Parallel feeder fault and the result if there is difference in conductor lengths

[JFletcher]

Not taking sides here however a detailed mathematical analysis of current flow is appreciated, it goes to the heart or intent of the code, if not the word of it.

As noted, conductors are rarely of exactly the same length... I would go so far as to say they are never exactly the same length in parallel installations. Also already as noted, there will be slight differences in current flow upon parallel conductors, which are more pronounced as the run gets shorter. To wit: a 30-foot run with two left or right hand sweeps in 4 inch conduit, with the conduits laid out in a 4 by 1 configuration, the inside conductors could be as much as 12ft shorter than those on the outside. This is a pretty extreme example, however it is something to think about when laying out conduit and using parallel conductors

 

[Besoeker]

Let's end the comments there please...

Finegood..
I'm sure you and others understand...

 

[electrofelon]

Every time I have measured, i get current differences in parallel conductors that I had installed and know they were pretty much the exact same length. Not as severe as the OP, but not insignificant - maybe like +/- 5-8 amps centered around 100. I suspect termination impedence or inductive effects are the cause? Or maybe meter inaccuracy?

 

[don_resqcapt19]

Every time I have measured, i get current differences in parallel conductors that I had installed and know they were pretty much the exact same length. Not as severe as the OP, but not insignificant - maybe like +/- 5-8 amps centered around 100. I suspect termination impedence or inductive effects are the cause? Or maybe meter inaccuracy?

Even the physical arrangement, proximity to the other circuit conductors, has an effect on the impedance of the conductors and the current division among the conductors of each phase.

 

[jim dungar]

Even the physical arrangement, proximity to the other circuit conductors, has an effect on the impedance of the conductors and the current division among the conductors of each phase.

Which may be partly why the NEC does not provide any additional allowance. Control what you can, so the out of control is somewhat tolerable.
I believe if we allow 10% difference, we will eventually have the discussion of why not 11%

 

[winnie]

On this issue the NEC is actually very liberal.

When I've looked at things like ampacity tables for bus bars and the like, parallel sets of bars are only credited for a portion of the multiple, eg. https://stormpowercomponents.com/resources/ampacity-charts/maxiflex where the ampacity of 2 bars in parallel is 1.72x that of a single bar, and 3 in parallel is 2.25x that of a single bar.

I don't believe that this is because of magnetic interactions between the components or mutual heating, but rather an allowance for uneven current distribution.

-Jon

 

[Phil Corso]

Wyo721...

Here’s a rather simple numerical solution:

Ia = Current in cable A.
Ib = Current in cable B.
It = Ia + Ib.
Za = Impedance of cable A.
Zb = Impedance or cable B.
Zt = Za + Zb

Then the current distribution is: Ia = It x (Zb/Zt), and Ib = It x (Za/Zt).

Further simplification… let Za = Z, and Zb = Z+ΔZ, where ΔZ = length difference. Then,

Ia = It x (Z+ΔZ)/(2xZ+ΔZ) and Ib = It x (Z/(2xZ+ΔZ)

If you want additional detail, for example using cable parameters found in NEC, Table 9, let me know!

Regards, Phil Corso

 

[ggunn]

Which may be partly why the NEC does not provide any additional allowance. Control what you can, so the out of control is somewhat tolerable.
I believe if we allow 10% difference, we will eventually have the discussion of why not 11%

Agreed. The NEC says to make them the same length. Do that to the best of your ability and move on.

 

[LarryFine]

It seems that the difference can be an issue with conductors that are loaded at or near their ampacity, but less so if they are not. It all depends on calculations and final selected sizing.

 

[Ingenieur]

Finegood..
I'm sure you and others understand...

finegood
I'm sure they do

btw, 'what' do they understand? lol

 

[Ingenieur]

Not taking sides here however a detailed mathematical analysis of current flow is appreciated, it goes to the heart or intent of the code, if not the word of it.
....

without an explanation/derivation/mathematical example, an 'opinion' on electrical engineering is not worth much
don't tell me, show me

 

[drcampbell]

The two conductors are connected at their ends, creating two resistors connected in parallel. Because the two ends of each wire are the same terminals, the voltage drop is the same for each conductor.

The current in each conductor is inversely proportional to its impedance. If one conductor has 50% more impedance, (as in the rather-extreme example above) the other conductor carries 50% more current. (60% vs. 40% of the total) This is true even if the conductors are not the same size or material.

If the amount of imbalance is less than the margin between the wire's ampacity and the actual current it's carrying, there's absolutely no cause for concern.
(and since published ampacity includes a generous margin of safety, there will need to be a lot of imbalance before an actual problem arises)

The same concept applies for three or more parallel conductors. It's still just resistors in parallel, only with more variables in the equation(s).

In the real world, the quality of the connections is likely to be more influential than the impedance of the wire itself, (except for very long runs) which is probably the rationale for assigning less than 200% ampacity to two wires in parallel and less than 300% ampacity to three wires in parallel.

 

[Ingenieur]

The two conductors are connected at their ends, creating two resistors connected in parallel. Because the two ends of each wire are the same terminals, the voltage drop is the same for each conductor.

The current in each conductor is inversely proportional to its impedance. If one conductor has 50% more impedance, (as in the rather-extreme example above) the other conductor carries 50% more current. (60% vs. 40% of the total) This is true even if the conductors are not the same size or material.

If the amount of imbalance is less than the margin between the wire's ampacity and the actual current it's carrying, there's absolutely no cause for concern.
(and since published ampacity includes a generous margin of safety, there will need to be a lot of imbalance before an actual problem arises)

The same concept applies for three or more parallel conductors. It's still just resistors in parallel, only with more variables in the equation(s).

In the real world, the quality of the connections is likely to be more influential than the impedance of the wire itself, (except for very long runs) which is probably the rationale for assigning less than 200% ampacity to two wires in parallel and less than 300% ampacity to three wires in parallel.

imo the v drop across wires of different length are not =
v drop = L x R puL x i
we know the i in each leg is = (kcl), and R puL is the same, but if L is different for each leg v drop must by different in each leg
the sum (leg v drop) = source voltage (kvl), they are in series, not parallel

I've done research on connectors and splices
if properly made there are usually in the uOhm range, so even a few dozen adds very little

 

[drcampbell]

imo the v drop across wires of different length are not = ...

Go ahead, 'splain how two parallel conductors can originate & terminate at the same two nodes but have different voltages across them.

 

[Ingenieur]

Go ahead, 'splain how two parallel conductors can originate & terminate at the same two nodes but have different voltages across them.

???

 

[Besoeker]

Go ahead, 'splain how two parallel conductors can originate & terminate at the same two nodes but have different voltages across them.

Good question. I suppose they could be slightly different lengths owing to differences in bend radii.
To the extent that it would cause one of the parallel runs to exceed it's current carrying capacity? Would you intentionally run them that close to their limit?
And who has ever measured a significant difference in the length of the runs?
In the words of that lovely Peggy Lee song "Is That All There Is"

Yet it gets so convoluted here at times.

 

[electrofelon]

Go ahead, 'splain how two parallel conductors can originate & terminate at the same two nodes but have different voltages across them.

Different impedence. There is more to impedence than just resistance.

 

[Ingenieur]

Different impedence. There is more to impedence than just resistance.

I was looking at it as a ph-ph fault with one ph being longer

I guess you could look at it as each ph has >1 conductor and their lengths vary
at that point a current divider
assume 2 cond/ph
ia2= Za1/(Za1 + Za2) x isc
ia1= Za2/(Za1+ Za2) x isc

ia2 = (Za1/(Za1+ Za2)) (ia1/(Za2/(Za1 + Za2))

ia2/ia1 = Za1/Za2
Zax = Lax x Rpu, Rpu same for both
ia2/ia1 = La1/La2

10% difference in Length, shorter will have 10% more isc

but if 3 ph per set and centered on lugs symmetry should mostly equalize lengths

 

[ggunn]

Different impedence. There is more to impedence than just resistance.

I think that what he is driving at is that the conductors are tied together with a low resistance connection at the load end so they must be at virtually the same voltage at that point irrespective of length differential. With voltage thus clamped and resistance fixed by the length of the conductor, the only variable free to move is current, which is, of course, the problem with parallel conductors of different lengths.

 

[topgone]

I think that what he is driving at is that the conductors are tied together with a low resistance connection at the load end so they must be at virtually the same voltage at that point irrespective of length differential. With voltage thus clamped and resistance fixed by the length of the conductor, the only variable free to move is current, which is, of course, the problem with parallel conductors of different lengths.

Very good analysis. A universal truth!:thumbsup: