sheaths and the like).

It was, in fact, in doing a PAT on the lead that I found the subject of this thread. I always make a point of testing mains electrical things that I make or repair along the lines of the proper PAT procedure.

 

[Guineafowl]

I doubt it's a low capacitance type of cable
How much capacitance would it take to get a "FAIL" ?

1250 volts / 4ma = 312.5 kohms
at 50 hz that's 10,186 picofarads
which would be just 51 picofarads per meter
or 15.5 picofarads per foot
which is in the realm of low capacitance audio cable.

I'd say your tester presumes it's testing short extension cords..

old jim

Yes, the tester (Clare v154) is really for appliances and normal extension leads. The DC tester, a Megger BM10, has a range of DC outputs from 50V to 1000V and is the tool for testing long cable runs. Also very handy, at lower voltages, for testing transformer and motor windings, and for similar checks on car alternators at the 50V setting. Other uses - capacitor leakage, diode reverse breakdown, MOV breakdown...

 

[anorlunda]

We have not mentioned voltage yet. You can look up the expected DC voltage drop from the wire gauge tables. You can also look up the minimum voltage needs of the welder. The same table will give you ohms per meter, so you can calculate the I²R losses. Temperature depends on thermal conductivity which is hard to estimate. That is one place where codes are useful because they replace difficult calculations with rules-of-thumb.

For an AC line, series resistance and series reactance (inductance) cause voltage to drop from source to load. But shunt capacitance (i.e.. capacitance to ground) tends to increase voltage on the load end; in extreme cases higher than the source voltage, and highest at zero load. The net, whether load voltage is higher or lower under load, depends on the balance between the numbers. If you have a boom box plugged in at the welder end, you might fry it with high voltage spikes as the welding arc stops.

If the cable lays on the ground, at times it will become an underwater cable. A conduit would mitigate that.

If the welder rectifies the AC, then you're going to have high harmonic content in the current, that amplifies the inductance and capacitance problems. An inexpensive half-wave rectifier would be the worst case.

 

[Guineafowl]

For an AC line, series resistance and series reactance (inductance) cause voltage to drop from source to load. But shunt capacitance (i.e.. capacitance to ground) tends to increase voltage on the load end; in extreme cases higher than the source voltage, and highest at zero load. The net, whether load voltage is higher or lower under load, depends on the balance between the numbers. If you have a boom box plugged in at the welder end, you might fry it with high voltage spikes as the welding arc stops.

I can do real-world tests at the moment, because the cable is looped round back to the workshop so that both ends are accessible. I have a kill-a-watt equivalent that will tell me the voltage drop at a certain current.

Series resistance is about 0.5 ohm, which I think is pretty good for such a long length.

This shunt capacitance increasing voltage is very interesting - is it acting with the series inductance to produce a boost converter?

The welder is one of these new-fangled inverters that produces a DC output. It’s reasonable quality - Parweld XTS 160 - so I hope the output is full-wave rectified. Power factor (quoted as cos phi, but I think the main factor is distortion, not displacement) is 0.7.

We’re building a tree-house for the kiddies using steels between two trees as the main bearers. Both of us are too old for such things as boom boxes...

 

[anorlunda]

This shunt capacitance increasing voltage is very interesting - is it acting with the series inductance to produce a boost converter?

LOL that is an interesting parallel. That never occurred to me, but perhaps yes. But remember it happens only if the C/L ratio is high enough. I can't estimate the C or the L for your cable.

I would guess that a welder is tough and not easily fried by spikes. But I would avoid plugging anything else in at the load end. I would also call off welding during and just after rain.

 

[Babadag]

If the cable is manufactured according to IEC 60502-1 no requirement for IR measurement
as routine test or after installation. Only as type test on samples IR measurement is required.
For PVC IR=36.7 MΩ.km at 20oC but only 0.037 MΩ.km at 70oC.
Less IR values are expected on site after installation.
The cable cross section it is as shown here.

conductor diameter 3.3 mm
insulation thickness 0.8 mm[PVC].
armor diameter 15.4 mm
overall diameter 18.6 mm
The capacitance of one conductor with respect the other and armor 275 pF/m for PVC with εr=4.5.
The calculated current when applied 1250 V 50 Hz approx. 40 mA [200 m length]

 

[Guineafowl]

The calculated current when applied 1250 V 50 Hz approx. 40 mA [200 m length]

I think that’s our answer - trip at 4mA.

I was really testing to ensure adequate Earth bonding to the armour. The IR was part of the test sequence of the machine, but it was a surprise when it failed.

 

[essenmein]

An aside, if the tester gives you the current and you can change the test voltage, then reducing the voltage would give a proportional reduction in current if its due to capacitive impedance, but will not be a linear relation if its breaking over somewhere (ie actual insulation failure).

 

[Babadag]

In my opinion, if you used Clair v154 tester at 1250 V 50 Hz then it was not IR test executed but:
AC HiPot Test Programmable Voltage Range 0.10 kV-5.00 kV (10V/Step) Frequency Independent of Supply 50 or 60 Hz Voltage Display and Accuracy 0.10 kV-5.00kV ±1% ± 5 Counts Voltage Display Resolution 0.01kV Current Display and Accuracy 0.01mA-20.00mA ±1% ± 5 Counts Current Display Resolution 0.01mA Selectable Range of Pass/Fail Levels* 0.01mA-20.00mA Maximum Current Output 20.00mA @5kV Optional Arc Detection 9 Levels
according to Instruction Manual:
http://www.seaward.co.uk/downloads/clare_hal_combined_user_manual_455a553_154_538_rev12_v1.pdf
page no.90

20 mA for 6 kV then 4mA for 1200 V. So 1200 V it is ok for 20 m only. For 200 m you need 40 mA at least-may be.

 

[Guineafowl]

In my opinion, if you used Clair v154 tester at 1250 V 50 Hz then it was not IR test executed but:
AC HiPot Test Programmable Voltage Range 0.10 kV-5.00 kV (10V/Step) Frequency Independent of Supply 50 or 60 Hz Voltage Display and Accuracy 0.10 kV-5.00kV ±1% ± 5 Counts Voltage Display Resolution 0.01kV Current Display and Accuracy 0.01mA-20.00mA ±1% ± 5 Counts Current Display Resolution 0.01mA Selectable Range of Pass/Fail Levels* 0.01mA-20.00mA Maximum Current Output 20.00mA @5kV Optional Arc Detection 9 Levels
according to Instruction Manual:
http://www.seaward.co.uk/downloads/clare_hal_combined_user_manual_455a553_154_538_rev12_v1.pdf
page no.90

20 mA for 6 kV then 4mA for 1200 V. So 1200 V it is ok for 20 m only. For 200 m you need 40 mA at least-may be.

That manual’s for a different machine. Mine is like this:

The Clare/Seaward HAL is much more modern and on a different level in terms of cost and functionality.

The old 152 and 154 do an excellent 25A test of the Earth path, to ensure not only continuity but also integrity of the Earth path. Any feeble/frayed/corroded connections will be revealed.

 

[Babadag]

Thank you, Guineafowl.