Circuit Wire For Long Run

[Cookie]

Ok, say you've got a lighting run with the last fixture being 1,280 meters away. 1,800 watts per phase 3 phase, 277/480Y volts though in this example we can use 230/400Y for the ease of discussion.

My question is- what size wire do I run for each phase and earth? What disconnection time do I use? Do I use a Type B, C or D breaker? What should my max R1+R2 come out to be?

 

[telectrix]

if digging in for 1 lamp post, why not fix 2 and stickk a propeller on the other one. wind power.

 

[pc1966]

At 50/60Hz??

For large conductor yes, but its only about 20%-ish for 1000mm^2 so usually can be ignored. But also the inductive term starts to become more significant relative to the DC R1+R2, but even that is only about 10%-ish for 150mm cables.

 

[Cookie]

For large conductor yes, but its only about 20%-ish for 1000mm^2 so usually can be ignored. But also the inductive term starts to become more significant relative to the DC R1+R2, but even that is only about 10%-ish for 150mm cables.

Oh man, you guys have some serious brains!

 

[Cookie]

I used this:

American wire gauge - Wikipedia

en.wikipedia.org

Looking at details it says 68F = 20C which is room temperature around my part of the world.

Also probably they are DC, starts to become a source of error for large conductors as skin effect kicks in.

I should know about this link! ????

Anyway, do you know of a table taking skin effect into account? For a while I've desperatly been searching for an AC resistance at 30*C.

 

[Cookie]

I take this to mean an RCD for disconnection is a no-no?

 

[pc1966]

I should know about this link! ????

Anyway, do you know of a table taking skin effect into account? For a while I've desperatly been searching for an AC resistance at 30*C.

At AC power frequencies the skin depth in copper is about 9mm, so any conductor that is less than around 20mm diameter it has quite a small effect. Somewhere there will be equations to allow its computation for different conductors but most folks in engineering take the easy route of getting a feel for when you can ignore something, and ignoring it.

For example it is unlikely you will know any AC system parameter to better than 5% uncertainty, let alone 1%, so once effects get down to the 5%-ish region they can be often ignored for most purposes.

The temperature coefficient for copper is around 0.00393 so each 10C increase in temperature has a 3.93% increase in resistance. Which is about 4%. Which is why the UK regs have a 0.8 factor for going from measured at 20C to Zs value working at 70C.

Yes, our regs do have detailed formulae for non-standard cases such as measurement very cold/hot and/or operation at unusually high temperatures, but that vast majority of industrial cases are just like commercial and domestic in the 20C test & 70C assumed max operation is perfectly applicable.

In terms of generic AC values we have tables in the regulations with the values to assume. For example this if for single cables in various configurations giving both the DC and AC "resistance" values (more correctly the voltage drop due to cable impedance):

 

[Soi disant]

For large conductor yes, but its only about 20%-ish for 1000mm^2

Just a tad bigger than I'm used to thinking about.... ?

 

[pc1966]

I take this to mean an RCD for disconnection is a no-no?

No, the UK regulations generally do not specify how you achieve disconnection, only the times, etc, needed for safety. But they do list approved device standards for doing so (e.g. BS / EN ones on fuses, MCB, RCD, etc).

That section you quoted (411.3.2.5) has been updated but really it was a clause to deal with cases where the usual fault disconnection (OCPD or RCD) are not feasible so other means of protecting against shock need to be considered (supplementary bonding, etc).

The note about "disconnection may be required for reasons other than protection against electric shock" is (I think) to point out you might need it for overload/fire protection anyway, even if that fails to meet the maximum disconnection times discussed in that section about shock protection.

But beyond that, there are other rules permitting no OCPD or special cases when disconnection might be more dangerous, e.g. support electromagnet in factory, etc.

 

[pc1966]

In terms of generic AC values we have tables in the regulations with the values to assume. For example this if for single cables in various configurations giving both the DC and AC "resistance" values (more correctly the voltage drop due to cable impedance):

If you look closely at that table you can see where the skin effect comes in by comparing the DC volt drop with the resistive term 'r' in column 5, but more generally it is the total impedance 'z' that would be used for volt drop if dominated by the cable.

Though close to a transformer, etc, where reactance is also a big factor you would be separately adding the r & x terms and finally computing z = sqrt(r^2 + x^2) to get your impedance.

 

[Soi disant]

At AC power frequencies the skin depth in copper is about 9mm, so any conductor that is less than around 20mm diameter it has quite a small effect.

 

[Cookie]

No, the UK regulations generally do not specify how you achieve disconnection, only the times, etc, needed for safety. But they do list approved device standards for doing so (e.g. BS / EN ones on fuses, MCB, RCD, etc).

That section you quoted (411.3.2.5) has been updated but really it was a clause to deal with cases where the usual fault disconnection (OCPD or RCD) are not feasible so other means of protecting against shock need to be considered (supplementary bonding, etc).

The note about "disconnection may be required for reasons other than protection against electric shock" is (I think) to point out you might need it for overload/fire protection anyway, even if that fails to meet the maximum disconnection times discussed in that section about shock protection.

But beyond that, there are other rules permitting no OCPD or special cases when disconnection might be more dangerous, e.g. support electromagnet in factory, etc.

Well, even in an IT system you still need to open device under a multi phase fault.

Thank you for explaining this to me, but I still think disconnection should not be accomplished via RCD, as RCDs can fail.

 

[Cookie]

At AC power frequencies the skin depth in copper is about 9mm, so any conductor that is less than around 20mm diameter it has quite a small effect. Somewhere there will be equations to allow its computation for different conductors but most folks in engineering take the easy route of getting a feel for when you can ignore something, and ignoring it.

For example it is unlikely you will know any AC system parameter to better than 5% uncertainty, let alone 1%, so once effects get down to the 5%-ish region they can be often ignored for most purposes.

The temperature coefficient for copper is around 0.00393 so each 10C increase in temperature has a 3.93% increase in resistance. Which is about 4%. Which is why the UK regs have a 0.8 factor for going from measured at 20C to Zs value working at 70C.

Yes, our regs do have detailed formulae for non-standard cases such as measurement very cold/hot and/or operation at unusually high temperatures, but that vast majority of industrial cases are just like commercial and domestic in the 20C test & 70C assumed max operation is perfectly applicable.

In terms of generic AC values we have tables in the regulations with the values to assume. For example this if for single cables in various configurations giving both the DC and AC "resistance" values (more correctly the voltage drop due to cable impedance):
View attachment 84241

Alright, because in all honesty I'm trying to get these values down to 30*C:

 

[pc1966]

Just a tad bigger than I'm used to thinking about.... ?

To be fair it is a lot bigger than I ever have to deal with.

Also as you get to such large conductors there are good reasons to consider parallel sets instead. Not only are they a bit easier to wrangle, they have a higher current rating for the same overall CSA as cooling is usually better. However, you do have to consider cycling the sets to keep current flow, etc, balanced and you also need to look carefully about the OCPD implications for a fault half-way along one of the set, etc.

 

[pc1966]

Well, even in an IT system you still need to open device under a multi phase fault.

Thank you for explaining this to me, but I still think disconnection should not be accomplished via RCD, as RCDs can fail.

As I said, I would consider RCD as my 2nd choice unless there is no reasonable alternative. So for a TT set-up you have no reasonable alternative so it is a RCD.

Also you can take steps to improve the reliability:

• Having two RCD in series is on obvious case (so domestic TT example with a 100mA delay incomer and 30mA per circuit RCBOs).
• Fitting SPD to reduce the chance of RCD electronic damage due to lightning surges, etc.
• Having and enforcing a 6 month periodic trip-test is another good idea.

 

[pc1966]

Alright, because in all honesty I'm trying to get these values down to 30*C:

If it is 75-30 = 45C colder then it would be 17.7% lower for the resistive part.

For big cables with X dominating Z the overall difference is going to be less, as that won't change much with temperature.