Battery interconnection wire size

[Sunsun]

Perhaps I should have been clearer. I'm wondering what, if anything, protects the battery to battery interconnects cables if a cell shorts out. I assume a large amount of current will flow through the interconnect cables to that shorted cell.

That did happen to me while we were away on vacation and had to spend an extra week out waiting for a used replacement rear end for an 87 Dodge camper van.

You are correct, when a cell shorts the battery heats up as the electrolyte boils off. Then a Cascade effect happens where the two adjacent cells boil off their electrolyte.

The inverter did not shut down due to low voltage and the 250 Amp DC panel breaker did not trip (Breaker between the battery bank and inverter).

And that was just one shorted cell in a large 1380 amp battery bank. A 4/0 cable can handle 302 amps and a 2/0 cable 190 amps. I don't think they are protected if a cell shorts.

I consider myself lucky that I Dodged a bullet on that one and got home before any fire started.



Chris

[NewBostonConst]

Thanks for all the replies. Mike's 'if one wire fails' point is good practical advice and has convinced me to go ahead and buy the 4/0 cables even though parallel 2/0 wires would have more load carrying capacity than one 4/0 wire.


Of course except for a short situation, its unlikely that much current would flow though those interconnect wires. But may as well do the "safety first" thing.

This does bring up another question. Does the main breaker to the inverter from the battery protect the interconnect cables when a cell shorts out?

Chris

Your question doesn't make sense. So I will try my best.

The cable from the battery to the inverter has a circuit breaker. It normally protects the batteries from the invertor pulling to much current and damaging them. Shorted cells usually will cause the battery to heat up and have low voltage which eventually cause the inverter to shut down. Shorted cells can cause a fire but I don't think the breaker will trip. The interconnect cables are the ones that go between AC breaker and the inverter? I don't they will trip either if the battery shorts.

[Sunsun]

Thanks for all the replies. Mike's 'if one wire fails' point is good practical advice and has convinced me to go ahead and buy the 4/0 cables even though parallel 2/0 wires would have more load carrying capacity than one 4/0 wire.


Of course except for a short situation, its unlikely that much current would flow though those interconnect wires. But may as well do the "safety first" thing.

This does bring up another question. Does the main breaker to the inverter from the battery protect the interconnect cables when a cell shorts out?

Chris

[bob-n]

I don't know NEC well enough to comment there. I'm not an electrician and am not giving electrical advice.

foo1bar is certainly right that there isn't much need in residential for parallel.

SunEagle's point made me think. So many things can go wrong. Connections can fail. Stranded wire can have a broken strand. Solid wire can have a dent or crimp. Installation can be defective. I guess that those issues are already addressed by testing and statistics. Hundreds of millions of homes are built to the NEC and UL/CSA approved devices are submitted to rigorous testing. For everyone's sake, we hope that all of that data is incorporated into the code. For example, the NEC includes a restriction on the number of bends in conduit to minimize the risk to the wire.

Parallel wiring situations are certainly not as rigorously tested. The issues that can arrise in parallel wiring are more complex because there are more variables. Not all connectors are engineered and tested for more than one wire.

So if it isn't in the code, don't do it. There is no cost benefit to using two smaller wires instead of one larger wire. Do it right. SunEagle's original advice ("I would never run parallel wires...") is very good, as with all of his advice. End of discussion.

But as an engineer, I don't see the risk. SunEagle's point of one wire being weaker than the other seems to be the same risk as part of a larger wire being damaged. We could argue that the odds of one of two connections failing are higher than the odds of one of one connection failing because there are more points to fail. Mike90250 makes the point that one single failure is easier to detect than one of two failing. The most common analogy is that many people drive with one headlight burned out and don't know it, but when both fail, you surely know it.

I could also argue that having a second connection is more fault tolerant. If one connection is defective and becomes resistive, the other will take more current and reduce the burden on the bad connection. If a single connection fails intermittently, it may arc, leading to a fire. If one of two connections fails intermittently, the other will take the load and there won't be any arc at all. The other may not be reliable to handle the load continuously for a long time, but it would be reliable for short intervals of overload. Again with the headlight analogy, you can't see as well with only one headlight compared to two, but you can see much better with one than with none.

Ampster's point about current sharing being based on resistance is also valid. If things are built correctly and there are no defects, two wires in parallel will be just as good as one of twice the size, even if the two wires are slightly different length. A small difference in length will result in a small difference in current balance, exceeding the rating of the wire, but by an insignificant amount. Nothing is perfect. NEC has margin built in for this very reason. A 15A circuit is only allowed to conduct 12A continuously.

But back to the original point: Don't do it.

[foo1bar]

NEC doesn't allow parallel wires in residential, but does in commercial. Not that I really care if you do it. But paralleling wires like that looks like a hack job.

NEC allows parallel in residential. (If you think I'm wrong, point me to what code section says I'm wrong)

But in practicality - there isn't much need in residential, as parallel wires is at least 1/0, and they must be same electrical characteristics though (same length, cross section, etc) (NEC 310.4 and NEC 310.10)

So why bother with parallel in residential if you're just doing 200A service anyhow. Just get 2/0 and be done. (or 4/0 aluminum)

In general, I'd say if you can meet your needs with 4/0, I would use 4/0 rather than do parallel.

Or in the case of batteries, I would probably look at doing appropriately sized copper bus bars that were
But that's me.

[SunEagle]

Help me translate that into Ohm's law or physics. I think what you are saying is that if one wire has more resistance than the other wire with less resistance could overheat if the current exceeds the rating of that wire. I suppose one could calculate how much resistance it would take to force that current using I=V/R. My guess is if one used the same size wire of the same length and terminated them properly the current difference would be trivial and the risk of overheating one conductor would be reduced.
I admit my question to @Takis was a little snarky or seemed like baiting him to test his logic. I have more knowledge of your experience to be able to shed some light on this for the greater good. Apparently according to one source it is allowed under NEC for commercial. I am not asking for an endorsement.

2 identical wires in parallel should each carry 1/2 the current. But if one of those wires is weaker then the other it may carry more current and exceed it's rating.

[Ampster]

For power cabling both wires need to be exactly the same resistance or the current will take the path of least R and overheat the wires.

Parallel wiring is ok but it has to be exactly the same type and length along with being properly terminated.

Help me translate that into Ohm's law or physics. I think what you are saying is that if one wire has more resistance than the other wire with less resistance could overheat if the current exceeds the rating of that wire. I suppose one could calculate how much resistance it would take to force that current using I=V/R. My guess is if one used the same size wire of the same length and terminated them properly the current difference would be trivial and the risk of overheating one conductor would be reduced.
I admit my question to @Takis was a little snarky or seemed like baiting him to test his logic. I have more knowledge of your experience to be able to shed some light on this for the greater good. Apparently according to one source it is allowed under NEC for commercial. I am not asking for an endorsement.

[SunEagle]

I'm confused by the comment about not running parallel wires. Why?

By my way of thinking, more copper is always better and will result in less voltage drop. Even if one wire is skinny and one is fat, they will still carry current in the same direction. If the fat wire was enough, the skinny additional wire will give safety margin.

Current capacity is roughly set by the amount of copper. Surface area to dissipate heat is another variable, but if that is the constraint, two thin wires may actually be better than one fat one because they have more surface area and are spaced apart.

If you need #10 and have two #12 conductors in parallel, I think you will still be OK. They may be slightly different diameter or slightly different length, so they won't share the current precisely equally, but the current mismatch will be no worse than the wire mismatch. I=V/R where V is the voltage drop from one end to the other and R is the resistance of the wire under consideration.

In my work, we do wires in parallel of different size all the time, but that's in a different field, not high-power solar.

For power cabling both wires need to be exactly the same resistance or the current will take the path of least R and overheat the wires.

Parallel wiring is ok but it has to be exactly the same type and length along with being properly terminated.

[NewBostonConst]

NEC doesn't allow parallel wires in residential, but does in commercial. Not that I really care if you do it. But paralleling wires like that looks like a hack job.

[Mike90250]

I'm confused by the comment about not running parallel wires. Why?..........

Because there is the chance a wire comes loose or goes bad, or a crimp falls apart. Then the remaining wire has to be sized to carry the load, so when sized right, using a single larger gauge is better than using 2 just acceptable gauges. A single wire failure is noticed right away, if a wire in a double run fails, nobody notices until the fire starts.

[bob-n]

SunEagle: What is the reason why you don't recommend parallel wires?

Takis: Warning: Ampster is baiting you.

Current doesn't take the shortest path. Just like water flow, current flow will split between parallel paths. Most will take the easiest path (fattest wire), but in this sense, many parallel wires are no different from wide wire.

Takis does have a point, however. If there is a very short, skinny wire and a very long fat wire, the skinny wire might wind up carring too much current. That's not what I call parallel paths, because the lengths are very different.

As Ampster implies, it all comes down to algebra and Ohm's law. If the wire lengths are within 10% of each other, then the current will share to a similar tolerance.

[Takis]

Can you explain that in terms of Ohm's law or physics for the benefit of future readers.

To be honest, no.
I've just posted as a simplified answer.

[Ampster]

Because the current always chooses the shortest way.
Difference in wire length would be devastated.

Can you explain that in terms of Ohm's law or physics for the benefit of future readers. Can we assume those wires are terminated at the same place? As a hint, the resistance of a thinner longer wire would be greater than that of a shorter fatter wire.

[Takis]

Because the current always chooses the shortest way.
Difference in wire length would be devastated.

[bob-n]

I'm confused by the comment about not running parallel wires. Why?

By my way of thinking, more copper is always better and will result in less voltage drop. Even if one wire is skinny and one is fat, they will still carry current in the same direction. If the fat wire was enough, the skinny additional wire will give safety margin.

Current capacity is roughly set by the amount of copper. Surface area to dissipate heat is another variable, but if that is the constraint, two thin wires may actually be better than one fat one because they have more surface area and are spaced apart.

If you need #10 and have two #12 conductors in parallel, I think you will still be OK. They may be slightly different diameter or slightly different length, so they won't share the current precisely equally, but the current mismatch will be no worse than the wire mismatch. I=V/R where V is the voltage drop from one end to the other and R is the resistance of the wire under consideration.

In my work, we do wires in parallel of different size all the time, but that's in a different field, not high-power solar.