Continuous Duty - Disconnect vs. Breaker

[ggunn]

Ok consider a commercial rooftop system with 450 amps of inverter output current. I am using a larger system because it will have higher utility fault currents whereas a resi system may not have an issue. I come down off the roof with inverter output circuits to a combiner panelboard on the side of the building. Next to that I have a non fused disco. The panelboard and disco are 600 A. After the disco, the feeder goes through the exterior wall into the electrical room where it connects to the interconnect breaker (supply or load side, shouldn't matter for this discussion). The SCCR of the switch is 10k with this configuration. This is likely too low and a code violation. Do you agree with this or do you think my reasoning is incorrect?

Can you point me to the language in the NEC where you think the violation lies? I think this scenario does not address the issue at least as far as the ampere rating of the switch is concerned, since 450A is too much for even a 100% rated 400A switch and 600A (the next size up) is more than 125% of 450A. If your inverter output were 380A, the question would be if a 400A unfused switch would suffice, or would you have to use a 600A switch. I maintain that the 400A switch would be sufficient.

 

[jaggedben]

... I interconnect [25A through 30A switches virtually every day and I have never had an AHJ challenge it.

Mind my asking what brand? And if it's documented on the label that they are rated for 100% continuous use? I've been using 60A switches in exactly the situation you describe because I didn't realize I might have another option.

 

[electrofelon]

Can you point me to the language in the NEC where you think the violation lies? I think this scenario does not address the issue at least as far as the ampere rating of the switch is concerned, since 450A is too much for even a 100% rated 400A switch and 600A (the next size up) is more than 125% of 450A. If your inverter output were 380A, the question would be if a 400A unfused switch would suffice, or would you have to use a 600A switch. I maintain that the 400A switch would be sufficient.

I agree that there is no derating for the switch. In my example I said the switch was 600 just because that was the next size up (it just so happened to also cover a,125% adder if there was one). I am talking about the SCCR of the switch, analogous to the interrupting rating of an OCPD. IIRC, nec 110.9 covers IR and 110.10 covers SCCR.

 

[electrofelon]

Mind my asking what brand? And if it's documented on the label that they are rated for 100% continuous use? I've been using 60A switches in exactly the situation you describe because I didn't realize I might have another option.

It doesn't need to be labeled for 100% continous use. There is no continuous derating for switches

 

[ggunn]

I agree that there is no derating for the switch. In my example I said the switch was 600 just because that was the next size up (it just so happened to also cover a,125% adder if there was one). I am talking about the SCCR of the switch, analogous to the interrupting rating of an OCPD. IIRC, nec 110.9 covers IR and 110.10 covers SCCR.

That's a completely different question from what I was talking about. If you have an inverter to interconnect that has (picking a number out of the air) 380A rated output, you can use a 400A unfused switch, but if the switch is fused, it must be a 600A switch (the next size up). You cannot put 500A fuses in a 400A switch.

I don't see why the SCCR of the switch is an issue. Under normal operation the inverter cannot produce more than its rating and hence will never overload the switch. Under fault conditions if the interconnection OCPD does not shut the circuit down the wire insulation will fail before the switch does, and you wouldn't use a manual switch to clear a fault, anyway.

 

[GoldDigger]

The SCCR does not relate to an interrupt rating. That is AIC.
The SCCR assures that the switch will not fail (in a potentially hazardous way) under fault current from the service side before the service OCPD can open.

 

[jaggedben]

It doesn't need to be labeled for 100% continous use. There is no continuous derating for switches

Is there a code section for that?

 

[electrofelon]

Is there a code section for that?

I think the burden of proof is on you to show that there IS a derating for continuous loads required for a non fused safety switch Note that the NEC 80% continuous derating we are familiar with is for OCPD's branch circuits, feeders, and service entrance conductors. I think you will find them in 210.19(A)(1), 210.20(A), 215.2(A)(1), 215.3, 230.42. It does not apply generally to all equipment. Class 320 ("400 amp") meter sockets are derated 80% for continuous use, but that is a product standard or UL thing not an NEC requirement.

 

[electrofelon]

That's a completely different question from what I was talking about. If you have an inverter to interconnect that has (picking a number out of the air) 380A rated output, you can use a 400A unfused switch, but if the switch is fused, it must be a 600A switch (the next size up). You cannot put 500A fuses in a 400A switch.

I agree with that. Sorry for the confusion, there are two topics going on in the thread now.

I don't see why the SCCR of the switch is an issue. Under normal operation the inverter cannot produce more than its rating and hence will never overload the switch. Under fault conditions if the interconnection OCPD does not shut the circuit down the wire insulation will fail before the switch does, and you wouldn't use a manual switch to clear a fault, anyway.

I see it as the same analogous issue as coordinating the interrupting rating of an OCPD to the available fault current. Things that dont have interrupting duties, still must be able to withstand any fault current imposed upon them, and as with interrupting ratings, it doesnt matter if there is an OCPD ahead of the device (unless there is a series rating) since that high fault current will still exist until that OCPD clears.

 

[ggunn]

I see it as the same analogous issue as coordinating the interrupting rating of an OCPD to the available fault current. Things that dont have interrupting duties, still must be able to withstand any fault current imposed upon them, and as with interrupting ratings, it doesnt matter if there is an OCPD ahead of the device (unless there is a series rating) since that high fault current will still exist until that OCPD clears.

What possible use could there be for an unfused disco to be able to interrupt 10kA of current when the conductors that feed it are only rated for 100A? Or whatever?

 

[electrofelon]

What possible use could there be for an unfused disco to be able to interrupt 10kA of current when the conductors that feed it are only rated for 100A? Or whatever?

The unfused disco doesn't interrupt a fault so it doesnt have an interrupting rating. However it does have a SCCR. If there is a fault on the PV side of that disco (feeder, combiner panelboard, inverter out output circuit, etc) the utility will feed that fault and potentially 10's or 100's of thousands of amps will flow until the fault is cleared. The disco needs to be able to handle that fault current without catastrophically failing. That is what the SCCR rating is. Lots of people overlook SCC ratings. I admit to never really thinking about it for things like meter sockets and unfused discos until we were nailed on it last summer.

 

[iwire]

But you would still have high fault currents from the utility side resulting from an inverter failure or fault on the inverter output circuit. You can't just cosnider the pV side.

That's what the OCPD at the interconnect is for.

That OCPD does not lower the fault current.

 

[iwire]

Is there a code section for that?

How about posting the code section that requires derating switches?

 

[iwire]

What possible use could there be for an unfused disco to be able to interrupt 10kA of current when the conductors that feed it are only rated for 100A? Or whatever?

Forget about interrupting, the switch has to rated high enough it does not turn into a grenade when fault current is passed through it.

 

[ggunn]

Forget about interrupting, the switch has to rated high enough it does not turn into a grenade when fault current is passed through it.

I understand, but in the presence of very large fault current from the service the conductors will short out long before the switch fails.

 

[iwire]

I understand, but in the presence of very large fault current from the service the conductors will short out long before the switch fails.

How do you know that?

 

[ggunn]

Lots of people overlook SCC ratings. I admit to never really thinking about it for things like meter sockets and unfused discos until we were nailed on it last summer.

Can you give me details on what equipment you installed for what application and what aspect of the inspection you failed? I assume (I hope) that your being nailed was an inspection failure and not a catastrophic system failure.

I am not looking to simply win an argument here; I just want to design safe and code-compliant PV systems. I have had hundreds of systems pass inspections in several jurisdictions and this issue has never come up.

 

[Carultch]

The unfused disco doesn't interrupt a fault so it doesnt have an interrupting rating. However it does have a SCCR. If there is a fault on the PV side of that disco (feeder, combiner panelboard, inverter out output circuit, etc) the utility will feed that fault and potentially 10's or 100's of thousands of amps will flow until the fault is cleared. The disco needs to be able to handle that fault current without catastrophically failing. That is what the SCCR rating is. Lots of people overlook SCC ratings. I admit to never really thinking about it for things like meter sockets and unfused discos until we were nailed on it last summer.

Is it possible/practical to retrofit existing unfused disconnects to convert them into fused disconnects?

 

[Carultch]

Can you give me details on what equipment you installed for what application and what aspect of the inspection you failed? I assume (I hope) that your being nailed was an inspection failure and not a catastrophic system failure.

I am not looking to simply win an argument here; I just want to design safe and code-compliant PV systems. I have had hundreds of systems pass inspections in several jurisdictions and this issue has never come up.

It was probably an inspection failure. SCCR and KAIC ratings are an easy thing to overlook, as it is a very confusing topic to learn correctly.

I can't imagine that this really would be a catastrophic system failure, because the unfused switch is likely built with exactly the same blade assembly and enclosure as the fused switch. Plus, you can put the fuses in a separate enclosure, and still take credit for the higher SCCR rating of the disconnect. By the standard way you are supposed to connect it, the blade is also on the line side of the fuses, even when the fuses are an integral part of the disconnect. It doesn't make sense that the blade would fail at 11 kA when fuses are not present in the circuit, while not failing until 201 kA when fuses are present in the circuit. The difference is simply too extreme.

 

[ggunn]

It was probably an inspection failure. SCCR and KAIC ratings are an easy thing to overlook, as it is a very confusing topic to learn correctly.

That's what I figured as well, but I would be very interested in learning the specifics.