3 sp circuits off a tp isolator

[johnboy6083]

Then you have big issues anyway, even on a standard 3ph submain, as the voltage will rise with respect to earth.
as I've stated, I'd definately not install like this, and I would whinge if I found it, however my opinions don't count on a report. It's facts, experience, science and regulations that matter.

i agree you can't guarantee the loads on socket outlets, however we aren't in the realms of data suites ect here, just general socket outlets.

 

[netblindpaul]

Sorry to say Paul but if I understand the OP correctly then you're wrong. Let's say for example that there is a 20A load plugged into each of the three phases (60 total current). The three line conductors are out of phase with each other, therefore the resulting current in the neutral from any one of these lines will be out of phase also. In very simple terms; at the point in which L1 is drawing its 20A, L2 and L3 aren't drawing, therefore the neutral current is 20A, 120 degrees (150th of a second) later L2 is drawing 20A whilst L3 and L1 are drawing nowt. Neutral current = 20A. And so on...

No they are not,
They are 120 degrees out of phase in the line conductors, we have no idea where they are in the N.
You cannot make simple assumptions like this.

 

[netblindpaul]

Thing is johnny, you don't know where the sockets are, or what is going to be plugged into them.

I stand by my arguments and no one has put anything forward to change my mind yet.

You have a single 1x2.5mm sq carrying the return current from 3 x 20A single phase circuits, they are NOT 3 ph at the point of use, if they were then I would never have commented.
This is not acceptable, show me a reg that says it is, or give me some REAL science that proves that the N can never be overloaded.

 

[D Skelton]

The maximum neutral current will be equal to the maximum line current on one phase at any one time. Maximum LINE current = 20A then maximum neutral current = 20A

 

[D Skelton]

If you didn't have blinkers on you'd understand mate.

 

[netblindpaul]

Can you please explain that in fundamental terms on 3 separate single phase circuits please?
Remember these circuits ONLY share an mcb they are otherwise not linked on the phase side, only the N side, so each can take 20A, yes, if they are perfectly balanced then you "could" argue a cancelling affect, though how would this cancelling current affect each individual single phase load?
To cancel the current has to go elsewhere other than down the N conductor, thus it has to flow back down the other two phases, how will this return current affect the electronic single phase loads connected to these single phase socket outlets?

You cannot flow N current from a single phase circuit back down a different single phase circuit to cancel out N current without affecting the characteristics of both separate single phase circuits.

 

[netblindpaul]

If you didn't have blinkers on you'd understand mate.

I don't have blinkers on, you cannot flow N return current back down a separate phase conductor without affecting the separate single phase circuit, we are NOT talking about a 3ph circuit here.
We are discussing 3 single phase circuits with a shared N, the only way the N current can balance is by affecting the other 2 1ph circuits.
This cannot be allowed, each 1ph circuit MUST be independent.

 

[D Skelton]

We are discussing 3 single phase circuits with a shared N

No, we are discussing 3 OUT OF PHASE single phase circuits with a shared N.

Again I say, if there is a 20A load plugged in on each phase, when L1 is drawing 20A, L2 and L3 are drawing NOTHING, therefore the neutral current is 20A.

150th of a second later 20A will be drawn down L2 whilst L3 and L1 draw NOTHING.

At any one time only one phase will be drawing 20A therefore the neutral current can ONLY be 20A.

 

[netblindpaul]

No you area assuming that the N return currents are in phase with the phase currents.

 

[Guest55]

i agree with pauls & jonnys sentiments that a shared neutral is bad practice / design

if single phase & unbalanced loads ( the sockets ) are connected to a 400v supply then i'd expect seperate neutrals for each circuit back to the origin.
whether the devices are 1 or 3 pole is neither here nor there in my mind , it would be a C3 defect if i was inspecting it.

 

[D Skelton]

Because they will be, they are single phase circuits!

 

[netblindpaul]

OK,
I can see where those of you who say that the N current will be delayed by 120 degrees, this would be the case with balanced 3 phase loads of known characteristics.
It cannot be guaranteed to be the case with 3 off single phase socket outlet circuits spread across 3 phases, you have no control over what the loads are or what they an "do" to the N return current, you cannot assume that the N return current will be phase balanced.
There are too many non-linear loads in current installs.
If the loads are perfectly balanced then why do we not use 1/2 size N conductors for 3ph boards all the time now?

 

[D Skelton]

Forget balancing, it is completely irrelevant to this discussion.

Even still, take a TP+N load, L1 = 100A, L2 = 0A, L3 = 0A then N = 100A.

The neutral current can never exceed the line current!

If you can guarantee a specific Ib to stay at a certain value per phase then it is completely permitted to calculate the size of the neutral conductor required and install that.

 

[netblindpaul]

It's not a TP+N load though!
Where did you get it is?

 

[D Skelton]

It doesn't matter, the L current per phase will be out of phase with the others, and so will the neutral current.

Bad design for sure, but absolutely no danger.

 

[netblindpaul]

Yes it could be out of phase, but by how much, and, why are all 3 phases bound to be loaded at the same time and by the same amount and with a suitable phase shift to ensure that the N cannot be overloaded, you cannot guarantee that they will, you can get sufficient phase shift to overload the N, remember there are so many non-linear loads around now and the OP does not define what is connected.
There are too many unknowns to give this an OK.

 

[D Skelton]

You're not going to get any plug in non-linear loads without suitable PF correction fitted within. Even still, any negative effects caused by slight inductance or capacitance will be negligable.

 

[netblindpaul]

So you are categorically stating that if 2 off 20A single phase loads were plugged into 2 separate single phase circuits they could NEVER overload the 2.5mm sq N conductor?

 

[johnboy6083]

The calc I posted proved it. Harmonics notwithstanding!

 

[netblindpaul]

So you can't then, end of argument.

 

[Rockingit]

You're not going to get any plug in non-linear loads without suitable PF correction fitted within. Even still, any negative effects caused by slight inductance or capacitance will be negligable.

I can understand where Paul is coming from, with a purely theoretical view, but as a practical application I have to go with this ^^^ ..... We're talking about some sockets in a dado trunk, not industrial connections.

 

[netblindpaul]

Thing is you don't know what is going to be plugged in.

 

[D Skelton]

So you are categorically stating that if 2 off 20A single phase loads were plugged into 2 separate single phase circuits they could NEVER overload the 2.5mm sq N conductor?

No, I'm just saying it's HIGHLY improbable.

Think of TP+N lighting power track systems. These share a common neutral and are protected by TP OCPDs. Commonly these are also fluorescent fittings, an assumption can be made that a standard rate of cap faliure within the fittings will equal itself out over time across the three phases thus causing very little cause for concern. Still, installations like these need regular maintenance where things like this SHOULD be picked up if the guys are doing their job properly.

Like I said, the circuit brought to discussion in the OP is badly designed with it's use in mind, however it would cause me no concern if I came across the same thing on an install.

 

[Rockingit]

Thing is you don't know what is going to be plugged in.

Not sure what I'll be having for breakfast, either! You can over-think things sometimes.

 

[netblindpaul]

Stuff lighting!

Have you ever analysed the non-linearity of power loads in even a simple domestic install these days?

Think SMPSU's, your fluoro fittings, motor loads, vsd's in washing machines, yes, there will be pfc, & each should individually meet the requirements of the LVD & the EMCD, but cumulatively this can have a significant affect on the connected systems, so i stand by my points that the loads can sum as they can be way out of phase.
Take an accurate power analyser and check a few installs and see what you get at the tails, then guarantee me that you can never overload the N in this scenario.

 

[44ALLAN]

Fantastic thread guys.............well done to all.

Great stuff

 

[D Skelton]

You're definitely over thinking things now!

We were talking about three 20A circuits, not a local distribution network lol!

 

[netblindpaul]

What local distribution network, all of these things can be found on a small domestic install these days, I'm not going to back down and no one has yet given proof that the N cannot be overloaded in the scenario given by the OP.
So until this is done I stand by my arguments, I DESIGN, I DO NOT GUESS values, so the N can be overloaded, & if I was being paid for my consultancy then I would be bothered to calculate things, as I am not, it;s down to you lot to prove otherwise, which, as yet, you have failed to do.

 

[tom79bmth]

Well how come the supply distributors don't have a bigger neutral than the phases,

 

[netblindpaul]

Because it is not needed in that scenario, because it is not what is being discussed here, this is a very different situation.

 

[Guest55]

is it over ?
are we done yet ?
any more calcs to throw in ?

phew!
;-D

 

[lewlec]

Ok I may have missed some thing here, say you have a tp&n supply connected to a switch fuse with say a 4 core swa feeding a tp&n db and all loads on that db are single phase, would that be incorrect? Would you oversize your neutral?

 

[netblindpaul]

No it's not a tp+n supply & it's not to a db, it 3 off final circuits fed from the same 3p mcb, with a common N to all circuits, and the cct's are socket outlet circuits.

 

[Guest55]

*facepalm*

 

[D Skelton]

What local distribution network, all of these things can be found on a small domestic install these days, I'm not going to back down and no one has yet given proof that the N cannot be overloaded in the scenario given by the OP.
So until this is done I stand by my arguments, I DESIGN, I DO NOT GUESS values, so the N can be overloaded, & if I was being paid for my consultancy then I would be bothered to calculate things, as I am not, it;s down to you lot to prove otherwise, which, as yet, you have failed to do.

I never said never, I just said given the OP, it was highly unlikely.

On another note, PF is not cumulative in the sense that if a system is made up of loads at a PF of .9, the overall PF is still .9. The average domestic install is expected to average out at roughly .9 - .95, three of those domestic installs across three phases and the slightly negative effects caused by the non-linear loads within each dwelling become meaningless. The exact reason for this can be found within our discussion regarding our three 20A socket circuits.

We all suffer from a lack of understanding at times (myself included! ) you have since seen sense but still seem to be fixated on being theoretically correct. It seems to me that you're just not looking at the bigger picture.

 

[netblindpaul]

I am not talking about PF, I am mainly talking about distortion of the supply, I have never said anything about PF.
Supply distortion these days has a much bigger affect on installs than PF as this is normally taken care of, but, distortion IS cumulative.

 

[Silly Sausage]

Fantastic thread guys.............well done to all.

Great stuff

Hahaha...must have a proper read tomorrow!
The original Q just appears to be an unbalanced Star with a Neutral, what's the problem?!?! :sifone::smilielol5:

 

[D Skelton]

Is it? On a large scale you would not be wrong to expect that distortion would equal itself out across the phases.

 

[44ALLAN]

Hahaha...must have a proper read tomorrow!
The original Q just appears to be an unbalanced Star with a Neutral, what's the problem?!?! :sifone::smilielol5:

If you sort it out let me know mate.................I've agreed with at least 10 differing versions up till now lol

 

[netblindpaul]

TO be honest, whist there seems to be a lack of belief from posters here with regard to what I am saying, there seems to be a significant lack of understanding of what I am trying to put across as well, I don;t know how else to describe things.

Thing is see, I'm not a spark, I operate mostly as an Engineer, & consultant thus I am paid for my opinions, & I don't start throwing arguments about unless I can back them up, because that is how I make my living, and I'm doing OK, thank you and have been for a few years now, based solely on my own personal opinion, which is hen backed up by fact once the money starts flowing, why should I do otherwise when it is my living.
I'm not proposing anything as stupid as payment here, I'm here for the fun of it. but, I will never post anything that I cannot back up, and no one has managed to prove with science that my arguments are incorrect yet, once they do I will give an apology, until then I stand by them.

In the OP, the N conductor CAN be overloaded.
The circuit is not acceptable.
If anyone can PROVE otherwise in all scenarios then I will bow down, until then I await proof.

 

[johnboy6083]

I actually enjoy decent threads like this, where we all have different takes, but with no ill manners to be seen.

 

[netblindpaul]

Well I thought I got pretty ****ty at one point so I WILL apologise for that, but, I stand by my engineering arguments even if my attitude stunk!
If it did, sorry.

 

[D Skelton]

In the OP, the N conductor CAN be overloaded.
The circuit is not acceptable.
If anyone can PROVE otherwise in all scenarios then I will bow down, until then I await proof.

And pigs CAN fly, albeit thrown out of a hot air baloon strapped to a hang glider!

You are theroetically correct, just not realistic.

 

[johnboy6083]

I don't doubt your conviction Paul, but in my mind, I can't see the theoretical difference between the OPs scenario, and the scenario of a feeder feeding houses on different phases in the street.

 

[Silly Sausage]

I think it depends on how it's wired (to state the obvious!)
Trying some sketches as we speak.
Can't see the Neutral carrying 60A though...

 

[44ALLAN]

Set the circuit up and clamp it mate then tell us what the neutral was taking...........I'll trust you.

3 single phase circuits feeding 3 heaters(all on different phases.....or should that be lines !) through same mcb with one neutral for all 3 circuits.

 

[netblindpaul]

Ahh, now johnny, there you have a good point, they don't comply with BS7671 as a start! #
They don't have to!
Also, the scenario is again different, because they will approximate phase balance across their feeder, and, the N conductor at each house will be protected load end by the fuse in the head, & the N conductor in the street will be either earthed "PME" or protected supply end by a suitable OCPD, which will prevent unbalanced currents overloading the N.
In the OP this is not the situation.

 

[netblindpaul]

OK maybe not 60A in all scenario's but it could take that depending on the loads...

Also no one has come back with a calculation that can prove that the N cannot be overloaded regardless of the connected loads yet.

 

[johnboy6083]

The Sqrt(((A^2)+(B^2)+(C^2))-((AB)+(AC)+(BC))) should prove that. I'm fully aware that not all loads are linear, and socket outlet circuits will always be prone to this, whatever configuration they are installed in.

You are absolutely correct in saying that we cannot guarantee the nature of the load, but this is true for most installs nowadays.


This is an interesting debate though.

 

[netblindpaul]

Sorry johnny, not convinced!

Also the DNO argument is not along the same lines.

You have to look at the origin over current devices, and the circuit conductor sizes in relation to the prospective loads...