Surface mount sockets

[Dave Plowman (News)]

;-)

Now you know where dribble gets the plantpot from. As if...

BTW the 'many thousands' on reflection is an exaggeration. Make that hundreds.

[Tim Watts]

I did say that. I also said you cannot test an RCD properly with a DVM too (as an aside on testing, not to do with pyro).

There are many classes of DIY job. The ones that could be tested with nothing more than a DVM could include small scale PVC internal works.

Ones I think are best tested with the correct gear (at least R1+R2, end-end ring and RCD checks) would include new circuits with any complexity, including rings and lighting (the impedance tests *are* useful for discovering bad terminations), *anything* outside and pyro.

Except RCD tests. They do go wrong and it isn't totally impossible that a faulty one has made it onto the shop shelves (note recent protective device recall by well know manufacturer).

It is still a very good idea.

No, because it is *one* set of terminations that I can inspect manually, and putting the back of the plug on doesn't disturn them in the same way that putting a socket back into the wall does - that is where I had my broken neutral. All 6 conductors survived a pull test, the wires are neatly looped so as to collapse in a helical fashion when the socket was returned into its 35mm backbox. And the bloody wire still fractured. I suspect the screw had bitten it too much or something.

A clear case where the most careful workmanship in the world still resulted in a fault.

Re the plug, also I can visually inspect the lead. I still do a rough and ready appliance test on some of my stuff, eg external cement mixer, extension leads, Class 1 second hand stuff: (500V IR with Megger and earth loop impedance test, and R1+R2 as well for extension leads).

You should have said before. That is worth knowing, in which case the OP should at least do that.

Have you?

No, because as yet I have had no reason to use pyro - but I am well aware of its failure modes and *I* wouldn't be happy unless I'd checked them. You're happy because you've done loads before.

Which is entirely possible with cable of unknown origin (ebay).

Compound fault on non RCD circuit (entirely possible):

Earth fails + L-E leaky. Pyro sheath and connected metalwork now live.

Of course, the earth failing is always bad, and compound faults are very rare, but it has been known for people to die due to one in a million compound faults.

Why take the risk - eliminate that which is known to be reasonably possible.

Yes - I do it myself...

[Dave Plowman (News)]

Then dry it out before using. It's very obvious when damp has penetrated - the MI ceases being a powder.

I'd say if you have an installation where this is possible it wants sorting and quickly. The whole idea of a safety earth is it can carry any current possible safely before the protection device activates. Anything else would be a nonsense.

In which case simply don't attempt any DIY. Rely on a pro doing it and all the possible safety checks. Even just changing a tap washer could introduce contaminates to drinking water.

You, as a pure DIYer, have all the 17th edition test gear? Which is regularly calibrated?

[Tim Watts]

Yes. Megger 1552. As you mention it, calibration happens to be due again

- had it done immediately I got it just to be sure the instrument was in good nick.

[Tim Watts]

I should add that the cost of instrument and first calibration = about 2 PIRs. I do my calibration every 2 years. I might drop that to 3 if the next one shows little or no shift from the last. I do dead-check it agains a resistor box before each session too.

Cost-benefit analysis - pays for itself if you can persuade the BCO to accept your EIC.

[Dave Osborne]

[ARWadsworth]

There are not often many faults until the kitchen fitter starts work

Adam

[dennis]

Do you know how hard it is to dry MI cable out? It wicks water in easily, getting it out again is hard.

[Dave Plowman (News)]

How often have you used it, Dennis?

[dennis]

Once. However the guy that was teaching me how was very good and he said its almost impossible to dry out and I believe him. I have never tried to dry it out and never will. The laws of physics backup what he said about it being hard to dry out. Not impossible but it would be quicker to remanufacture it IMO.

[Dave Plowman (News)]

Why would it be possible for something to get wet, but not be dried?

Neither have I. Simply never needed to.

Which laws would those be?

>

[John Rumm]

I have got separate 16th editions testers - Megger LT5 loop tester, BMD1 insulation and continuity tester, and CBT2 RCD tester. Not ideal, but cheap enough on eBay for most semi serious DIYers.

(I will probably upgrade them to a combined 17th edition tester at some point - the lack of a non trip loop test is a bit of a PITA)

[Tim Watts]

Hi John,

You may find my comments on the Megger 1552 useful then :)

1) Good instrument overall - easy to use, backlight, nice case that keeps everything together.

2) RCD auto test is handy.

3) There are reports of the selector know giving grief though - something to watch for if going 2nd hand.

4) Although the leads can be nulled for the dead impedance tests, they cannot for the PFC/Live LI tests which seems a fundamental design error. The calibrator will include the leads and standard probes when he calibrates the instruments, but if you use different probes (eg fused), leads or a wander lead then you need to make manual corrections.

6) No earth spike test, but Whitfield did describe a DIY test method for this occasion involving a car batter IIRC.

7) 3 year warrenty - it is possible to buy for a good price on ebay and still have some left.

[John Rumm]

Thanks for that...

A straight earth loop impedance reading on a TT system with any main bonds or other source of fortuitous earthing ought to give you that[1], I would have thought?

[1] given the resistance of any wiring up to the spike is likely to be negligible compared to the spike itself.

[Tim Watts]

I reckon so - because I fathomed *if* I ever TT'd anything, that's what I'd do.

Worth noting for anyone else reading that if we are doing a *live*[1] or PFC test through the earth rod, the rod will reach a potentially dangerous touch voltage for the short time the tester dumps upto X amps though it. Also, there will be a strong potential gradient formed in the vicinity of the earth rod - bad for people and more bad for animals.

So if trying this then my advice would be to cordon off the area around the spike for a radius of at least 10 feet whilst doing the test.

[1] A dead impedance test will not be worth anything as, in my view, it will be too susceptible to stray currents that may be present in the ground (eg other earth spikes, MEN bonding spikes on the main supply, spikes in lampposts or whatever. Proper TT testers use multiple electrodes to cancel out such effects which is why they're a bit special and not often included in standard multitesters (thought there is one IIRC that does include this feature).

[Andrew Gabriel]

It doesn't just get wet, the insulator turns from Magnesium Oxide into Magnesium Hydroxide. This can in principle be reversed, but not in a conventional oven (not hot enough) and not inside a semi- sealed copper tube where the magnesium oxide is under pressure and the released water can't get out, so it will simply absorb back in again when the magnesium oxide cools. I guess if you heated it up (to > 332°C, when Magnesium Hydroxide decomposes) for the same number of days that it was exposed to moisture, you would get half the moisture out via the path it got in (to a first approximation).

It's a not uncommon problem that moisture gets into mineral insulated heating elements and generate earth leakage. In this case the hot element can displace it to the colder outer jacket and stop the earth leakage, but when it's cooled, the moisture will disperse back through the insulation and generate earth leakage again. (PAT testing does not regard a temporary leakage of this type as a fault, providing it clears when the element heats up.)

[js.b1]

A combined tester saves on individual calibration, but conversely gives a single point of failure.

You may want to keep hold of your LT5 a bit longer. LT5 are favoured for reliably verifying Ze at CU/DB origin because their measurements are highly reproduceable. Regulations permit Ze to be determined by measurement, or enquiry (DNO), or assumption based on supply type limits for TN-C-S (PME,

0.35ohm) or TN-S (0.80ohm). Zs can then be found by addition of R1+R2 direct measurement via an IR tester's 200mA continuity test. For TT you need a high current test or ideally a proper Earth Rod Tester, subject to safety requirements being met as mentioned.

Non-trip testers use a very small current test, 15mA. With rusty SWA, aged conduit or loose terminals that 15mA may indicate a compliant low EFLI for a final circuit when a true L-E short would actually experience higher resistance - potentially elevating touch voltages or even not causing OC fault disconnect. Yes RCD should discount a L-E fault, and permit longer cable run by negating EFLI limits, but good practice should still consider how good the tool is - calibrated or not.

Early non-trip testers used very poor algorithm, on a sensitivity analysis they were very poor. For example if you had same CU/DB fluorescent lighting then the two position of the MCB toggle could produce a 0.4ohm Ze or 0.8ohm Ze, pass or fail, as desired and results were not particularly reproduceable.

Later non-trip testers had better algorithms, so less sensitive AND results are reproduceable. However the short comings of non-trip EFLI testing remains, so for critical circuits bridging the RCD and using the higher current non-trip is preferable where touch voltage is of concern.

The two critical test tools are RCD tester & IR/Continuity tester, RCDs do drift out of spec & do fail. With increasing numbers of RCD & RCBO the need to routinely test has increased.

[John Rumm]

Yup, not saying I would necessarily get rid of the LT5. Although I understand some of the modern testers can also do high current tests as well as non tripping ones.

Yup agreed. I also prefer to design such that things are safe even without the RCD where possible. (although obviously LE faults on TT depend on it)

Indeed...

In fact lets lob a general interest question in here. I have seen several several techniques for bridging RCDs. Which do people favour?

[js.b1]

Confirm Live, confirm Dead, backout busbar screw on an RCBO, slide in MK "split-load" tail tab & tighten screw moderately, connect other end to DIN rail fuseholder with HRC fast-blow 32A fuse, move live to DIN rail live out, test, reverse.

I have only used it once to confirm Ze at origin CU, thereafter R1+R2 measurement since all RCBO.

You can just use VDE shielded test leads or insulated croc clip or Suck n See SS130 4mm plug top adapter. Obviously test probes will spark since a high current EFLI test is 24A typically. Others much prefer a 13A socket to Hirschmann shielded 4mm plugs for the SS130 or VDE croc clips to light ceiling rose, with DIN rail fuse holder inline if not fused leads (very important re flash burns on an accidental dead shorth re probes slip).

Cost about 45ukp to make up an R2 wander lead, various other leads, DIN Rail HRC fuse to 4mm plugs, SS130.

The industry lacks some REALLY stong clips for FTE sized conductors re weight of cable pulling things off, so I use washing line clips to take the load off them & long leads. "Klips", hook etc are too weak and the contact area sucks to be honest at very low ohmage levels.

[Dave Osborne]