Ground Wire sizing in Transformerless Grid Tied

[bbartmasse r]

Hi All,

I've been confused about sizing the GEC/EGC in grid tied systems for some time now. My company has just defaulted to having a #8 ground (at least) everywhere in the system to cover our butts.

My question is concerned with Grid tied, Transformerless(ungrounded), rooftop mounted PV arrays, and specifically with the wire run from the inverter to the main panel. Let say for example the circuit requires #10 conductors and we are running romex. In the past we would have pulled an additional #8 ground with the romex from the inter to the panel. My suspicion is that this is unnecessary?

Will the #12 wire that is included with 10-3 romex be sufficient in this case and most importantly, meet code.

Thanks for your input in advance

 

[Carultch]

Hi All,

I've been confused about sizing the GEC/EGC in grid tied systems for some time now. My company has just defaulted to having a #8 ground (at least) everywhere in the system to cover our butts.

My question is concerned with Grid tied, Transformerless(ungrounded), rooftop mounted PV arrays, and specifically with the wire run from the inverter to the main panel. Let say for example the circuit requires #10 conductors and we are running romex. In the past we would have pulled an additional #8 ground with the romex from the inter to the panel. My suspicion is that this is unnecessary?

Will the #12 wire that is included with 10-3 romex be sufficient in this case and most importantly, meet code.

Thanks for your input in advance

If it meets the sizing algorithm in 250.122, it is sufficient as the EGC, whether it is on the AC or DC sides. The GEC is not required for non-isolated inverters, due to the fact that they carry the same EGC system to both sides, and do not use it to derive the absolute voltage of either polarity.

Since #12 has a maximum ampacity for a 20A circuit, and the maximum OCPD for a #12 EGC is 20A, I do see an application where the combination of #10 main and #12 EGC is code compliant. Therefore, I wouldn't expect a manufactured cable to be made that way.

Most multiconductor cables I see that are #10 and less in size, have an equally sized EGC.

 

[Smart $]

If it meets the sizing algorithm in 250.122, it is sufficient as the EGC, whether it is on the AC or DC sides.

What about compliance with 250.166 or the larger of 250.122 and 250.166 as called for in 690.47?

The GEC is not required for non-isolated inverters, due to the fact that they carry the same EGC system to both sides, and do not use it to derive the absolute voltage of either polarity.

I've heard mention of the concept but I do not see it in Code. Exactly where is it?

 

[jaggedben]

What about compliance with 250.166 or the larger of 250.122 and 250.166 as called for in 690.47?

That would be 2011 or earlier, and is where the minimum #8 comes from. But under 2014 it's quite explicit that an ungrounded system can be sized to 250.122. (There's a strong argument that it still needs to be installed to 250.64 standards for GEC's, however.)

Under 2017, thankfully, all that is going away and we'll no longer need anything except equipment grounding, except that a structure with an array will require a GES even if it otherwise wouldn't.

I've heard mention of the concept but I do not see it in Code. Exactly where is it?

It's in the UL standard. An inverter with no grounded DC conductor does not require a GEC per UL. Per the 2014 NEC it still does but it can be sized to 250.122. Per the 2017 it doesn't.

***

The one thing that yet again throws a monkey wrench into things is 690.47(D), for those on the 2014 code. If your AHJ insists on enforcing it, and you're trying to invoke the exception that allows you to avoid adding an electrode within 6ft, then you'll need that #8 to satisfy 250.166. As discussed in other recent threads, following 690.47(D) without invoking the exception is not - to put it as politely as possible - really the most advisable thing to do.

FWIW, I dealt with one AHJ who insisted even on the 2011 code on a #8 from the array to the GES based on 250.8 and 250.166, and claimed that 690 had nothing to do with it because that only spoke to inverter grounding.

 

[bbartmasse r]

If it meets the sizing algorithm in 250.122, it is sufficient as the EGC, whether it is on the AC or DC sides. The GEC is not required for non-isolated inverters, due to the fact that they carry the same EGC system to both sides, and do not use it to derive the absolute voltage of either polarity.

Since #12 has a maximum ampacity for a 20A circuit, and the maximum OCPD for a #12 EGC is 20A, I do see an application where the combination of #10 main and #12 EGC is code compliant. Therefore, I wouldn't expect a manufactured cable to be made that way.

Most multiconductor cables I see that are #10 and less in size, have an equally sized EGC.

Thank you! and my mistake, the NM cable does have a #10 ground, not #12