Di-Log 9083P 17th Edition tester

[Kilovoltkid]

I just bought the above and found it won't carry out the RCD tests on a brand new house. Specifically, it registers either >25V or >50V, suggesting there is a high fault loop impedance path that would raise the earth potential to an unsafe value. I returned the tester thinking it was faulty as I borrowed a Fluke 1653 which worked fine giving me the x1/2, x1 and x5 values without a problem. I was told by Di-Log that the 17th edition is more strict than the 16th on such matters and the tester correctly refuses to continue with the test whereas the older Fluke must not compatible with the new regulations. Does anyone know of a similar problem and how to solve it?

 

[Spark123]

What were the Ze & Zs results like?
Has the 17th edition changed ohms law?

 

[Kilovoltkid]

Ze was 0.34 Ohm at the consumer unit and Zs was 0.46 Ohms on the longest downstairs ring circuit; Ze was 0.21 measured by Southern Electric at the meter. Typical RCD readings were 27.1ms at x1 and 29.7ms at x5 for the ring circuit (as measured with the Fluke).

I think the point Di-Log was making is that older tester (Robin for example) would chuck a high charge onto the lines making the installation possibly dangerous to others whilst the test is occurring or worse leaving it that way afterwards. It seems there is a BSEN relating to test instruments that has altered and the 17th edition simply refers to it.

 

[ban-all-sheds]

I was told by Di-Log that the 17th edition is more strict than the 16th on such matters and the tester correctly refuses to continue with the test whereas the older Fluke must not compatible with the new regulations.

...the 17th Edition does not introduce
any significant changes in the
requirements for the testing of RCDs
even where they are installed to
provide automatic disconnection in
the event of a fault.

 

[Spark123]

Ze was 0.34 Ohm at the consumer unit and Zs was 0.46 Ohms on the longest downstairs ring circuit; Ze was 0.21 measured by Southern Electric at the meter.

Why the large variation between your Ze and the DNO Ze?

I think the point Di-Log was making is that older tester (Robin for example) would chuck a high charge onto the lines making the installation possibly dangerous to others whilst the test is occurring or worse leaving it that way afterwards.

An RCD tester needs to deliver 0.5I, I (and for up to 30mA devices 5I) for a timed period. It may monitor the voltage rise on the protective conductor and cut the test if this exceeds 50v. In a TN system with the impedances you mention the resulting voltage for the test will be very small - for a Zs of 0.46ohms even with 150mA the result is less than a volt.

 

[Kilovoltkid]

Didn't witness SE measuring Ze so I only have their word they actually carried out test rather than cited typical values but there are long tails of approximately 6m installed.

I have written to Di-Log (so I have it on paper) asking how their unit test RCD times and how that differs from Fluke. I intend to try out the tests again once unit is returned taking detailed comparisons with Fluke. At the moment I am still puzzled as to why one unit works on a given installation and the other doesn't.

 

[jason1958]

There are 3 ways of doing EFLI measurement.

1) Calculation

- Ask DNO what service type & Ze (eg, TN-C-S & 0.35ohm)
- Measure R1+R2 (or r1+r2 etc as appropriate based on f.c. type)
- Zs = Ze + (R1+R2)

2) 20-25A whack via EFLI meter doing High Current EFLI test

- This requires a non-RCD protected circuit
- It does confirm at least the earth is better than carbon tracks

3) 15ma squeak via EFLI meter doing RCD no-Trip EFLI test

This is where the fun starts...
- Firstly your earth path might actually be carbon & vapourise before CPD disconnects the fault
- Secondly EFLI is calculated via a complex algorithm which is greatly affected by noise - sensitivity analysis is poor

Sensitivity...
Test is performed at 15ma, does not take much noise to affect your signal.
- Eg, any connected devices (washing machine, fridge, PC have filters) which can affect the result (and give a voltage on the line)
- Eg, any noise on the line from fluorescent circuits on the same CU/DB with CPD left "on" can affect the result (earlier Megger CM500/LT7 very sensitive here)

Accuracy...
Each machine has its own accuracy.
- Some are +/- 10%.
- Some are +/- 2.5% with 4 digits - so low impedance readings error margin is quite significant

Calibration...
Realise most meters are only calibrated at quite high impedance
- Calibration may be 0.72ohms displayed under 0.70ohm test
- Thus accuracy below that is subject to unknown error

Whats good on low current EFLI testing...
- Recent Fluke are actually quite reliable
- Megger MFT1552 etc are also well regarded
- Socket & See PDL200/234 & 310 are also consistent & accurate
- Older Megger CM500/LT7 can be too sensitive & unreproduceable

Good?
i) results from high-current tests are ball-park with low-current tests
ii) results are reproduceable. Sensitivity analysis is good

Batteries can be very important.
- Some machines suffer drift as batteries get down to half full


Low current tests are sensitive.
Sensitive to high noise, high harmonics, close to TX, lots of fluorescent lights, any loads left connected somewhere with filter networks (washing machines, alarms). As you get below about 0.7ohms they are indicative, and any really low reading (0.40ohm & below) on some meters can be fiction in that a high current test will give quite different results.

It can be annoying - on a TN-S supply it is easy to get into pass/fail/pass/pass/fail/fail on repeated tests with some low-current efli tests on high rated CPD. So if in doubt, bridge out & use a high current test (realising the risks). Alternatively measure Ze & then R1+R2 - the regs permit "dead calculation" which with a battery of RCBOs and a noisy supply might actually be necessary with some older meters.


So if you are finding low-current EFLI test results vary, I am not that surprised.

I recall a PDL-420 gave 0.40-0.62ohm on low current test for Zs, when an LT7 on high current test gave 0.21ohm (which was close to DNO test figure). The LT7 on low current test gave 0.6-0.8-1.2ohm, but its algorithm is known to be extremely sensitive. Some makers say check batteries are fresh & repeat test 3 times taking the average. I think even some of the older Fluke (1651?) benefited from that.

If that is a new DiLog I would perhaps look at the MFT 1552 etc.
I prefer 1-use machines (Megger for RCD, Robin for IR+Continuity).


The full set of tests is x0.5 (no trip), x1 (trip <200ms), x5 (trip <40ms), with test performed at 0-degree & 180-degree phase rotation (up/down cycles in AC voltage waveform). Some meters can not do the phase rotation bit, cutprice multi-testers leave some bits out (Kewtech?).

 

[Taylortwocities]

Jason, fantastic overview!

I do despair sometimes. We are driven to buy, and keep calibrated, expensive test equipment. The reason is to ensure that we are installing systems that meet the requirements of stringent standards.

Testing circuits that are protected by RCDs (and thats all of them in this 17th ed world) is largely now guess work as many testers give widely varying results for "non-trip" RCD tests verus what you get from a
Ze +R1 + R2 calculation.

I have a Metrel EasiTest and that gives a good 0.5 ohm difference between non-trip and Ze +R1 + R2

PS. Before you ask, that Ze value is with all bonds in place..... anyone know what Z that is??

 

[ebee]

Do you want to call it ZeB?
(Ze with Bonding in place)

Or Zzg

zg = zola gorgon

if you put some zola in a bowl ang mix in some gorgon you end up with zolagorgon which is silly because there is no such thing.
But mix them the other way round and you end up with a well known cheese

The only reason you`d take a measure of Ze with bonds in place is to get max PFC not a measure of Z in itself

 

[Taylortwocities]

Indeed, but what I was trying to do was to compare a Zs measurement done with a no-trip test function against what the measurement is if done another way.

The other way would be to strap out the RCD and do a Zs meaurement.
This value would be the ZeB + R1 + R2 value, wouldn't it?

 

[ebee]

Yup I can see what you`re doing here