Transformer OCPD

[ggunn]

Is it your interpretation that 690.9(D) allows us to not require a secondary OCPD, even when 240.21(C) would require it?

The Exception to 690.9(D) is simply stated and succinct, and it makes sense to me. A short on the inverter (secondary) side of the transformer will pull enough current through the primary to trip the OCPD between the transformer and the service when it is protected on the primary only side per T450.3(B). The inverter cannot source enough current to endanger the transformer or wiring under any circumstances if the equipment is sized correctly. It seems pretty straightforward to me; is there a set of conditions I am not taking into account?

 

[wwhitney]

The Exception to 690.9(D) is simply stated and succinct, and it makes sense to me. A short on the inverter (secondary) side of the transformer will pull enough current through the primary to trip the OCPD between the transformer and the service when it is protected on the primary only side per T450.3(B). The inverter cannot source enough current to endanger the transformer or wiring under any circumstances if the equipment is sized correctly. It seems pretty straightforward to me; is there a set of conditions I am not taking into account?

That seems in accordance with the 2020 NEC, under the assumption that there are no loads on the secondary side of the transformer. If there are loads, then addressing possible overload conditions is not as simple as sizing the equipment and conductors to the maximum inverter output current.

But the 2023 NEC introduced 705.30(F), which requires compliance with 240.21(C). And I don't think the 690.9(D) exception is sufficiently specific to amend 705.30(F). [90.3 is a bit unclear on this question, it says that Chapters 5-7 may amend Chapters 1-7, so what happens when 690 tries to amend 705, and 705 tries to amend 690?] Probably why it was removed in the 2026 NEC.

Cheers, Wayne

 

[jaggedben]

In light of Wayne's observation about the next code cycle, I'll refrain from trying to parse how exactly the 2023 code should be read with respect to 90.3. Deleting 690.9(D) is an appropriate solution. It should have been deleted and/or its content transferred to article 705 back when article 705 was created (2011 iirc). Because it's not PV specific (and certainly doesn't apply to PV DC circuits!); it's a rule that should apply just as appropriately to the AC output of an ESS or other more exotic source such as small wind or fuel cell or hydro. It's long been my opinion that no such rules belong in 690 except in as much as they reference to another article such as 705 or 710. That's mostly been true now since 2011 except for a few things that have been 'missed'. The exception to 690.9(D) should have been deleted or moved to 705 when the main content was changed to refer to 705. 705 now covers this by referring to 240.21(C). In 2026 any confusion will be gone, which is good.

 

[ggunn]

That seems in accordance with the 2020 NEC, under the assumption that there are no loads on the secondary side of the transformer. If there are loads, then addressing possible overload conditions is not as simple as sizing the equipment and conductors to the maximum inverter output current.

But the 2023 NEC introduced 705.30(F), which requires compliance with 240.21(C). And I don't think the 690.9(D) exception is sufficiently specific to amend 705.30(F). [90.3 is a bit unclear on this question, it says that Chapters 5-7 may amend Chapters 1-7, so what happens when 690 tries to amend 705, and 705 tries to amend 690?] Probably why it was removed in the 2026 NEC.

Cheers, Wayne

Code entanglement aside, under what scenario could using the Exception to 690.9(D) interpreted as I have present a hazard to equipment or personnel? It is still there in the 2023 NEC.

 

[jaggedben]

Code entanglement aside, under what scenario could using the Exception to 690.9(D) interpreted as I have present a hazard to equipment or personnel? It is still there in the 2023 NEC.

Broadly, any hazard that would result from not following 240.21(C) in any installation. (That is, regardless of whether the secondary has any sources connected vs. just ordinary loads.)

 

[ggunn]

Broadly, any hazard that would result from not following 240.21(C) in any installation. (That is, regardless of whether the secondary has any sources connected vs. just ordinary loads.)

Less broadly? The simplest case would be a single inverter with an integrated AC disconnect connected directly to the secondary of a 208V delta primary 480/277V wye secondary transformer which is protected on the primary side with OCPD 125% or less of the transformer-rated current per T450.3(B).

 

[jaggedben]

Less broadly? The simplest case would be a single inverter with an integrated AC disconnect connected directly to the secondary of a 208V delta primary 480/277V wye secondary transformer which is protected on the primary side with OCPD 125% or less of the transformer-rated current per T450.3(B).

I'm not super familiar with 240.21(C) but it appears that would only be allowed if the integrated disconnect has overcurrent protection. If the secondary were delta not wye I think it's ok.

 

[ggunn]

I'm not super familiar with 240.21(C) but it appears that would only be allowed if the integrated disconnect has overcurrent protection. If the secondary were delta not wye I think it's ok.

The secondary must be wye for most 480V inverters and the integrated AC disco would not have OCP for any that I have seen, but I am still not seeing the hazard and it seems to me that the Exception to 690.9(D) explicitly allows this.

 

[jaggedben]

The secondary must be wye for most 480V inverters and the integrated AC disco would not have OCP for any that I have seen, but I am still not seeing the hazard and it seems to me that the Exception to 690.9(D) explicitly allows this.

As Wayne said, under the 2020 NEC you can credibly make that argument. Under the 2023, it's much less clear; Can an exception in 690 really override a requirement in 705? I'm inclined to say no, although I acknowledge there's a legalistic, if unprincipled, argument that it can because of "shall be permitted". Under the 2026 NEC, however, it will not be allowed, so take note.

As far as the hazards, that's above my pay grade, but I vaguely recall previous discussions about this. I assume they had a technical substantiation at some point for allowing primary side protection for deltas but not wyes. Something about the ratio of secondary L-N fault current to primary L-L fault current iirc.

 

[wwhitney]

I assume they had a technical substantiation at some point for allowing primary side protection for deltas but not wyes. Something about the ratio of secondary L-N fault current to primary L-L fault current iirc.

The issue is the multiple voltages present on the secondary side with a wye.

If you have a 2-wire primary and secondary, or 3-wire delta primary and secondary, then there is only one voltage present between any two primary wires, and also only one voltage present between any two secondary wires. The currents on the primary and secondary will be in inverse ratio to the voltages, and so if you have a 100A primary OCPD, and a 2:1 primary:secondary voltage ratio, then the secondary conductors are effectively protected at 200A.

Whereas with a wye secondary, you have two different voltages present, in ratio of sqrt(3); for a 120/240 single phase secondary, you have two different voltages present in a ratio of 2. The worst case current ratio for an overload or fault would be based on the lower secondary voltage. But for carrying balanced power, you'd normally want to size your conductors based on the higher secondary voltage. So relying on the primary OCPD to protect the secondary conductors would entail oversizing your secondary conductors on a factor of sqrt(3) or a factor 2.

I don't believe there is any technical reason that 240.21(C) couldn't provide an allowance to rely on primary OCPD in these cases if you oversize your secondary conductors appropriately. Perhaps it's a matter of the allowance not being useful, as you'd rarely choose to oversize rather than provide secondary OCPD. Or maybe the CMP would just view the resulting rule as too complicated.

Cheers, Wayne

 

[pv_n00b]

If you look back in previous versions of the NEC the original purpose of 690.9(D) was to define how to apply 450.3 to a power transformer with an inverter on one side and the utility on the other. The purpose of the exception was to allow a power transformer to be exempt from 450.3 requirements, when considering the inverter side the primary side, if the rating of the transformer was not less than the inverter(s) total rating. Then in the 2023 NEC 690.9(D) was rewritten to point to 705.30(F) which has two things, one is that it defines the primary side for OCPD purposes to be the side with the largest source. The utility in every PV system I have seen. This mimics the intent of older versions of 690.9(D) in that we don't have to apply 450.3 to the transformer for both sides being the primary in turn. But it confuses the issue because 705.30(F) also has a second requirement relating to conductor protection that 690.9(D) never had. This messes up the exception in 690.9(D) that was only to apply to 450.3. Historically 690.9(D) never was applied to any requirement for OCPD protection of conductors connected to the transformer. NEC 690.9(D) has been removed from the 2026 NEC in the second draft.

 

[wwhitney]

If you look back in previous versions of the NEC the original purpose of 690.9(D) was to define how to apply 450.3 to a power transformer with an inverter on one side and the utility on the other. The purpose of the exception was to allow a power transformer to be exempt from 450.3 requirements, when considering the inverter side the primary side, if the rating of the transformer was not less than the inverter(s) total rating.

By this I take it you mean avoiding applying 240.21(C) to both the utility side and the inverter side of the transformer, on the basis that either side could be considered the secondary. Interestingly, 705.30(F) no longer has the requirement that the rating of the transformer is not less than the total rating of the inverters.

Cheers, Wayne

 

[pv_n00b]

By this I take it you mean avoiding applying 240.21(C) to both the utility side and the inverter side of the transformer, on the basis that either side could be considered the secondary. Interestingly, 705.30(F) no longer has the requirement that the rating of the transformer is not less than the total rating of the inverters.

Cheers, Wayne

Nope, not what I mean at all. 690.9(D) was never intended to impact 240.21 requirements, only the application of 450.3.
By changing the wording of 690.9(D) in 2023 to point to 705.30(F) the purpose of 690.9(D) was lost and the exception in 690.9(D) became meaningless. It was a mistake by the CMP and it looks like it will be fixed in the 2026 NEC by deleting 690.9(D). But since 705.30(F) does not address the application of 450.3 to transformers with sources on both sides that is now lost too.

 

[wwhitney]

Nope, not what I mean at all. 690.9(D) was never intended to impact 240.21 requirements, only the application of 450.3.

OK, I guess I'm not following. Let's call the current choices in 705.30(F) (utility side is primary, inverter side is secondary) "normal" and the opposite choice "reversed."

What's an example of a transformer installation that would comply with both 450.3 and 240.21(C) in the "normal" point of view, but which fails 450.3 in the "reversed" point of view? It seems to me that the need to comply with 240.21(C) in the "normal"" point of view is going to lead to 450.3 compliance in the "reversed" point view via the 125% primary only option, at least for low voltage transformers subject to 450.3(B).

Now a requirement to comply with 240.21(C) in the "reversed" point of view would definitely impose additional limits, as the normal primary conductors do not have length limitations. But 705.30(F) makes it clear you don't have to do that.

Cheers, Wayne

 

[pv_n00b]

You keep trying to pull 240.21 into 690.9(D) and historically the connection never existed. Here is 690.9(D) from the 2020 NEC.

(D) Power Transformers.
Overcurrent protection for a transformer with a source(s) on each side shall be provided in accordance with 450.3 by considering first one side of the transformer, then the other side of the transformer, as the primary.
Exception: A power transformer with a current rating on the side connected toward the interactive inverter output, not less than the rated continuous output current of the inverter, shall be permitted without overcurrent protection from the inverter.

The purpose of 690.9(D) was to provide guidance on the application of 450.3 only. This section is logical, internally consistent, and the exception makes sense. What was changed in the 2023 NEC disconnected it from 450.3 and introduced 240.21 in 705.30(F). 690.9(D) is no longer consistent and the exception makes no sense in the current context. The change to 690.9(D) in 2023 was an error that was identified and will be corrected in the 2026 NEC. There is no way to apply the 2023 690.9(D) to the application of 240.21. It does not work.

 

[ggunn]

You keep trying to pull 240.21 into 690.9(D) and historically the connection never existed. Here is 690.9(D) from the 2020 NEC.

The purpose of 690.9(D) was to provide guidance on the application of 450.3 only. This section is logical, internally consistent, and the exception makes sense. What was changed in the 2023 NEC disconnected it from 450.3 and introduced 240.21 in 705.30(F). 690.9(D) is no longer consistent and the exception makes no sense in the current context. The change to 690.9(D) in 2023 was an error that was identified and will be corrected in the 2026 NEC. There is no way to apply the 2023 690.9(D) to the application of 240.21. It does not work.

But I will ask again: What is the hazard? The ostensible purpose of the NEC is to provide rules to protect people and property from electrical hazard; what is it in this case? I am not stating unequivocally that there is no hazard, only that I do not see it.

 

[SteveO NE]

Is it your interpretation that 690.9(D) allows us to not require a secondary OCPD, even when 240.21(C) would require it?

You are assuming the transformer, and the secondary conductors are inside - if they are outside take a look at (C)(4). My issue has always been OCPD on the inverter itself that makes this impracticable to have absolutely no OCPD on the secondary side, but it doesn't mean the exception isn't used because the exception is for primary protection on the secondary side. I still don't see how the 2026 argument even changes this. Because these are secondary conductors you have to put protection within the limits established for those conductors. In the absence of 690.9(D), which end of the conductor do you start counting your conductor length from? I think this is the disconnect from this entire thread...we are talking about two different things here, primary and secondary protection.

Otherwise, to answer questions you asked in the early portions of the thread, you could theoretically have a scenario where the conductors are protected but the transformer is not - the code pretty clearly makes a distinction between the things you are protecting, conductors, busbars, and the transformer, are all considered separately in the design. Step 1, no transformer secondary protection is required, step 2 PV is not considered primary so no primary protection required on the PV side of the transformer, step 3 conductors are rated for X amps - OCPD is rated for X amps unless the transformer meets the specific exemptions that allow it to be protected by the primary, step 4 panelboard or other equipment is protected in accordance with 408.36(B).

As for everyone else's discussion about why one inverter would be different from multiples, first I believe its reading into it beyond what was intended, possibly a typo because it does in fact allow you, as written to look at each separate inverter separately, though that would be ridiculous. However, if they actually intended it to be a single inverter it would be on the basis of contributory fault current, islanding, and phase loss phantom voltage issues that become exponentially more problematic with multiple inverters. If that were to be the problem I would say they likely would have done a catchall like with busbar sizing that allows the engineer to calculate that out.

 

[wwhitney]

The purpose of 690.9(D) was to provide guidance on the application of 450.3 only.

OK, with further reflection, I guess the point is that 450.3 effectively requires you to have a single primary side OCPD. So if you have to consider the inverter side as primary, and have multiple inverters, that would force you to install a single set of inverter-side conductors on the transformer going to a single OCPD, which then connects to the multiple inverters.

2020 NEC 690.9(D) Exception says you don't have to do that if the sum of the ratings of the inverters does not exceed the transformer rating (only reasonable way to read it for multiple inverters). Now 2023/2026 NEC 705.30(F) says you never have to do that. So it's less restrictive.

But it seems to me that 240.21(C) always applied, so PV did not provide any extra allowances for omitting secondary conductor OCPD.

Cheers, Wayne

 

[SteveO NE]

But it seems to me that 240.21(C) always applied, so PV did not provide any extra allowances for omitting secondary conductor OCPD.

Yeah, I didn't think anyone was actually saying that and it was as disconnect in the thread. The conductor needs to be protected; the question is from what end of the conductor, and how. For instance, if you look at a branch circuit it's obvious that your 10' tap rule is talking about from the panelboard its fed from, not 10' from the utilization equipment...that would be silly. When you add source on both sides, its not so obvious, which side of the conductor needs protection? When those are considered secondaries for the purpose of OCPD the unprotected conductors are the conductors between the transformer and the first means of protection on that run of conductors, or if otherwise allowed it is protected from the primary side of the transformer.

The reason this existed in the first place was silly arguments that used to take place. I think it was under 2011 that I had an inspector require we put OCPD within 10' of the transformer, and again on the other end of the line treating the inverters as the primary source on the conductor because the code left it up to interpretation of first consider it as primary then consider it as secondary during the design process. This was always silly because the contributory fault current from the PV inverters is marginal, but despite being silly was not obvious to those that don't understand the system to that level which is the case for many AHJ's not primarily electrical and just blindly regurgitating code as a prescriptive permissive design manual that it was never intended to be.

 

[ggunn]

OK, with further reflection, I guess the point is that 450.3 effectively requires you to have a single primary side OCPD. So if you have to consider the inverter side as primary, and have multiple inverters, that would force you to install a single set of inverter-side conductors on the transformer going to a single OCPD, which then connects to the multiple inverters.

The inverter side of a 480/277V wye to 208V delta interconnection transformer is the secondary.