Understanding the relationship between TNCS, TNS and TT

[HappyHippyDad]

I have never actually had to make an electrical system into a TT. I have added rods in order to get an effective Ra but never had to (for example) install a TT on a new garage or shed.

It's best I give an example with regards my question..

A new garage has a supply taken to it from the house. The house is PME and this has been extended to the garage, lets say 4mm 3 core SWA. The 3rd core is perfectly adequate as an earth. Also, the installer puts a copper rod in the ground and attaches this to the earth bar in the new garage CU. Is this OK? Is this actually better? Is it now dangerous?

I've got a few ideas but I'd rather not influence the line of thought and just leave it as above.

Cheers all.

 

[Simon47]

Just it would be far more cost-effective for the DNO to take additional steps ...

What additional steps might they take ?
It can't be monitored for at the substation - all they'll see there is a change in loads that probably is insignificant relative to the total load.
Monitoring the PEN-true earth current might give a clue - but ONLY if consumers downstream of the PEN break do have effective earth rods so as to be able to create a significant earth leakage current. And this would need careful management so as to avoid nuisance tripping. It would also be negated (or at least significantly reduced) in the presence of multiple PME earth points - since I learned a bit more about it, I've started noticing PME earth conductors down the side of poles.
Realistically the only place it can be monitored for is at the last property on a circuit or the last tap off in the street. That's a lot of points to monitor, and would be prohibitively expensive.
As an aside, some years ago at a previous job, our local DNO (forget what they were called back then) asked if we'd host a monitoring box - it plugged into the mains and a phone line, and would phone home if the power went off. Presumably they'd investigate if they got several calls - one call might just be an internal issue in the property.

In retrospect, smart meters could have done that and actually been useful, but that was not considered and cost saving on the design (coupled with general incompetence of the whole project management) mean it probably never will be.

Indeed, one of the benefits (cost savings) used in the business case for smart meters was the ability to monitor supply voltage at many points - and thus improve network management and fault responses. Unfortunately, it seems that the people who created the spec for the meters didn't talk to the people who would be using the measurements - and the result is that the voltage measuring ability of the meters is not sufficiently accurate (or it might not even be mandated, can't remember now) for the required purpose. AIUI, last time the business case was re-evaluated, this benefit was still being counted

Am still of the view that the rods down stream of the break won't matter a jot but need to fully tease it out.

How can they NOT have an effect ?
If we exclude the "very large resistance so ineffective" issue, because that's much the same as "no earth rod" really.

So lets start from the situation with no earth rods. If a PEN break occurs, the neutral and MET in each downstream property will take up a (dynamic) voltage (relative to earth) which is dependent on the loads attached to each phase.
Basically, the phase currents must sum to zero as there's no neutral to carry the imbalance.
If we start with the simple case of all resistive loads, then the current in each phase is proportional to the local P-N voltage - a phase with a larger set of connected loads will see a reduced P-N voltage until the load current reduces (and the other phase currents increase) such that the phase currents sum to zero. The extreme situation is a load on one phase, and nothing on the other two - the loaded phase will see zero P-N voltage while the other two phases will see 415V. At the other extreme, the loads are all equal, so the neutral will be roughly at earth potential and all properties will see a normal 240V supply.
Add in non-linear loads like SMPSUs and it gets quite complicated. As the P-N voltage reduces, the load current will increase - so in a network with all such "negative impedance" loads, the tendency would be for the neutral voltage to shift rapidly towards one of the phase voltages until the attached loads shut down due to undervoltage.

Now lets introduce one or more earth rods. Now downstream of the PEN break there is a N-E connection, and thus a connection via the earth to the DNOs earthed neutral system. This will be a complex path involving the substation earthing arrangements, any PME earth points in the network, and any customer earth rods upstream of the PEN break.
We no longer have the requirement that the phase currents sum to zero as there's a (relatively high resistance) path for the neutral current. So assuming we have imbalanced loads, the phase with the highest load will still see the lowest P-N voltage. But the effect will be less due to some of the imbalance returning to the substation neutral via the earth.
Lets take the example values mentioned above - 10 houses and a 20ohm earth rod each. Excluding the resistance of the connecting cables, that's a 2ohm connection to earth. Lets say the phases are imbalanced by 10A, that 10A will try and return via the 2ohm earth, raising the local neutral to around 20V relative to earth. So one phase will now be at something like 220V and the others will be correspondingly increased (by a bit less than 20V - you need to draw the phasor diagram).

So adding the local earth rods has turned a "will blow up lots of stuff on 2 phases and maybe set houses on fire" event into a "incandescent bulbs burn brighter and pop but otherwise nothing newsworthy" event.

Clearly the actual effect seen in any event is going to depend on the combined (parallel) earth rod resistances, and the nature of the loads attached at any point in time. And loads will be dynamic - someone putting the kettle on will change the situation significantly.

 

[LastManOnline]

What additional steps might they take ?
It can't be monitored for at the substation - all they'll see there is a change in loads that probably is insignificant relative to the total load.
Monitoring the PEN-true earth current might give a clue - but ONLY if consumers downstream of the PEN break do have effective earth rods so as to be able to create a significant earth leakage current. And this would need careful management so as to avoid nuisance tripping. It would also be negated (or at least significantly reduced) in the presence of multiple PME earth points - since I learned a bit more about it, I've started noticing PME earth conductors down the side of poles.
Realistically the only place it can be monitored for is at the last property on a circuit or the last tap off in the street. That's a lot of points to monitor, and would be prohibitively expensive.
As an aside, some years ago at a previous job, our local DNO (forget what they were called back then) asked if we'd host a monitoring box - it plugged into the mains and a phone line, and would phone home if the power went off. Presumably they'd investigate if they got several calls - one call might just be an internal issue in the property.


Indeed, one of the benefits (cost savings) used in the business case for smart meters was the ability to monitor supply voltage at many points - and thus improve network management and fault responses. Unfortunately, it seems that the people who created the spec for the meters didn't talk to the people who would be using the measurements - and the result is that the voltage measuring ability of the meters is not sufficiently accurate (or it might not even be mandated, can't remember now) for the required purpose. AIUI, last time the business case was re-evaluated, this benefit was still being counted

How can they NOT have an effect ?
If we exclude the "very large resistance so ineffective" issue, because that's much the same as "no earth rod" really.

So lets start from the situation with no earth rods. If a PEN break occurs, the neutral and MET in each downstream property will take up a (dynamic) voltage (relative to earth) which is dependent on the loads attached to each phase.
Basically, the phase currents must sum to zero as there's no neutral to carry the imbalance.
If we start with the simple case of all resistive loads, then the current in each phase is proportional to the local P-N voltage - a phase with a larger set of connected loads will see a reduced P-N voltage until the load current reduces (and the other phase currents increase) such that the phase currents sum to zero. The extreme situation is a load on one phase, and nothing on the other two - the loaded phase will see zero P-N voltage while the other two phases will see 415V. At the other extreme, the loads are all equal, so the neutral will be roughly at earth potential and all properties will see a normal 240V supply.
Add in non-linear loads like SMPSUs and it gets quite complicated. As the P-N voltage reduces, the load current will increase - so in a network with all such "negative impedance" loads, the tendency would be for the neutral voltage to shift rapidly towards one of the phase voltages until the attached loads shut down due to undervoltage.

Now lets introduce one or more earth rods. Now downstream of the PEN break there is a N-E connection, and thus a connection via the earth to the DNOs earthed neutral system. This will be a complex path involving the substation earthing arrangements, any PME earth points in the network, and any customer earth rods upstream of the PEN break.
We no longer have the requirement that the phase currents sum to zero as there's a (relatively high resistance) path for the neutral current. So assuming we have imbalanced loads, the phase with the highest load will still see the lowest P-N voltage. But the effect will be less due to some of the imbalance returning to the substation neutral via the earth.
Lets take the example values mentioned above - 10 houses and a 20ohm earth rod each. Excluding the resistance of the connecting cables, that's a 2ohm connection to earth. Lets say the phases are imbalanced by 10A, that 10A will try and return via the 2ohm earth, raising the local neutral to around 20V relative to earth. So one phase will now be at something like 220V and the others will be correspondingly increased (by a bit less than 20V - you need to draw the phasor diagram).

So adding the local earth rods has turned a "will blow up lots of stuff on 2 phases and maybe set houses on fire" event into a "incandescent bulbs burn brighter and pop but otherwise nothing newsworthy" event.

Clearly the actual effect seen in any event is going to depend on the combined (parallel) earth rod resistances, and the nature of the loads attached at any point in time. And loads will be dynamic - someone putting the kettle on will change the situation significantly.

My apologies. Got my "downstream" and my "upstream" mixed up. Just had two 14 hr days. Look forward to reading your post thoroughly tomorrow

 

[pc1966]

It can't be monitored for at the substation - all they'll see there is a change in loads that probably is insignificant relative to the total load.
Monitoring the PEN-true earth current might give a clue - but ONLY if consumers downstream of the PEN break do have effective earth rods so as to be able to create a significant earth leakage current. And this would need careful management so as to avoid nuisance tripping.

As you point out, it is very hard to detect an open PEN section at the substation alone. The overall "earth leakage" might be able to show it, but given the multiple earths in place normally that is also unlikely to be apparent from one monitor point.

Monitoring remote branches is the obvious method, and given the rise of EVs and the low likelihood of DNO cables and substations being upgraded to allow ~30% of households to switch on 10kW+ charges in the evening, it would seem the smart approach is to have the EV charges reporting to the local substation in some sort of standardised manner.

That way the charge rate can be adjusted to keep cable/transformer limits OK, and the reported voltages then could be used to detect a PME fault region. Given the EV charges have to deal with the open PME risk somehow, it would seem a good way to cover this aspect for other households.

 

[Vortigern]

When I started training some 45 years ago I often came across installations that had no earth whatsoever. Just L - N, it struck me that was a better system as if there was a break in the N the electrics wont work in the installation and there would be none of the complications as above. Adding an earth has brought its own problems I think it is a redundancy. Thinking outside the box, I say let's do away with earth all together and avoid all the scenarios above and solve the problem in a single stroke. It is ironic that Earth was introduced as an extra safety system and has become one of the most confusing and misunderstood systems, which brings a host of problems/dangers that we could solve by getting rid of it. What say you all?

 

[stevethesparks]

When I started training some 45 years ago I often came across installations that had no earth whatsoever. Just L - N, it struck me that was a better system as if there was a break in the N the electrics wont work in the installation and there would be none of the complications as above. Adding an earth has brought its own problems I think it is a redundancy. Thinking outside the box, I say let's do away with earth all together and avoid all the scenarios above and solve the problem in a single stroke. It is ironic that Earth was introduced as an extra safety system and has become one of the most confusing and misunderstood systems, which brings a host of problems/dangers that we could solve by getting rid of it. What say you all?

I liked your post as have thought over similar myself in the past. However, such a system would produce its own problems with faults to earth, particularly where several faults may be apparent in different locations where 3 phase is concerned. At least being tied to neutral, only presents a problem of 230Vac to earth (250Vac in reality).
Perhaps DP switching on all protection devices, and with pme systems the DNO providing a switching device that disconnects the supply and pme in fault conditions.
I'm sure this will get some laughs, but what say you?

 

[Vortigern]

would produce its own problems with faults to earth

How would Earth fault occur, in my scenario there is no earth? As to disconnection, yes that is what it is all about in fault conditions, ADS all the way!

 

[stevethesparks]

Where metal pipes arise from true earth, along with metal structures etc, there will always be an earth - even if not directly connected or referenced to the centre tap. It will always act as a parallel conductor waiting for a fault to give it potential.
Edit: Without that reference, it would be very difficult to monitor such a leakage fault.

 

[davesparks]

How would Earth fault occur, in my scenario there is no earth? As to disconnection, yes that is what it is all about in fault conditions, ADS all the way!

Faults in seperated/ earth free/ isolated systems or whatever you want to call them are a nightmare to find due to the lack of an earth reference.

If a fault occurs making the metal case of a piece of equipment live then it will be perfectly safe to touch because it is not referenced to earth so you cannot get a shock. But it also is very hard to detect and protect against, it is likely to go unnoticed until a second fault occurs causing either a very loud bang, or another piece of metal within touching distance to become live at a different voltage and someone receives a fatal shock.

It becomes very difficult to locate faults without the common reference of earth to refer to.

Any degradation of insulation can bring exposed metalwork up to the line voltage.

Plus of course all of the equipment which relies on the earth for other things like screening, leakage current from filters, RCBO's, surge protection etc

We are all familiar with the effects of capacitively coupled voltages/ghost voltages on unconnected and unearthed conductors. Those kind of voltages will be appearing on alletallic containment if it isn't earthed.

 

[Lucien Nunes]

Expanding on the above, there would be major problems with an earth-free public supply. All works fine until one customer has a fault to earth. They won't necessarily know about it or fix it, because it doesn't stop their system working and no breakers trip, but the system is no longer earth-free. So the next customer using the same supply doesn't know whether there's no earth, his phase is earthed, the neutral is earthed or even some other phase is earthed which would mean his 'line' was 400V above earth and his 'neutral' 230V above earth.

TN systems using an earthed neutral solve this problem by forcing one wire, the neutral, always to be near earth. A line-earth fault can be hazardous, but when it occurs in the first customer's installation, it is rapidly disconnected by his breakers, so it doesn't persist unnoticed or cause a hazard for another customer. A neutral-earth fault could persist, but it doesn't upset the supply for the next customer and is low-risk in itself.

IT (insulated from earth) supplies are OK where the likelihood of multiple faults is negligible, e.g. when supplying only one load, and where one operator has sight of all the possible locations of a fault. Where multiple loads are to be powered from one supply that has no solidly-earthed wire, such as factory DC systems, it is normal to use high-impedance earth fault monitoring equipment, so that the first fault can be found and fixed before another one occurs. But that won't work on a public supply because the DNO can't go round switching everyone off to find the culprit every time anyone has a leaky oven element or a wet plug.

The reason for DP fusing that we still see in old installations, is that once upon a time the neutral was not guaranteed to be solidly earthed. It could be far enough from earth that an N-E fault would send enough current through a neutral cable within an installation to damage it, therefore the neutral cables had to be fused too.
[automerge]1597577966[/automerge]
To clarify this last point, imagine that customer 1 has a neutral-earth fault on a 6A circuit, which is symptomless and persists. Then customer 2 has a line-earth fault on a 100A circuit. If the neutral is not solidly earthed by the DNO, then all the current required to blow customer 2's 100A fuse has to pass through customer 1's 6A circuit, probably setting it on fire. Hence the fused neutrals.

 

[Vortigern]

@Lucien Nunes reading your statements above confuses me. You are talking about a "fault to Earth" and a "Line-Earth" but if you had no Earth none of those events could occur. If a live worked its way loose from its terminal then the item would stop working so something would be done about it. I am not truly advocating dispensing with Earth just flagging up the contrast of such a system. And I am really only talking about it in the context this has been raised in this thread to whit; PEN disconnection and earth within an installation which I take to be domestic. The public supply side of things are slightly different. Manufacturers could make all appliances double insulated removing the danger of a live to metallic casing. I take on board @davesparks points, given that I am talking about domestic installations. I can see it would present danger in other forms with no Earth, yet with the confusion around Earth and bonding that presents probably as much danger in it's own right, maybe?

 

[pc1966]

@Lucien Nunes reading your statements above confuses me. You are talking about a "fault to Earth" and a "Line-Earth" but if you had no Earth none of those events could occur.

But they will. Exterior cable damage, faults to plumbing that is extraneous, etc.

IT supplies have the place, but usually it is the likes of ships and hospital theatre supplies when you need to keep working after a single fault. They have insulation monitoring systems in place and maintenance regimes that will act to fix the 1st fault when it is safe to do so.

That simply would not work on an large public network.

 

[Lucien Nunes]

if you had no Earth none of those events could occur

To ensure there were no faults to earth you would have to make the planet non-conductive, prevent any rain falling and ban all metal. As it is, as soon as one customer has a fault to something extraneous, it could be a screw through a cable that is in contact with building steelwork, or a central heating pump that has an insulation failure to the casing which is attached to an extraneous pipe, or just some water in an outside light, presto, there's an earth reference that the supply cables then take to every other customer on the system.

There's another issue that we haven't mentioned yet, which is a problem even without a fault, and that is normal earth leakage through stray capacitance. All conductors have stray capacitance to the mass of earth, and if energised with AC can pass a small current to (or from) earth through that capacitance. Once you get a large system (a whole street, for example) connected together, even if all the appliances are Class II, and there are no faults, the capacitive reactance from each wire to earth will tend to form a potential divider that holds earth at the midpoint due to symmetry. So you will still have 230V to earth from each line, through some unspecified impedance which, once you factor in all the leakage from SMPSUs etc, will source enough current to deliver a hazardous shock.

 

[davesparks]

@Lucien Nunes reading your statements above confuses me. You are talking about a "fault to Earth" and a "Line-Earth" but if you had no Earth none of those events could occur.

Yes they can, and they will, a low insulation resistance of anything which is also in contact with the ground, maybe an external light such as a lamppost, bollard or those fashionable flush floor mounted lights, is sufficient to start making an earth reference for one conductor.

 

[LastManOnline]

What additional steps might they take ?
It can't be monitored for at the substation - all they'll see there is a change in loads that probably is insignificant relative to the total load.
Monitoring the PEN-true earth current might give a clue - but ONLY if consumers downstream of the PEN break do have effective earth rods so as to be able to create a significant earth leakage current. And this would need careful management so as to avoid nuisance tripping. It would also be negated (or at least significantly reduced) in the presence of multiple PME earth points - since I learned a bit more about it, I've started noticing PME earth conductors down the side of poles.
Realistically the only place it can be monitored for is at the last property on a circuit or the last tap off in the street. That's a lot of points to monitor, and would be prohibitively expensive.
As an aside, some years ago at a previous job, our local DNO (forget what they were called back then) asked if we'd host a monitoring box - it plugged into the mains and a phone line, and would phone home if the power went off. Presumably they'd investigate if they got several calls - one call might just be an internal issue in the property.


Indeed, one of the benefits (cost savings) used in the business case for smart meters was the ability to monitor supply voltage at many points - and thus improve network management and fault responses. Unfortunately, it seems that the people who created the spec for the meters didn't talk to the people who would be using the measurements - and the result is that the voltage measuring ability of the meters is not sufficiently accurate (or it might not even be mandated, can't remember now) for the required purpose. AIUI, last time the business case was re-evaluated, this benefit was still being counted

How can they NOT have an effect ?
If we exclude the "very large resistance so ineffective" issue, because that's much the same as "no earth rod" really.

So lets start from the situation with no earth rods. If a PEN break occurs, the neutral and MET in each downstream property will take up a (dynamic) voltage (relative to earth) which is dependent on the loads attached to each phase.
Basically, the phase currents must sum to zero as there's no neutral to carry the imbalance.
If we start with the simple case of all resistive loads, then the current in each phase is proportional to the local P-N voltage - a phase with a larger set of connected loads will see a reduced P-N voltage until the load current reduces (and the other phase currents increase) such that the phase currents sum to zero. The extreme situation is a load on one phase, and nothing on the other two - the loaded phase will see zero P-N voltage while the other two phases will see 415V. At the other extreme, the loads are all equal, so the neutral will be roughly at earth potential and all properties will see a normal 240V supply.
Add in non-linear loads like SMPSUs and it gets quite complicated. As the P-N voltage reduces, the load current will increase - so in a network with all such "negative impedance" loads, the tendency would be for the neutral voltage to shift rapidly towards one of the phase voltages until the attached loads shut down due to undervoltage.

Now lets introduce one or more earth rods. Now downstream of the PEN break there is a N-E connection, and thus a connection via the earth to the DNOs earthed neutral system. This will be a complex path involving the substation earthing arrangements, any PME earth points in the network, and any customer earth rods upstream of the PEN break.
We no longer have the requirement that the phase currents sum to zero as there's a (relatively high resistance) path for the neutral current. So assuming we have imbalanced loads, the phase with the highest load will still see the lowest P-N voltage. But the effect will be less due to some of the imbalance returning to the substation neutral via the earth.
Lets take the example values mentioned above - 10 houses and a 20ohm earth rod each. Excluding the resistance of the connecting cables, that's a 2ohm connection to earth. Lets say the phases are imbalanced by 10A, that 10A will try and return via the 2ohm earth, raising the local neutral to around 20V relative to earth. So one phase will now be at something like 220V and the others will be correspondingly increased (by a bit less than 20V - you need to draw the phasor diagram).

So adding the local earth rods has turned a "will blow up lots of stuff on 2 phases and maybe set houses on fire" event into a "incandescent bulbs burn brighter and pop but otherwise nothing newsworthy" event.

Clearly the actual effect seen in any event is going to depend on the combined (parallel) earth rod resistances, and the nature of the loads attached at any point in time. And loads will be dynamic - someone putting the kettle on will change the situation significantly.

OK. Firstly the values you use of 20 ohms per rod won't apply in TNC-S countries such as ROI, NZ, Aus as we are required to sink a 1.2 Mt rod only which in good conditions will usually measure approx 100 ohms and in poor conditions 250 to 300 (and upwards). There is no requirement to measure its resistance. So a more realistic value for the 10 homes would be 10 ohms upwards. So what does that 10 ohms achieve under broken neutral circumstances?
Yes house lights will glow. But for all those houses in mid afternoon with virtually no load connected (fridge) you will likely find the touch voltage at the rod is well over 50 volts? Now what if the neutral breaks at house number 5, well the touch voltage doubles again.
Frankly it's a lottery. As, alluded to in an earlier post (Pete 1966) its a problem caused by the DNO and really should be rectified by them.
Lastly (and my apologies again for my confusing "upstream" with "downstream") the rods installed upstream of the break are of negligable value to the installations downstream of the break. Bottom line in my view is if the earthing is, nt done properly a, safety feature becomes a, potential hazard. (like everything else in our profession)
[automerge]1597609741[/automerge]

When I started training some 45 years ago I often came across installations that had no earth whatsoever. Just L - N, it struck me that was a better system as if there was a break in the N the electrics wont work in the installation and there would be none of the complications as above. Adding an earth has brought its own problems I think it is a redundancy. Thinking outside the box, I say let's do away with earth all together and avoid all the scenarios above and solve the problem in a single stroke. It is ironic that Earth was introduced as an extra safety system and has become one of the most confusing and misunderstood systems, which brings a host of problems/dangers that we could solve by getting rid of it. What say you all?

Well now. That's a thought that has occurred in many a sparks mind I would say. It certainly has in mine. Its used in special locations and still to be found in general use in some countries but its a, system that's fallen out of favour. I dont really have the qualifications to argue one way or the other but the simplicity of the system does appeal.

 

[Simon47]

Monitoring remote branches is the obvious method, and given the rise of EVs and the low likelihood of DNO cables and substations being upgraded to allow ~30% of households to switch on 10kW+ charges in the evening, it would seem the smart approach is to have the EV charges reporting to the local substation in some sort of standardised manner.

That way the charge rate can be adjusted to keep cable/transformer limits OK, and the reported voltages then could be used to detect a PME fault region. Given the EV charges have to deal with the open PME risk somehow, it would seem a good way to cover this aspect for other households.

Nice idea, but we've seen how well the comms is working for all the smart meters ! In theory that is one thing smart meters are supposed to enable, but ...
a) It needs the EV charger (not the socket unit) to have the intelligence and the comms to talk to the meter. Bear in mind that (most of) these home EV charging points are really just a high power socket with some protection devices - there's no smarts in them, they are in the on-board charger.
b) The user to actually pair them. Unless you make it so it won't work without, then most users won't - and what about charging away from home ?
c) The whole "smart" system to reach down to that level - which it doesn't and isn't ever likely to do.
d) It relies on the EV being plugged in to detect the fault.
So the most logical place to put the detection is in a device that's installed in every premise (if TPTB have their way), is powered on all the time, already has a network connection, and will include all the smarts. Except that they missed the boat in meter specs.

Faults in seperated/ earth free/ isolated systems or whatever you want to call them are a nightmare to find due to the lack of an earth reference.

My brother spent a while on an NGO aid ship - with emergency lighting running off a 240V battery bank. A pair of light bulbs indicated if the system was earth free (both dim) or if there was a fault (one bright, the other off).
They knew they had a fault, but had been struggling to find it - I suspect the system hadn't been designed (late 40s, early 50s) with that in mind.
The fault was found when someone reported a stairwell light not working. Turned out to have been miswired due to different people having different ideas how the core colours should be mapped to +, -, and earth in a floating DC system. Dunno why it wasn't picked up when the wiring error was done and the light didn't work.