PV system transformer

[ggunn]

I have managed to confuse myself about transformers in a PV application; I have a couple of questions.

The basic design is this 480/277V PV inverter => 480V to 208V transformer => unfused PV AC disconnect (outside) => 208V backfed breaker in MDP (in electrical room).

I have read through NEC 450, but most of the OCP information in it is about autotransformers (this isn't) and secondary ties (this isn't). It seems to me that I don't need any more OCP on either side of the transformer because the primary is connected to the service through a breaker and the secondary is connected to a current limited source; is that correct?

Also, the transformer manufacturer rep is recommending to the project manger that we use a wye to wye transformer; I normally use a 208V delta to 480V wye transformer for this kind of installation, largely due to discussions I have participated in on this forum. Is the transformer guy out to lunch?

 

[jim dungar]

NEC 450 is basically about primary side protection sizing, it requires a minimum of 125% but allows a larger size if appropriate secondary protection is provided.
For the most part, transformer secondary protection occurs once the secondary side conductors are protected per 240.21(C).

My guess is the manufacturer rep is not well versed in power systems created by PV inverters and is questioning how you can supply 120V loads without a secondary side neutral point.

 

[wwhitney]

As a starting point, for a PV inverter with, say, a maximum 100A continuous output current, at 480Y277V (i.e. it requires a neutral reference), we could instead design the rest of the system (transformer, feeders) for a 100A continuous 480Y277V load. What changes if any would be needed when substituting the PV inverter for the load? Does the transformer have to be labeled either for backfeed or for bidirectional energy transfer? Otherwise I can't think of any.

Cheers, Wayne

 

[jaggedben]

...
It seems to me that I don't need any more OCP on either side of the transformer because the primary is connected to the service through a breaker and the secondary is connected to a current limited source; is that correct? ...

As far as code requirements there's 705.30(C) in the 2020 code, which was very much changed from its counterpart 705.65(B) in previous codes. To my reading the 2020 language seems to agree with your quote above. The previous versions of the code require you to look at both sides as if they are both primary and secondary, which seems considerably more complicated.

 

[synchro]

As a starting point, for a PV inverter with, say, a maximum 100A continuous output current, at 480Y277V (i.e. it requires a neutral reference), we could instead design the rest of the system (transformer, feeders) for a 100A continuous 480Y277V load. What changes if any would be needed when substituting the PV inverter for the load? Does the transformer have to be labeled either for backfeed or for bidirectional energy transfer? Otherwise I can't think of any.

Unless directional relays are involved, overcurrent devices will typically not respond any differently whether the current is in-phase or out-of-phase with the voltage. Transformers, of course, are also bidirectional. They are typically asymmetrical for inrush because the primary windings are placed over the secondary windings, but that's a different issue.

I could see a battery charger for an electric vehicle behaving similarly to an inverter in the sense that it draws a controlled amount of current from the POCO supply, for the most part independently of variations in the POCO voltage as long as they're within the operational range . It's just that the direction of the current flow on the inverter is 180° from that of a normal load.
I think the direction of energy transfer is important for system considerations. But ampacity of conductors is fully specified by an RMS current level, and so the direction of AC current flow at any given point in time does not need to be considered (as it will never get past that "S" in RMS).

The basic design is this 480/277V PV inverter => 480V to 208V transformer => unfused PV AC disconnect (outside) => 208V backfed breaker in MDP (in electrical room).

... It seems to me that I don't need any more OCP on either side of the transformer because the primary is connected to the service through a breaker and the secondary is connected to a current limited source; is that correct?

Also, the transformer manufacturer rep is recommending to the project manger that we use a wye to wye transformer; I normally use a 208V delta to 480V wye transformer for this kind of installation, largely due to discussions I have participated in on this forum. Is the transformer guy out to lunch?

As Wayne brings up, the question is whether an inverter behaves sufficiently differently from passive load that overcurrent protection requirements with a transformer would be any different. After all, equipment may have internal breakers or fuses but that does not eliminate the requirement for overcurrent protection on the wye secondary of a delta-wye.

Just some food for thought.

 

[ggunn]

As Wayne brings up, the question is whether an inverter behaves sufficiently differently from passive load that overcurrent protection requirements with a transformer would be any different. After all, equipment may have internal breakers or fuses but that does not eliminate the requirement for overcurrent protection on the wye secondary of a delta-wye.

Well, yes, it does. A PV inverter has a maximum amount of current that it can source on the secondary side of the transformer, and the transformer itself and conductors between it and the inverter are sized to be able to safely handle current in excess of what the inverter can supply. 705.30(C) says OCPD is not needed; thx to jaggedben for the reference..

 

[Mchapman]

We did a similar project. It had (5) 60A 277/480V 3PH inverters. These inverters ran to a combiner panel. The combiner panel fed a step down transformer which then fed a fused disconnect on the secondary side. Then fed into the building MDP panel. The transformer did need to be wye-wye because the inverters required a neutral reference point. In our case we had primary side & secondary side OCP per table 450.3B, however I think you could get by with primary side only OCP. In this case you will have a step down transformer with the inverters being considered the primary side.

 

[jaggedben]

We did a similar project. It had (5) 60A 277/480V 3PH inverters. These inverters ran to a combiner panel. The combiner panel fed a step down transformer which then fed a fused disconnect on the secondary side. Then fed into the building MDP panel. The transformer did need to be wye-wye because the inverters required a neutral reference point. In our case we had primary side & secondary side OCP per table 450.3B, however I think you could get by with primary side only OCP. In this case you will have a step down transformer with the inverters being considered the primary side.

FWIW...

Probably best to avoid using the terms 'primary' and 'secondary' in these situations, especially without defining them first. Otherwise it confuses a lot of people. Also the new 2020 code language I referenced above now says to consider the side with the higher available fault current to be the primary, so utility is still primary in on-grid applications. My understanding from this forum is that from a practical and listing standpoint what really matters is which side the transformer is energized from (i.e. it should be energized on the primary side).

 

[pv_n00b]

As far as code requirements there's 705.30(C) in the 2020 code, which was very much changed from its counterpart 705.65(B) in previous codes. To my reading the 2020 language seems to agree with your quote above. The previous versions of the code require you to look at both sides as if they are both primary and secondary, which seems considerably more complicated.

I had not looked at it critically but it looks like under the 2020 NEC 705.30(C) overrides 450.3 and does not require secondary protection even if 450.3 calls for it as long as the transformer is rated for the PV current.

 

[pv_n00b]

We did a similar project. It had (5) 60A 277/480V 3PH inverters. These inverters ran to a combiner panel. The combiner panel fed a step down transformer which then fed a fused disconnect on the secondary side. Then fed into the building MDP panel. The transformer did need to be wye-wye because the inverters required a neutral reference point. In our case we had primary side & secondary side OCP per table 450.3B, however I think you could get by with primary side only OCP. In this case you will have a step down transformer with the inverters being considered the primary side.

A Dyg transformer with the WYE facing the inverters provides a neutral reference. There is no reason to go Ygyg on systems smaller than around 1MW. Using a Ygyg transformer can introduce problems and is only done to allow OCPD on the secondary side to detect ground faults on the primary side. Dyg transformers will blind the secondary OCPD from detecting primary side ground faults.
A lot of people mistakenly believe that a 4-wire service has to be connected to a WYE transformer winding. All commercial transformers that provide 208/120V distribution in office buildings from 480/277V services are Dyg.

 

[winnie]

Is the inverter designed to directly power 277V L-N loads, or is it a 480V output inverter which requires a neutral for reference only?

Will the PV system be expected to power the customer with the mains disconnected?

A wye-wye transformer couples L-N voltage on one side to L-N voltage on the other side. A wye-wye transformer will _require_ neutral connections on both sides; rather unlike a delta-wye transformer used in 'reverse', where connecting the neutral to the wye 'primary' could cause huge circulating currents and problems.

Sometimes utilities _require_ wye-wye transformers for their 'effective grounding' requirements.

-Jon

 

[ggunn]

Is the inverter designed to directly power 277V L-N loads, or is it a 480V output inverter which requires a neutral for reference only?

The only load powered directly by the inverter is a 3P-20A breaker for monitoring. I run a minimum sized neutral to the solar load center where that breaker is but the inverters themselves do not require a neutral at all; they reference voltage to the EGC.

Will the PV system be expected to power the customer with the mains disconnected?

No.

Sometimes utilities _require_ wye-wye transformers for their 'effective grounding' requirements.

I have only encountered that once.

 

[winnie]

IMHO if the inverter cannot supply 277V L-N loads, then connecting the inverter to a wye:wye transformer will somehow permit the inverter to supply 120V L-N loads. The proposed wye:wye transformer turns 120V L-N loads into 277V L-N loads. Thus the thinking suggested in post 2 (attributed to the hypothetical mfg. rep) is not relevant.

Barring utility input requiring such, I don't see any reason to use a wye:wye transformer. I don't see a problem either...this is not the common mistake of using a delta:wye transformer in reverse.

-Jon

 

[pv_n00b]

While sometimes utilities require a Ygyg xfmr outside of that they should be avoided. There's a great old transformer handbook out there on the internet called The Whys of The WYEs that talks about the problems. It's an old GE manual and worth a read.