Unusual Zs reading when testing circuit with RCBO

[Octopus]

Been testing a modified circuit today.

Ze = o.20
R1 + R2 = 0.77
Zs = 1.77

So the reading is about 0.80 higher than I was expecting... Nothing on this (lighting) circuit switched on..... I was testing at the end of a smoke alarm addition.

I've read elsewhere, that its a good idea to do a Zs on the outgoing terminal of the RCBO to look for the "addition" this device is giving.

Is this correct? Anyone be able to point me in the direction of any previous threads, or guidance?

Thanks

 

[Andy78]

Never seen the internal resistance of a device to be that high, but it would certainly be worth investigating.

I get a bit of variance with the low current Zs test on my MFT, but nothing in the order of 0.8 ohms

 

[Ian1981]

Been testing a modified circuit today.

Ze = o.20
R1 + R2 = 0.77
Zs = 1.77

So the reading is about 0.80 higher than I was expecting... Nothing on this (lighting) circuit switched on..... I was testing at the end of a smoke alarm addition.

I've read elsewhere, that its a good idea to do a Zs on the outgoing terminal of the RCBO to look for the "addition" this device is giving.

Is this correct? Anyone be able to point me in the direction of any previous threads, or guidance?

Thanks

I take it this was on a non trip loop setting?
I.e. 15mA test current?
Perhaps 'noise' on the circuit or the internal components of the rcbo causing it

 

[westward10]

Type B 6A and below can cause this on high current tests but that is a lot. Ditch the rcbo and retest on high current.

 

[Octopus]

I take it this was on a non trip loop setting?
I.e. 15mA test current?
Perhaps 'noise' on the circuit or the internal components of the rcbo causing it

Yes - non trip setting

 

[Ian1981]

I take it this was on a non trip loop setting?
I.e. 15mA test current?

The

Yes - non trip setting

Three-wire no-trip testing requires connections to be made to the live, neutral and earth conductors, and typically uses a test current of 15 mA or less. Its benefits are that there is no need to bypass RCDs and RCBOs during testing, which saves time, and that it does not trip MCBs. There are, however, a few limitations.


The results are not as consistent as those obtained with high-current testing, the instrument takes longer to make the measurements, the results can be affected by external factors; also the RCD internal impedance may occasionally affect the result. In addition, even though this is a no-trip test, it sometimes happens that existing and test earth currents add together to trip the RCD or RCBO.
These limitations only affect a minority of cases, however, and three-wire non-trip testing is the preferred technique for circuits protected by RCDs and RCBOs where access to the live, neutral and earth conductors is readily available.

Taken from the megger site.

 

[soulman]

Hi,
I have had this problem before, spoke to megger, if the Zs is much higher than anticipated at the end of the circuit, Take a Zs at the incoming side of the rcd/rcbo & take a Zs at the outgoing side = resistance of the rcd device. Deduct this resistance from your measurement at the light.
Cheers

 

[LankyWill]

I would investigate further, seems a bit high even for RCD/RCBO malarky. Im with @westward10 chuck a standard B- type in and retest. Try another circuit on the same board and see if you get a variance there. The joy of testing ey..

 

[Strima]

If you need to then measure the Zs. However it is still accepted to add your R1+R2 to your Ze.

A lot of people are now being trained to keep live testing to a minimum to reduce risk. As long as you have Ze/PSSC/PEFC/phase rotation etc then you can use your dead test results to complete your paperwork.

 

[telectrix]

once has a Zs on all circuits 350 ohms. suspected a TT with a crap rod, but investigation found that it was PME .Ze of 0.2 ohms.was an upfront RCD causing this. test on hi current this way.diagram courtesy of Richard Burns.:

 

[davesparks]

If you need to then measure the Zs. However it is still accepted to add your R1+R2 to your Ze.

A lot of people are now being trained to keep live testing to a minimum to reduce risk. As long as you have Ze/PSSC/PEFC/phase rotation etc then you can use your dead test results to complete your paperwork.

Which leads to poor earth connections going undetected and makes the whole point of testing rather pointless.

 

[Ian1981]

Which leads to poor earth connections going undetected and makes the whole point of testing rather pointless.

i would say dead testing will pick up poor earth connections as you are disconnecting any parallel paths which may influence the test.
A simple zs will involve leaving all cpcs and bonding conductors connected and will generally give lower than the calculated value of zs due to parallel paths and may conceal the fact that a cpc is not continuous on the tested circuit.
Both tests are applicable in initial verification except zs can be determined by calculation alone eliminating the need of live testing and keeping it to a minimum

 

[Deleted member 9648]

TBH I'm so used to inaccurate no trip low current tests (they are the rule rather than the exception) that I'd just accept the result is way under the maximum for the OCPD and crack on. As already stated dead tests have verified low resistance R1 and R2.
Spending time farting about bypassing RCD's on one circuit may be viable. But if you've got a large installation to verify.....it's going to be a late one.

 

[Octopus]

FWIW I'm back on site tomorrow so will test the Zs on the outgoing side of the RCBO out of interest.

Thanks for the input so far.

 

[Strima]

Which leads to poor earth connections going undetected and makes the whole point of testing rather pointless.

I sort of agree with you, but if the dead testing is done correctly then any loose earths should be identified at that stage.

Also live testing may not detect a poor earth if only testing at the furthest point, any spurs on the circuit may get missed.

TBH both methods have their merits and downfalls.

 

[Upton Sparks]

I've come across this as well on RCD and RCBO,

my understanding is that when doing the RCD test on most Meters on the non trip setting, the meter does this test by letting a very small current through down the earth conductor about 5 mill amps ish , and does this several time and then takes the mean value as the reading, therefore give a reading that doesn't always fall into the normal expectations

My advice would be to do your R1 + R2, record this value then take the Ze/ Zs at the DB and add these values together and record this as your value for the Zs for your circuit.

Just perform the Zs at the final point of the circuit for more of a confirmation that there is return path and piece of mind.

 

[davesparks]

I sort of agree with you, but if the dead testing is done correctly then any loose earths should be identified at that stage.

Also live testing may not detect a poor earth if only testing at the furthest point, any spurs on the circuit may get missed.

TBH both methods have their merits and downfalls.

Dead testing does not pass enough current to highlight any fault in the earth path, the 200mA of the continuity test will not reveal poor connections. The usual 6A of the high current Zs test used these days is only a little better, but unless we are to return to using the test rigs of the past or the ductor type tester then it's the best we can do.

A method which relies on calculation doesn't prove anything at any point in the circuit, it's purely mathematical and entirely pointless in that respect.

I can't see any merit to the calculation method, modern test equipment and good working practice reduces the risk associated with testing live equipment.

 

[davesparks]

I've come across this as well on RCD and RCBO,

my understanding is that when doing the RCD test on most Meters on the non trip setting, the meter does this test by letting a very small current through down the earth conductor about 5 mill amps ish , and does this several time and then takes the mean value as the reading, therefore give a reading that doesn't always fall into the normal expectations

My advice would be to do your R1 + R2, record this value then take the Ze/ Zs at the DB and add these values together and record this as your value for the Zs for your circuit.

Just perform the Zs at the final point of the circuit for more of a confirmation that there is return path and piece of mind.

Better to leave the box empty than to record a calculated value which gives a false impression of a measured value.

 

[Upton Sparks]

Better to leave the box empty than to record a calculated value which gives a false impression of a measured value.

not at all.. and I disagree, leaving a blank box is worst thing to do... unless you put a long comment on as to why you've left it blank. .

there are 3 ways to obtain a result,
1.Measurements
2.Calculation
3.Enquire

your perfectly in your rights to calculate the result, as long as you do your dead test correctly.

Also as someone else had already mention above the point about this method being pushed as best practice for testing as it reduce the need for live testing, and anyone who's had to write some RAM's (risk assessment and method statement) will know the grief you've got in justify why your working live.

 

[davesparks]

not at all.. and I disagree, leaving a blank box is worst thing to do... unless you put a long comment on as to why you've left it blank. .

there are 3 ways to obtain a result,
1.Measurements
2.Calculation
3.Enquire

your perfectly in your rights to calculate the result, as long as you do your dead test correctly.

Also as someone else had already mention above the point about this method being pushed as best practice for testing as it reduce the need for live testing, and anyone who's had to write some RAM's (risk assessment and method statement) will know the grief you've got in justify why your working live.

I don't my see the issue in risk assessing live testing, it's certainly not the same thing as working live.
The risk is controlled by a high level of training and competence along with modern test equipment having being built to a high standard of safety. Meters have built in sensing to detect incorrect connection and not perform the test, along with fused shrouded leads.
Then if you identify a particular risk from say the design of some old equipment you need to test then you can use the alternative Zs test method to move the point at which you connect to the live supply to a point where it is safer to do so.
I'd be more concerned about the assessed risk of not carrying out live testing, the risk of poor earth connections going undetected because the calculated value makes it appear to be all ok.

 

[Strima]

■ it is unreasonable in all the circumstances for the conductor to be dead;

It needs to be live so I can test it as I'm doing live testing

■ it is reasonable in all the circumstances for the person to be at work on or near that conductor while it is live;

Those 6' test probes are really hard to control

■ suitable precautions (including, where necessary, the provision of personal protective equipment) have been taken to prevent injury.

I have received suitable training in the use of test equipment, my tester is GS38 compliant and I'm wearing suitable PPE such as hand and eye protection.

 

[Upton Sparks]

I'm not advocating not to do Live testing at all. I see the benefit in this, a proven measurement that the circuit will disconnect in the required time.

Now whether I agree with there the current guidance of how to test is neither here nor there.
I'm just specking of the health and safety world we living in, it's the employer responsibility to removed,reduce and then manage the risk's... That's the Law.

going back to the problem of, A difference in readings values from what is expected from measured, both readings would be a pass, only one is an unexpected value, and we can all see why its unexpected as we are all taught to do Zs=Ze+ R1+R2, so we are checking that our calibrated meter is working correctly, as it does its magic...

you can in this instance spend the next hour or so investigation if there is a loss earth or not. you can pull the whole circuit apart if your that way inclined, buzz out each leg measure it, and put it back together and add all your measure legs up, do your testing again and see where you are...... and maybe you had a lose connection or maybe your back to square one.

But in my Experience its the RCD tests which most of the time cause this slight difference.

there is nothing wrong in calculation your results and proving it or supporting it by a live test to prove earth continuity.
which would then be recorded on the Cert as comment.

 

[Ian1981]

Not sure there's even a disagreement here as either calculation or direct measurement by live testing is both allowed and an acceptable method advocated by scheme operators and by bs7671, GN3 and multiple guidance books.
Scheme operators like the calculation to eliminate the need for live testing but it's upto the individual testing.

 

[Octopus]

So I tried testing the same circuit on the outgoing side of the RCBO.... hardly any different to the Ze....

So tested another new completely new circuit today.

R1 + R2 = 0.10
Ze = 0.20
Zs = 1.01

Hum ..... a bit puzzling

So next step will be next week, when a mate is lending me his spare MFT to compare results.......

I'll keep you posted.

Thanks for the input so far.

 

[Octopus]

So I borrowed my mates tester, and on 1 quick test:

My Metrel tester Zs = 0.52 ( i used a plug and tried my 13A plug lead set, plus my kewtech plug in with the leads - the results were similar

Mates tester Zs = 0.38

So a variance between testers, not as much as had seen, but a difference all the same

I recall somebody a while ago mentioning a "spray" to use on tester connections - anyone know what it is?

 

[Andy78]

Electrical contact spray ?

 

[Andy78]

What brand was your mate's tester Murdoch ? And were the comparison tests low current "no-trip" tests ?

 

[Sparky Ninja]

I know a company that uses this..

Aerosol - ESC Electrical Contact Cleaner Spray - 500ml - [AT-1487] - http://www.thesafetysupplycompany.co.uk/p/8954323/aerosol---esc-electrical-contact-cleaner-spray----500ml---at-1487.html?gclid=CjwKEAjw4vzKBRCt9Zmg8f2blgESJADN5fDgvfGfr2lGv4q91BRXOKGuZVQs0qVr5sEUVOuKGhBU6RoCMKXw_wcB

on their supply leads in a factory.

I don't think you really need to worry too much about it though.

Fundamentally, it appears as if the protective devices are introducing significant impedances to question the effectiveness of your protective measure. If I was in that position (assuming I can personally verify the accuracy of my instrument) I would accept and record the readings measured. If these are too high I would try a replacement and would contact the manufacturer for their spiel, and maybe record it on the documentation.

Using calculation to compare your measured value, is for purposes exactly as you are demonstrating here.

For the argument of referring to calculation instead...

Opting to refer to calculation over measurement would be naive. In a real world scenario, a fault current is going to flow down your measured impedance, not your calculated impedance. Calculations, while useful, never compensate for connection resistances, contact resistance or arc gaps.

Yes, the H&S world, Training courses, Guidance Note 3 and the incoming 18th edition all suggest EFLI can be verified by calculation with the measured protective conductor continuity and supply impedance loop. But these fault currents don't flow down imaginary numbers.

Also, the disconnection of a protective conductor when verifying continuity will remove the parallels. When we measure Zs we want the parallels.We need to start paying more attention to the pefc and pscc (and therein the pfc) within our circuits and not just at the origin or submains. The 18th edition is planning on introducing terms like:

• rated short-time withstand current, Icw (already there)
• rated peak withstand current, Ipk
• rated conditional short-circuit current, Icc

These have been around for powertrack busbar systems but are now going to be more commonly used for any assembly. So we will need to set our testers to pfc (which means it will do an EFLI test and play with ohms law) and make sure all the parallels are there to ensure we measure the highest pfc value.