electrical earthing again

Got very useful input from here last week - thanks again. My home electric circuits were tested as ok some weeks ago, including adequate earthing, when a survey was done for a new combi boiler. When they came to fit boiler, this was abandoned when initial tests showed earthing inadequate - nearly 5 ohms instead og < 1. In the interim new street mains and house supply gas pipes were replaced with plastic and one possibility is that my equipotential bonding to old iron gas pipes was helping earth my system, instead of gas pipes being earthed to my system.

Impedance of about 5 ohms was measured at earth connection on sockets. Same was measured at what was supposed to be the actual main earth; an earth wire from the consumer unit - what I call the fuse box!! - to an ordinary earth clamp

formatting link
round the pipe containing the incoming mains electricity supply. The theory, by both British Gas (boiler fitters) and Scottish Power (DNO for my area) is that the earthing supplied by this pipe is inadequate or has become compromised.

One electrician did try scraping the pipe a bit at this earth connection to get a better connection, without success. However, I've since noticed that what looks like copper is more like some ancient gunge/paint/corrosion which I've scraped down to a shiny metal surface and re-attached the earth.

I'd like to now measure the earthing impedance at this earth connection down the pipe and, if possible, at the sockets. After initial crude tests with 'socket and see', I think electricians previously tested it with multimeter. I know enough about volts amps and ohms to know what they are, but don't usually touch electrics - but I do enough about electrics to know what not to try!!

What I don't understand is how to find impedance down earth pipe and at socket with a multimeter. I'm not measuring the resistance across something but down something where I dont have access to the other end. Do I attach attach one probe to something non-conductive and the other to the pipe to get the resistance to leakage in one direction, and how do I safely get a reading at socket (and no, I do know enough not to poke into the live supply).

There is a delay in the DNO attending - I'd be interested to know whether things are safer meantime and to what extent what one electrician saw as the possible problem, might have been corrected.

Regards

Toom

Reply to
Toom Tabard
Loading thread data ...

In practice, you can't. You need a proper tester which just looks a bit like a multi-meter. It puts a very large load to earth on the circuit for a fraction of a second and measures the current flow, and from that calculates the impedance. They are quite expensive devices and need regular calibration.

Reply to
Dave Plowman (News)

Put a photo up on photobucket so we can have a look.

Reply to
ARWadsworth

Wouldn't a 3 bar electric fire between live and earth do the job? Where you measure the voltage (change) between earth and neutral?

Resistance = voltage / current through fire?

Reply to
Fredxx

I'd say your assessment of the earth being supported by the old metal underground pipes is correct. Forget all the above comments about using ground spikes on their own and get a pro with an earth fault loop impedance tester to check it correctly. Its very unlikely that an earth spike or multiple earth spikes will be adequate to solve your problem in most parts of the uk as you will typically have ground loop resistances in excess of 15 ohms and maybe as high as the hundreds. Unless the supply undertaking are willing to provide you with a low impedance earth terminal you should change to a TT system with RCD protection.

Reply to
cynic

The way in which an earth loop impedance meter will test, is to apply a known voltage to the earth, and measure the current that flows down it. From that it can compute the earth resistance. The older style meters did this by creating a large earth fault current (i.e. up to 25A or so). These give a good test of the earth under load conditions, but also trip any RCDs in the process. Modern "non trip" meters can also do low current tests, which while not quite as good, will give a reasonable reading.

To replicate this yourself, you would need a way of *safely* connecting live to earth via a current limiting load of some form, and then measuring the voltage dropped across the load and the current flowing through it.

The theory:

With a shunt load (S) of some sort, you could measure the supply voltage across it with a volt meter at (V). Say under normal circumstances you measured 240V, and 5A through the load (with the amp meter (A) in circuit with it). You could deduce the load had an internal resistance of about 240/5 = 48 ohms

S L ------- A ------\/\/\------ N | | | | ---- V ----

Now connecting that between L and E on a non RCD protected circuit:

S Ze L ------- A ------\/\/\------ E ----\/\/\---- | | | | ---- V ----

With a perfect earth (or at least a decent one) you would get very similar readings. With a Ze of any significance however, then you have in effect a potential divider made from two series resistances, of which you know one and can measure the effect of both.

So say you measure 4.5A at (A), you know the total resistance must be

240/4.5 = 53.3 ohms. Which suggests a Ze of 5.3 ohms, if the shunt load represents 48 ohms (which of course you can check, since it should now be dropping 4.5 x 48 = 216V at (V)).

The practice may be a bit harder depending on your equipment and what ranges your meter has. Many meters don't have much capacity on their AC current ranges (if they have one at all!), which would limit the size of shunt load you can use, and low current shunts will equate to high internal resistances, thus reducing the observable effect of the small external resistance of the earth. You would also need to take care with the shunt to choose something with a purely resistive load, and preferably one that does not change resistance too much with current.

Reply to
John Rumm

Nice explanation. But for the process to work, there is an underlying assumption that there is a practically zero ohm connection between N and E somewhere out there (i.e. that N and E are well bonded together at the substation or wherever the supply is coming from). That being the case, why would it not suffice simply to measure the DC resistance between N and E directly with an ordinary multimeter?

Another question is how nearly zero "practically zero" really is. Where highish currents are involved, there will be a voltage drop in the mains cable supplying the house, i.e. across the cable impedances Zl and Zn in the diagram below.

Zl L ----------------/\/\/\/\/\----------------- L

House N ----------------/\/\/\/\/\--------------+-- N Substation Zn | E ----/\/\/\/\--+ +--/\/\/\/\--+ Zeh | | Zes | Zee | +--/\/\/\/\--+

Zes is the earth impedance at the substation end which is presumed to be of very high quality (of the order of milliohms I guess), and Zeh is the earth impedance at the house end which is under test. Zee is the "real" (non-cable) earth path impedance which is presumed (isn't it?) also to be virtually zero.

Now when you connect your shunt in the first instance between L and N for calibration purposes, and subsequently move it to between L and E, what you are calculating from your measurements isn't Zeh, but it is the amount by which Zeh+Zee+Zes exceeds Zn. Zee+Zes can be disregarded as being of a lower order of magnitude, but I wouldn't expect Zn to be small enough to ignore when we're expecting Zeh to be less than 0.3 ohm.

Is it normal practice to measure Zn (actually to measure Zl+Zn and to assume Zn is half that) by measuring the voltage between L and N under no-load and known high-load conditions, and to take this into account when calculating Zeh?

Reply to
Ronald Raygun

That would work well, I think, in the absence of more sophisticated equipment.

Many years ago I was having problems with hum in a particular piece of equipment. The problem only occurred first thing in the morning, so I swapped out the obvious part for further examination. When I'd done so, every thing was perfect, as anticipated.

Until the following morning!

I missed that day's display but turned up bright and early the following day to witness the problem at its most severe. It defied all logic and I was desparate but luck was on my side. It was an office in old building and the corner I was working in wasn't very well lit - it was mid winter and natural light was virtually non-existent.

As I disconnected an earthed cable I saw a minute spark between the connector and the earthed chassis. I plugged it back in again and, almost immediately, I noticed the fault condition sharply reduce in intensity - and realised that a girl was just returning to her desk after turning down the fan heater they were using to supplement the central heating ...

I immediately realised that earth and neutral must be reversed somewhere along the line and connected a meter between the 'earthed' chassis and the cable that I knew was most definitely earthed down in the basement. Each kilowatt step on the fan heater produced a voltage drop of 500mV or so and the hum was obviously being injected by the hefty AC current passing through the screen of the cable.

I was deceived on my first visit, of course, because, by the time I'd got the kit going again, the temperature had risen sufficiently for the temporary fan heater to be switched off completely for the rest of the day ...

Reply to
Terry Casey

Indeed - and we are also assuming that the supply really is a TN-S one, where the "pipe" in question is actually the main earth conductor, and not just an extraneous bit of pipe!

I would not chance it due to the risk of there being a potential difference between the highly loaded neutral conductor, and (relatively) unloaded earth connection (caused by other properties on the same phase). Not all multimeters take kindly to being placed across a live circuit on their ohms range.

Indeed, agreed. I was not suggesting this was a method of accurately measuring a "good" Ze (which for TN-S ought to be under 0.8 ohms), but more a way of detecting if its still in the order of 5 ohms and well out of spec for TN-S, although passable on simple plug in testers.

Yup if looking for small numbers (i.e. in spec) certainly. If looking for 5+ ohms, then the supply impedance itself can probably be disregarded.

Being realistic, unless the OP is very keen, he needs someone to come and plug a proper earth loop tester in. (unless he fancies buying one!)

(Measuring the supply impedance itself can be the topic of another post!)

Reply to
John Rumm

In article , John Rumm writes

:o) I think we've all tried to measure the resistance of the mains in an unguarded moment!

Reply to
Mike Tomlinson

I had the magic smoke from a well used Maplin Gold DVM when measuring mains voltage - and it was on the correct setting.

Investigation showed brass dust from the switch tracks has caused it to flash over. Managed to fix it, though.

Reply to
Dave Plowman (News)

Yup, I still have most of the probe left to prove it ;-)

(amazingly the meter survived (an ICE Multitest 80), although that ohms range was always a bit suspect after)

Reply to
John Rumm

I like the fuse in my 17th edition tester.

Reply to
ARWadsworth

The definition of a semiconductor is an expensive fast acting device to open circuit in time to protect a cheap fuse

Reply to
cynic

You want to see the price of the fast acting fuses recommended to protect the solid state relays controlling dad's kiln, I think he has given up on the fuses as the relays are cheaper to replace.

Reply to
Andy Burns

Not so expensive these days.

Compare the cost of a diode and a fuse..

Reply to
The Natural Philosopher

I came across that in a 110 volt 3 phase to DC transformer many years ago. The fast acting fuses were failing frequently costing about £3 to replace. The diodes they were protecting were about 80p each.

After replacing the fuses with cheaper ones fast enough to protect for overload, neither the fuses or the diodes failed.

Reply to
<me9

Just checked the emails and these were £13 each, with minimum quantity of 10.

Reply to
Andy Burns

It's the same with a stage lighting dimmer pack we had at a local theatre. However it was much quicker to replace the very fast-acting fuse than the triac.

Reply to
Frank Erskine

On a rather larger scale, the substantial semiconductors used in the Traction Convertor in APT-P had pretty large (1) and expensive fuses which appeared to be blowing spuriously.

Upon opening them up for investigation, the internal structure was of multiple parallel thin copper strips, with sections punched out to create lots of necking points in series. Close examination showed that the necking points which had shared the arcing when it blew all had little blobs of fused sand (with which the fuse cartridge was filled). However, there were also a fair number of clean breaks, and it was eventually determined that they were failing mechanically due to vibration, and thus as the number of parallel paths reduced, the current capacity slowly degraded.

(1) In the region of 1000 V, several hundred amps.

Chris

Reply to
Chris J Dixon

HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.