parallel feeders

[kenman215]

Should be an easy one here... I have a 3000 Amp parallel service to do on a building I'm starting. I've done larger parallel services before, and have always observed making sure that each conductor of each respective phase is the same length as its counterparts. Typically, when I ran underground conduit to supply these services, I'd have my trenches dug extra wide, so I could eat up some of the difference in run length caused by having to stack conduits going into my gear.

For this service, though, the transformer is literally 15 feet from the building, at a 45 degree angle from the main switchgear. I did what I coukd, but I have a four foot difference between the shortest and longest run. That is a ton of slack to eat up in the TC. Anyone know if there is a threshold in the code for when you have to start worrying about inbalance? Or am I just stuck with having to have an ugly cabinet?

 

[GoldDigger]

If the inspector can see an obvious difference in length without any attempt to equalize he probably will not allow it. The NEC does not provide a guide.

 

[ADub]

If the inspector can see an obvious difference in length without any attempt to equalize he probably will not allow it. The NEC does not provide a guide.

Or he might say nothing at all. From the pictures I've seen here and on other sites it doesn't look like everybody's particularly concerned with making them exact.


Sent from my iPad using Tapatalk

 

[kenman215]

I know that a lot of guys want their cabinets to look pretty, and picturing what my TC will look like after losing all of the slack makes me throw up in my mouth a little. I wish someone could explain to me exactly how important making the parallels the exact length truly is. For now, I'll just do what I'm supposed to and make sure to bring breath mints on the day I make it.

 

[GoldDigger]

I know that a lot of guys want their cabinets to look pretty, and picturing what my TC will look like after losing all of the slack makes me throw up in my mouth a little. I wish someone could explain to me exactly how important making the parallels the exact length truly is. For now, I'll just do what I'm supposed to and make sure to bring breath mints on the day I make it.

Basically, you can take the ratio of the difference to the full length and that will give you the ratio of current difference to total current. As you suspected the shorter the wire the tighter the tolersnce.
This assumes that the resistance of the terminations is zero.
If the termination resistances are large compared to the wire resistance then the current ratio will correspond to the total resistance ratio.
If you can come within 5% you are probably OK. But that depends or how close you are to the ampacity of the conductors.

 

[ActionDave]

Anyone know if there is a threshold in the code for when you have to start worrying about inbalance? Or am I just stuck with having to have an ugly cabinet?

The code won't give you any wiggle room on length. There is an exception in 300.3 that allows you to run each phase in its own pipe if your using PVC. It makes it easy to keep each phase the same length and look nice. I don't know if it would work for your install. iwire has a pic that he has posted here more than once. Maybe he'll chime in after the Pats game.

I have never done it. Trim and land the first wire and use it for a measuring stick to know how much to cut off the rest and just wad up all the extra as neat as you can was the way I was taught. My first journeyman kept us out working under headlights more than once trying to make things as neat as possible.

 

[jap]

Should be an easy one here... I have a 3000 Amp parallel service to do on a building I'm starting. I've done larger parallel services before, and have always observed making sure that each conductor of each respective phase is the same length as its counterparts. Typically, when I ran underground conduit to supply these services, I'd have my trenches dug extra wide, so I could eat up some of the difference in run length caused by having to stack conduits going into my gear.

For this service, though, the transformer is literally 15 feet from the building, at a 45 degree angle from the main switchgear. I did what I coukd, but I have a four foot difference between the shortest and longest run. That is a ton of slack to eat up in the TC. Anyone know if there is a threshold in the code for when you have to start worrying about inbalance? Or am I just stuck with having to have an ugly cabinet?

Unless your the one terminating at the Padmount, how do you know that the Power Company is taking as much care as you did?


JAP>

 

[kenman215]

Unless your the one terminating at the Padmount, how do you know that the Power Company is taking as much care as you did?


JAP>

I'm terminating at the pad mount, therin lies the rub...

 

[jap]

Dangit......
Better the ugliness be on that end rather than the other.


JAP>

 

[GoldDigger]

Four feet out of 20 is quite a lot, but try to get an estimate of the connector resistances.

 

[mwm1752]

you can be as neat as possible but the POCO will still do the last cut -- It is always easier to make the lenghts the same If the POCO cuts each end the same lenght. More room in the vault

 

[jusme123]

...although the code requires it, paralleled conductors are rarely , if ever, the same length.

 

[don_resqcapt19]

Four feet out of 20 is quite a lot, but try to get an estimate of the connector resistances.

You can do the calculations just using the lengths. Since the conductors are of the same size and type, the resistance per foot is the same.
If you use 8 sets of 500 kcmil with longest length being 20' reducing them 6" per run, so that the shortest is 16.5 long and you have a 3000 amp load the three shortest runs will be operating above 380 amps.
The currents for each, assuming the 6" step change starting with the longest (20') would be:
340.8, 349.6, 358.8, 368.5, 378.7, 389.5, 401, and 413.1.

 

[jap]

You can do the calculations just using the lengths. Since the conductors are of the same size and type, the resistance per foot is the same.
If you use 8 sets of 500 kcmil with longest length being 20' reducing them 6" per run, so that the shortest is 16.5 long and you have a 3000 amp load the three shortest runs will be operating above 380 amps.
The currents for each, assuming the 6" step change starting with the longest (20') would be:
340.8, 349.6, 358.8, 368.5, 378.7, 389.5, 401, and 413.1.

But that's considering that you have zero resistance at the connections.


JAP>

 

[jap]

which is the irony that Golddigger was bringing up,, or the way I read it at least,,,, maybe not.

JAP>

 

[ron]

Four feet out of 20 is quite a lot, but try to get an estimate of the connector resistances.

In my experience, contact resistance is usually negligible if made up and torqued properly.

 

[Carultch]

In my experience, contact resistance is usually negligible if made up and torqued properly.

Negligible compared to how many feet of a feeder?

Negligible compared to 500 ft of feeder, that I'll concur.
Negligible compared to 10 ft of feeder, that I'll need to see some data to believe.


Remember, the voltage drop across each conductor within a paralleled group has to be equal per Kirchhoff's voltage loop law. If the resistances are not the same, the currents will adjust within each one, until the corresponding voltage drops are the same.

So if one path is only half as long as the other, and contact resistance is negligible, it could carry twice the current. And that will be 33% more current than you would anticipate, if it were divided uniformly as it should be. Extreme example, yes I know.

 

[jap]

In my experience, contact resistance is usually negligible if made up and torqued properly.

The power company here terminated thier end of a 4000 amp service not long ago and cut off more than one of their 2 hole compression lugs and reterminated before they were satisfied with the resistance readings they were getting.



JAP>

 

[don_resqcapt19]

But that's considering that you have zero resistance at the connections.


JAP>

That is assuming that the resistance at the connections is identical

 

[jap]

That is assuming that the resistance at the connections is identical

thats a better way of putting it,,,,


jap.