Connecting portable/temporary equipment with parallel portable cables

[Russs57]

So, I have a portable chiller. Two circuits, each around 450 amps, 3 phase 480. I am running parallel (portable/temporary) cables with cam lock connectors. Two cables per phase to each of the circuits. Only one cable for ground for each circuit. So a total of fourteen 4/0 cables. The cables are supplied with the chiller and an electrical contractor has run the cables. Lets call each conductor one fifty to two hundred feet in length. So I wouldn't say each circuit is the same length as one circuit is perhaps fifty feet longer. However each circuit has all cables being the same exact length. Both circuits are derived from the same circuit breaker.

Now to my question. Does it matter how the cables are laid out? By that I mean, for each circuit, do I want to run two A phases next to each other, followed by two B phases next to each other, followed by two C phases next to each, followed by one ground? All cables are as close to each other as possible and laid on concrete. They are outside in the weather for the last fifty to one hundred feet.

The reason why I ask is I'm not sharing current as equally as I would like. Connections at the supply side (crimped ring terminals bolted to breaker's bus bars with 3/8-16 bolts with belleville washers) have over heated but not outright failed (insulation is melted).

Safe to assume any current reading on the ground conductor is magnetic/inductive coupling and not a cause for concern regardless of magnitude?

I'm purposely not stating degree of current unbalance on phase conductors nor amount of current of ground conductor (all readings were done with clamp on type meter). I'd rather you guys tell me how much is too much.

Thanks in advance.

 

[ActionDave]

Not sharing current as equally as liked where? Between phases or between conductors of the same phase?

For just the motor phase to phase more than ten percent difference is bad, but for the whole chiller who knows.

What makes you think the melting insulation at the connections is because of current imbalance? It could be evidence of a bad connection.

 

[Russs57]

Not sharing equally between same phase conductors on same circuit. Seems strange to have two cables landed on the same bolted connection and one to be fine and the other to have a foot of melted insulation.

 

[Hv&Lv]

So you have two circuits feeding what?
all cables on the same phase are the same length.
the cam lock connectors were cleaned and aren’t hot?
all other connections were cleaned and tightened properly?

a 5% voltage imbalance can cause a 50% temperature rise in motor heating

your voltage imbalance is probably what’s driving your current imbalance so far out of whack on the motors.

check your imbalance.
Measure the three voltages, get the average. Take that # and subtract it from your maximum voltage.
take the difference and divide it by your average.
X 100 for percent.
whats you imbalance?

you can rotate your connections and see if it’s supply or motor causing the voltage imbalance.

 

[ActionDave]

How much is too much is when insulation is burning off. It doesn't matter what the amp meter says.

I'd be unbolting, cutting off and replacing some connectors or entire cables and re torquing everything.

 

[Hv&Lv]

How much is too much is when insulation is burning off. It doesn't matter what the amp meter says.

I'd be unbolting, cutting off and replacing some connectors or entire cables and re torquing everything.

:thumbsup::happyyes:

 

[Jraef]

Throwing this out there;

Assuming this is extra flexible cable, like Type W or DLO, are you aware that you cannot use standard crimp lugs and dies on them? They need specifically designed lugs and special dies for those lugs, because the high strand count changes the way the crimps compress. My guess is that someone didn't know that and used whatever would fit (maybe even trimmed some strands off in the process), so it has created a poor crimp connection and is over heating. It's something I have seen time and time again on high strand count conductors.

 

[synchro]

Throwing this out there;

Assuming this is extra flexible cable, like Type W or DLO, are you aware that you cannot use standard crimp lugs and dies on them? They need specifically designed lugs and special dies for those lugs, because the high strand count changes the way the crimps compress. My guess is that someone didn't know that and used whatever would fit (maybe even trimmed some strands off in the process), so it has created a poor crimp connection and is over heating. It's something I have seen time and time again on high strand count conductors.

Agree, and if this is the case make sure that all lugs are replaced not just ones that are overheating. This is probably self evident but I'm mentioning it so it doesn't get missed.

 

[GoldDigger]

If you have a 200' cable and a 150' cable, the ratio of the resistances will be 200/150. And since the voltages at the connection points must be equal the ratio of currents will be 150/200. To make the numbers nice, lets say that the total current is 350A. Instead of each conductor carrying 175A you will have one carrying 150A and one carrying 200A. That could easily overload the shorter cable and its connectors.
The closer you design the system to just being able to carry the load current with exactly equal current division, the easier it is for a length difference to produce an overload. l
Note also that, as noted in another post, any additional resistance above the design resistance of an interconnection as a result of improper termination will produce a localized temperature increase that adds to the conductor heating caused by unequal current distribution.

 

[winnie]

So, I have a portable chiller. Two circuits, each around 450 amps, 3 phase 480. I am running parallel (portable/temporary) cables with cam lock connectors. Two cables per phase to each of the circuits. Only one cable for ground for each circuit. So a total of fourteen 4/0 cables. The cables are supplied with the chiller and an electrical contractor has run the cables. Lets call each conductor one fifty to two hundred feet in length. So I wouldn't say each circuit is the same length as one circuit is perhaps fifty feet longer. However each circuit has all cables being the same exact length. Both circuits are derived from the same circuit breaker.

As long as the conductors forming a phase of each individual circuit are the same length, then length shouldn't make a difference. Are you sure that they are in fact two separate circuits, and not joined again in the equipment such that you actually have _4_ parallel conductors per phase?

Now to my question. Does it matter how the cables are laid out? By that I mean, for each circuit, do I want to run two A phases next to each other, followed by two B phases next to each other, followed by two C phases next to each, followed by one ground? All cables are as close to each other as possible and laid on concrete. They are outside in the weather for the last fifty to one hundred feet.

Magnetic coupling between the phases will tend to imbalance the current flow. Say you have the conductors arranged A A B B C C; the individual conductors in each phase set will have different 'loop area' with respect to the other conductors in the circuit, which means different inductance and different inductive reactance. It shouldn't make a huge difference, but it would be there. I would not expect the difference to actually cause problems. The 'ideal' arrangement would have each ABC set triplexed together, but I strongly doubt the benefit is worth the hassle.

Safe to assume any current reading on the ground conductor is magnetic/inductive coupling and not a cause for concern regardless of magnitude?

Any such current would require a closed loop path...which you do have with 2 EGCs each run with the different length circuit. If you are seeing significant magnetic coupling to the EGC then this gives you and idea of the sort of magnetic coupling that you might see causing unbalanced currents in your parallel conductors.

-Jon

 

[Russs57]

Thank you Winnie (or is it Jon?). That is the type of information I was seeking. For what it is worth, these are the current readings for the two different circuits. Circuit 2 is about 50 feet longer and both circuits are from the same circuit breaker. Circuit 1 has two screw compressors and all the condenser fans. Circuit 2 has two screw compressor and a centrifugal pump. Readings were done with a few different meters so I think they are valid. B and C phases seem to be the problem ones but I must admit I am surprised by the magnitude of the ground readings.

Circuit 1
A phase 202 and 198
B phase 225 and 169
C phase 202 and 156
Ground 79

Circuit 2
A phase 221 and 211
B phase 257 and 174
C phase 224 and 176
Ground 75

 

[Hv&Lv]

Thank you Winnie (or is it Jon?). That is the type of information I was seeking. For what it is worth, these are the current readings for the two different circuits. Circuit 2 is about 50 feet longer and both circuits are from the same circuit breaker. Circuit 1 has two screw compressors and all the condenser fans. Circuit 2 has two screw compressor and a centrifugal pump. Readings were done with a few different meters so I think they are valid. B and C phases seem to be the problem ones but I must admit I am surprised by the magnitude of the ground readings.

Circuit 1
A phase 202 and 198
B phase 225 and 169
C phase 202 and 156
Ground 79

Circuit 2
A phase 221 and 211
B phase 257 and 174
C phase 224 and 176
Ground 75

What are the voltage readings here? Voltage imbalance can really have a big effect on current imbalance

 

[synchro]

If possible on your setup, I suggest putting your clamp meter around both ground conductors to verify that there is no significant common mode current (i.e., current in the same direction) on the two ground wires. If you get a small reading then that verifies the currents you measured were circulating currents induced by the magnetic fields of the other wires, and not from some phase to ground currents within the chiller.

Since the inductive coupling that Jon mentioned can effect the current balance between the parallel wires on each phase, another proposed layout is:
Route the two cables for each phase together and twist them to the extent that is reasonable given your situation (length of the cables, cable stiffness and manageablity, etc.). So two A phase cables twisted, two B phase cables twisted, two C phase cables twisted. You could also twist the two ground wires (except for the extra length on the longer cable).
Doing this utilizes the same principle that twisted pairs provide for signals and audio, which is to prevent differential voltages or currents from being produced in the pair of wires by outside magnetic and EM fields. In your case the differential current would be the difference in measured currents between the two wires of each phase.

Just to give an example of the effect of inductive reactance, the following table shows that the impedance of 4/0 wires at 60 Hz increases by 35% with a spacing of 8 inches vs. 2 inches:

https://www.google.com/url?sa=t&rct=...CuBBlvK#page=8

 

[winnie]

I am going to second synchro's suggestions.

Think about the _loops_ of conductor that you have.

Any AC magnetic flux 'threading' the loop will induce current flow in that loop. You expect the ground wires to be connected at both ends, but since they are for separate circuits you are not running them together. Nice big loop with fat low resistance conductors, so a good chance for induced current. But the induced circulating current should be the same everywhere in the loop, so if you can put a clamp meter around the two ground wires such that the current will balance, then you will only read the 'residual' current which should be low.

Similarly your parallel conductors are connected at either end. The induced circulating current depends upon the AC magnetic flux threading the loop. If you bring the wires close together then you make the loop smaller, and if you _twist_ the wires together than the 'polarity' of the loop changes with each half twist.

A small nuance is that if you twist the wires, either as each phase twisted with itself or as the ABC triples, if you twist the different sets with different pitches (eg. one set has a twist every 5 feet, the next every 7 feet, etc) then you reduce coupling still further.

The 75 or so amps of current on the ground wires, if it is caused by magnetic coupling and not an actual ground fault, gives you an idea of the size of circulating currents you could see caused by the magnetic coupling in your situation.

-Jon (given name is Jon, e-mail address and forum id is winnie)

 

[oldsparky52]

Thank you Winnie (or is it Jon?). That is the type of information I was seeking. For what it is worth, these are the current readings for the two different circuits. Circuit 2 is about 50 feet longer and both circuits are from the same circuit breaker. Circuit 1 has two screw compressors and all the condenser fans. Circuit 2 has two screw compressor and a centrifugal pump. Readings were done with a few different meters so I think they are valid. B and C phases seem to be the problem ones but I must admit I am surprised by the magnitude of the ground readings.

Circuit 1
A phase 202 and 198
B phase 225 and 169
C phase 202 and 156
Ground 79

Circuit 2
A phase 221 and 211
B phase 257 and 174
C phase 224 and 176
Ground 75

I'd be checking connections with those readings.

 

[Russs57]

No way I'm going to get them to shut down long enough and allocate the labor to re-group and/or twist wires. They already think I am Chicken Little

I'll see if I can move the grounds close enough to get a reading on both conductors. Likely they already are at source end.

Thanks for all your help guys. This type of wiring isn't my wheelhouse.