connecting EGC and neutral for a line side tap (supply side connection)

[shortcircuit2]

I do not understand why it is required to do this at every single service disconnect (as in, what is the real life consequence in only doing it at one of the disconnects?), but somehow that is the way the code ended up being written.

The Main Bonding Jumper at each disconnect in a Service with multiple switches is needed because the jumper at each switch is sized for the Service Entrance Conductors that feed the switch. This becomes more applicable with larger switches and you size a wire bonding jumper according to the SE conductors feeding that switch.

If you were to install the bond at only 1 of the Service Disconnect Switches in a multiple separate switch service, it may be undersized the other switches. If there were a fault in multiple switches at the same time, the bonding jumper would be undersized for the fault incident. Also what if the only switch with the bonding jumper were removed...it would leave the others un-bonded.

It is important to bond at each Service Disconnecting Means to provide the low-impedance ground fault path back to the source with a bonding jumper correctly sized for the possible available fault current at that individual switch.

 

[jaggedben]

It does not matter. We are talking about a line side tap to connect a PV system, opening that switch will not kill power to the building.

Yes, it will.

I would have figured you understood how interactive inverters work, but perhaps you don't.

If you are trying to make about point about multiple disconnects, the PV disconnect is no different from another additional service disconnect allowed by 230.71.

If you are trying to make a point about about DC power being available somewhere else on the building (and it may not be, in any interesting way, with module level power electronics), that is also not a safety difference between a PV disconnect and another service disconnect, at least not with respect to the OP or anything in article 230 or 250.

 

[jaggedben]

Being compliant depends on whether the disconnect is technically a service disconnect or not. If not, it is not compliant. Even when it is a service disconnect, it would not be compliant with 250.6... but it seems all AHJ's ignore 250.6 when it comes to service bonding.

Points taken, but the question I was speaking to is whether there is any interesting difference between the PV disconnect and another disconnect that is undeniably a service disconnect. I submit that there is not.

As for AHJs ignoring 250.6, that's understandable considering a few different code sections that explicitly permit a parallel path. For example, if you must comply with 250.24(C) but are permitted to install per 250.64(D)(1) or (2), then whatever current travels on the GECs must not run afoul of 250.6, or else why do 250.64(D)(1) and (2) permit those methods? Just another case, I guess, where the code requires interpretation of it's logical inconsistencies.

 

[Smart $]

Points taken, but the question I was speaking to is whether there is any interesting difference between the PV disconnect and another disconnect that is undeniably a service disconnect. I submit that there is not.

As for AHJs ignoring 250.6, that's understandable considering a few different code sections that explicitly permit a parallel path. For example, if you must comply with 250.24(C) but are permitted to install per 250.64(D)(1) or (2), then whatever current travels on the GECs must not run afoul of 250.6, or else why do 250.64(D)(1) and (2) permit those methods? Just another case, I guess, where the code requires interpretation of it's logical inconsistencies.

A disconnect is a disconnect. Associated parameters may be changed, including what it is called, and through it all, it is still a disconnect.

250.6, 250.24, 250.64 are all requirements... but when it comes time to enforce one over the other, my best guess is AHJ's choose to put 250.6 at the bottom of the priority list because it is a judgement call while the others are not.

 

[shortcircuit2]

The danger of removing the ground-fault path back to the source outweighs 250.6 objectionable current.

PV Inverters just use the neutral for Utility sensing on most systems so there would be no objectionable current anyway.

 

[ggunn]

PV Inverters just use the neutral for Utility sensing on most systems so there would be no objectionable current anyway.

True in some cases but not all. An array of single phase inverters connected phase to neutral is not that uncommon.

 

[ggunn]

Yes, it will.

I would have figured you understood how interactive inverters work, but perhaps you don't.

I'm sorry, what? When inverters are connected to the supply side, opening the disco to the inverters does not interrupt power to the building; the PV disco and service disco are in parallel.

 

[jaggedben]

I'm sorry, what? When inverters are connected to the supply side, opening the disco to the inverters does not interrupt power to the building; the PV disco and service disco are in parallel.

If the PV system is also on the building, then the PV disco is just another disco connecting utility power to the building. If the PV disco is closed, there are AC conductors in the building that are supplied with utility AC power. It makes no safety difference if they are connected to a load or a source, as long as they are energized and utility can feed a fault.

Compare two services, both with a tap box after the meter leading to two discos. On Service A, the 2nd disco supplies only a PV system. On Service B, the 2nd disco supplies a single load. Is there any safety difference between the two 2nd discos that justifies applying any or all requirements in 230 or 250 to one and not the other?

As Smart$ said...

A disconnect is a disconnect. Associated parameters may be changed, including what it is called, and through it all, it is still a disconnect.

 

[Smart $]

If the PV system is also on the building, then the PV disco is just another disco connecting utility power to the building. If the PV disco is closed, there are conductors in the building that are supplied with utility AC power. I'm surprised this point is so confusing to talk about.

Compare two services, both with a tap box after the meter leading to two discos. On Service A, the 2nd disco supplies only a PV system. On Service B, the 2nd disco supplies a single load. Is there any safety difference between the two 2nd discos that justifies applying any or all requirements in 230 or 250 to one and not the other?

As Smart$ said...

A disconnect is a disconnect. Associated parameters may be changed, including what it is called, and through it all, it is still a disconnect.

For the record, yes I did say that, and is true. However, we are discussing conductors and current pathways that are associated with the disconnect, but not directly involved with the function of the disconnect itself.

 

[jaggedben]

For the record, yes I did say that, and is true. However, we are discussing conductors and current pathways that are associated with the disconnect, but not directly involved with the function of the disconnect itself.

I submit that the function is no different in any interesting way. The function is to disconnect the utility from the building in all cases.

 

[Smart $]

I submit that the function is no different in any interesting way. The function is to disconnect the utility from the building in all cases.

Only the ungrounded conductors. Disconnecting cuts off both power and the transfer of energy. Compliant bonding, sufficient arc flash containment and automated disconnect on fault are the essential differences. Is that enough of a safety issue to account for all the rhetoric? ... :angel:

 

[jaggedben]

Only the ungrounded conductors. Disconnecting cuts off both power and the transfer of energy. Compliant bonding, sufficient arc flash containment and automated disconnect on fault are essentially the only difference. Is that enough of a safety issue to account for all the rhetoric... :angel:

None of the things you've mentioned are a difference regarding safety. That is, there is no danger present in one that is not present in the other, in any way that calls for applying any sections of 230 or 250 differently. (Arc flash calculations might be different in nature, but the code requirement isn't.)

The only thing that could call for applying code differently is the wording of other parts of the code. :roll:

 

[c_picard]

The Main Bonding Jumper at each disconnect in a Service with multiple switches is needed because the jumper at each switch is sized for the Service Entrance Conductors that feed the switch. This becomes more applicable with larger switches and you size a wire bonding jumper according to the SE conductors feeding that switch.

If you were to install the bond at only 1 of the Service Disconnect Switches in a multiple separate switch service, it may be undersized the other switches. If there were a fault in multiple switches at the same time, the bonding jumper would be undersized for the fault incident. Also what if the only switch with the bonding jumper were removed...it would leave the others un-bonded.

It is important to bond at each Service Disconnecting Means to provide the low-impedance ground fault path back to the source with a bonding jumper correctly sized for the possible available fault current at that individual switch.

Thank you.

 

[iwire]

Yes, it will.

I would have figured you understood how interactive inverters work, but perhaps you don't.

I ran some of the first commercial PV system jobs in MA. I have installed a number of grid interactive systems both load side and line side connected. I have been to night classes for training.

I fully understand them and stand by my statement.

 

[jaggedben]

I ran some of the first commercial PV system jobs in MA. I have installed a number of grid interactive systems both load side and line side connected. I have been to night classes for training.

I fully understand them and stand by my statement.

As I said, so I would have figured, which is why I'm surprised by your position. Since you've declined to explain your reasoning, I remain baffled. :roll:

 

[iwire]

As I said, so I would have figured, which is why I'm surprised by your position. Since you've declined to explain your reasoning, I remain baffled. :roll:

I feel I did explain it.

The purpose of the supply side PV disconnect is not to disconnect the building from the utility.

You know you are free to disagree.

 

[jaggedben]

I feel I did explain it.

The purpose of the supply side PV disconnect is not to disconnect the building from the utility.

You know you are free to disagree.

And I'll state my disagreement one more time, because it's not different in any interesting physical respect from an additional service disconnect serving a load.

 

[iwire]

And I'll state my disagreement one more time, because it's not different in any interesting physical respect from an additional service disconnect serving a load.

Except it is not serving a load, it is serving a source unless you are designing systems with no output. That to me is both intresting and significant.

To each their own, the CMPs will end up clearfying things eitherway eventually.

 

[romex jockey]

You mean more electrical theory thrown under the CMP bus, as in on the '17 horizon , after all the ado the dimmer cabal made to create it, is a 404.2 exception allowing an ECG utilized where no neutral exists


~RJ~

 

[ggunn]

I feel I did explain it.

The purpose of the supply side PV disconnect is not to disconnect the building from the utility.

You know you are free to disagree.

FWIW, this looks like a "the grass is green vs. the sky is blue" argument to me. If the PV is interconnected on the supply side, the PV AC disco disconnects the PV system from the utility. It does not disconnect the building from the utility; there is another switch that does that. If the PV system is interconnected on the load side, then there is one switch that does both.