Ungrounded Vs Grounded Inverters

[Electric-Light]

Every part of a PV system is 'hot', regardless, when the sun is shining and the utility side breaker is on. Optimizer systems excepted.

I meant hot as in having non-isolated ties to the grid.

Interesting.
The page you linked to has a paragraph that begins "Line commutated inverters do no provide significant fault current." With that said, that page doesn't seem to be addressing the differences between grounded and ungrounded inverters.

I don't know where your understanding of 'bare minimum filtration components' comes from. I believe that both types of inverters would have to meet the same IEEE standards for their output. Some of those TL inverters are really heavy with capacitors, too.

Minimum as in omitting everything they can get away with to lower cost.

Well, they don't have any diagrams of ungrounded inverters in that PDF. That's not a particularly useful document for this discussion.

Yes it does. Ungrounded/Grounded seems to really mean isolated vs non-isolated in this context and it shows an example of each.

Even if what you say were true, fault current would be no worse than on an AC circuit of the same rating as the inverter output. So...not more hazardous than other AC circuits.

The transformer on left is the utility feeder. The neutral(knot) is grounded by the utility You can see that everything from the transformer on is directly connected. So, any part of the system coming in contact with grounded structure has the full fault current of the feeder available.

Also, a fault in the static converter can energize the grid side connection with battery voltage.

If you used a transformer with only your PV connected to it, the transformer servers as the galvanic isolation between point of coupling and solar.

 

[GoldDigger]

The transformer on left is the utility feeder. The neutral(knot) is grounded by the utility You can see that everything from the transformer on is directly connected. So, any part of the system coming in contact with grounded structure has the full fault current of the feeder available.

Also, a fault in the static converter can energize the grid side connection with battery voltage.

If you used a transformer with only your PV connected to it, the transformer servers as the galvanic isolation between point of coupling and solar.

That diagram looks to me a lot more like a DC power supply feeding a load than a DC to AC inverter feeding power into the grid. :huh:
If so, what relevance does it have to the initial discussion?

 

[jaggedben]

Yes it does.

It shows a very basic and incomplete diagram to explain why a non-isolated inverter cannot have a DC grounded conductor. It's not an actual diagram of an inverter.

If you want an actual diagram of a non-isolated inverter try ABB datasheets.
https://library.e.abb.com/public/1221f006ba594baa906ddd8f0b754f58/PVI-5000-6000-Datasheet.pdf
There are more here, including isolated ones.
http://www.abb.com/abblibrary/DownloadCenter/?showresultstab=true&categoryid=9AAC172308

Minimum as in omitting everything they can get away with to lower cost.

See diagrams above for what's actually in a typical inverter.

Ungrounded/Grounded seems to really mean isolated vs non-isolated in this context and it shows an example of each.

Non-isolated must be ungrounded. Isolated can be grounded or ungrounded.

That diagram looks to me a lot more like a DC power supply feeding a load than a DC to AC inverter feeding power into the grid. :huh:
If so, what relevance does it have to the initial discussion?

What he said.

 

[Electric-Light]

That diagram looks to me a lot more like a DC power supply feeding a load than a DC to AC inverter feeding power into the grid. :huh:
If so, what relevance does it have to the initial discussion?

It looks similar, but they're not diodes. They're a set of SCRs that commutate DC source facing "into" the grid using the AC line to set the triggering like a light dimmer so DC side can drain into the grid. Did you look at the drawing here?
https://books.google.com/books?id=qW_OBQAAQBAJ&pg=PA807

It shows a very basic and incomplete diagram to explain why a non-isolated inverter cannot have a DC grounded conductor. It's not an actual diagram of an inverter.

If you want an actual diagram of a non-isolated inverter try ABB datasheets.
https://library.e.abb.com/public/1221f006ba594baa906ddd8f0b754f58/PVI-5000-6000-Datasheet.pdf

Which is just like the diagram I posted, except more complicated. The AC side is grounded at the utility side unless you have a very unusual service.

Non-isolated must be ungrounded. Isolated can be grounded or ungrounded.

Yeah, because, the DC side will short out to the ground.

 

[GoldDigger]

It looks similar, but they're not diodes. They're a set of SCRs that commutate DC source facing "into" the grid using the AC line to set the triggering like a light dimmer so DC side can drain into the grid. Did you look at the drawing here?

Why then does the diagram not show a DC source, only a LOAD?
You can use commutated SCRs instead of diodes to implement a bridge rectifier too.

 

[Electric-Light]

Why then does the diagram not show a DC source, only a LOAD?
You can use commutated SCRs instead of diodes to implement a bridge rectifier too.

Dude, the second one does. You apparently did not look.

Anyways, the wording makes thing more complex. If they just say non-isolated, and isolated, it would be less confusing..

 

[GoldDigger]

Dude, the second one does. You apparently did not look.

There is only one drawing in post #121, and that is what I was referring to.
Adding more drawings later does not make that one any less wrong in its original context.

 

[Electric-Light]

There is only one drawing in post #121, and that is what I was referring to.
Adding more drawings later does not make that one any less wrong in its original context.

That one.
https://books.google.com/books?id=qW_OBQAAQBAJ&pg=PA807

It makes no difference in that the "DC" side can fault, because the system is not isolated.

 

[jaggedben]

Yeah, because, the DC side will short out to the ground.

It's really the AC side that has that problem. Another way to look at it is that it's a problem with any grounded system. (Why do we ground any conductors again? Not everyone does. Tradition and economics.)

 

[SolarSavage]

You guys sure don't disappoint.

Thanks for all the great information. I'm going to share this with the rest of my team. Knowledge is power!

 

[Zee]

In non-engineer speak...
[for transformerless inverters]...
..... any contact of any part of wiring on the solar side with a grounded surface can fault

I don' t think this is true.
In an ungrounded system Negative and positive "float" .......they have no potential to ground. They are ungrounded.
They have voltage, potential, only to each other. pos+ to neg-.

Am i right here?

This is why, I think during installation a transformer-less system is inherently MUCH safer.

On the other hand, In a grounded PV system, aka transformer situation, all the racking is already charged either + or -. A short from the other conductor could cause a shock.

 

[jaggedben]

I don' t think this is true.
In an ungrounded system Negative and positive "float" .......they have no potential to ground. They are ungrounded.
They have voltage, potential, only to each other. pos+ to neg-.

Am i right here?

When the inverter is not turned on there's no voltage to ground. (Except for a tiny bit that leaks capacitively through glass and frames, and forget I said that. )

I believe that when the inverter is turned on there is a potential of both conductors to ground, but only because the AC neutral is grounded. And since the completion of the circuit would flow through either loads or the utility side transformer, it's not a short circuit situation.

At least, that's the best I can dope out on the back of an envelope.

The key is that the inverter connects the DC side in an opposing direction to that of the utility voltage. So even though there's no transformer separation the DC conductors really can't fault to ground with AC voltage, because the PV is pushing against that. Besides which, the GFDI mechanisms will shut it down real fast.

This is why, I think during installation a transformer-less system is inherently MUCH safer.

Yeah. During operation, too.

 

[Engineer#4827502]

In a single ground fault in an ungrounded array, no current should flow through ground. If two (or more) ground faults occur, the faults can escalate into a short circuit or arcing situation. Any ground fault in an ungrounded array (or grounded array) needs to be detected and dealt with.

Also with the necessary ground fault detection systems on an ungrounded array, the system is not blind to ground faults on either of the poles. In a typical fused ground fault detection system on a grounded array, the system is blind to ground faults on the grounded conductor. This scenario has caused fires before.

 

[Electric-Light]

I don' t think this is true.
In an ungrounded system Negative and positive "float" .......they have no potential to ground. They are ungrounded.
They have voltage, potential, only to each other. pos+ to neg-.

Am i right here?

This is why, I think during installation a transformer-less system is inherently MUCH safer.

On the other hand, In a grounded PV system, aka transformer situation, all the racking is already charged either + or -. A short from the other conductor could cause a shock.

The terms non-isolated vs isolated are better description than grounded vs not. A static converter is somewhat like an auto-transformer. You can't really "unground" a grounded system with a static converter. Attached picture gives you a general idea of how its wired up in a non-isolated inverter.

For example take a corner grounded 480v source and feed it into a tapped single winding transformer tapped like:
0-240-360-480v.

You can build a platform with a steel plate on a bunch of 5 gallon buckets and feed it 120-N-120 power from it by bringing wires from 240, 360 and 480 and designating 360 as the mid point and bond the 360 to the metal plate. You'll have 120-N-120 power on the platform, but any part on the platform can fault to the real ground.

An isolated system is separated mechanically(motor gen sets) or magnetically (transformers) and do not fault to the real ground and the ground point do not relate to the primary side.

When the inverter is not turned on there's no voltage to ground. (Except for a tiny bit that leaks capacitively through glass and frames, and forget I said that. )

I believe that when the inverter is turned on there is a potential of both conductors to ground, but only because the AC neutral is grounded. And since the completion of the circuit would flow through either loads or the utility side transformer, it's not a short circuit situation.

At least, that's the best I can dope out on the back of an envelope.

The key is that the inverter connects the DC side in an opposing direction to that of the utility voltage. So even though there's no transformer separation the DC conductors really can't fault to ground with AC voltage, because the PV is pushing against that. Besides which, the GFDI mechanisms will shut it down real fast.

Yeah. During operation, too.

You could hang a metal flashlight off a power line. Flashlight continues to operate, but if any part of the flashlight comes in contact with the ground, there will be a fault.

In a single ground fault in an ungrounded array, no current should flow through ground. If two (or more) ground faults occur, the faults can escalate into a short circuit or arcing situation. Any ground fault in an ungrounded array (or grounded array) needs to be detected and dealt with.

Also with the necessary ground fault detection systems on an ungrounded array, the system is not blind to ground faults on either of the poles. In a typical fused ground fault detection system on a grounded array, the system is blind to ground faults on the grounded conductor. This scenario has caused fires before.

Just like your GFCI. If there's a short between neutral and ground in the extension cord plugged into the GFCI, part of the current will not return by neutral and trip it.

 

[jaggedben]

The terms non-isolated vs isolated are better description than grounded vs not.

They describe two different things. When Zee says he prefers ungrounded, he means exactly that. With respect to the point he's making, the fact that most such ungrounded systems are non-isolated is a coincidence. (There's no iron fast technical reason they have to be, either.)

...

For example take a corner grounded 480v source and feed it into a tapped single winding transformer tapped like:
0-240-360-480v.

You can build a platform with a steel plate on a bunch of 5 gallon buckets and feed it 120-N-120 power from it by bringing wires from 240, 360 and 480 and designating 360 as the mid point and bond the 360 to the metal plate. You'll have 120-N-120 power on the platform, but any part on the platform can fault to the real ground.

What's your point, exactly?

You could hang a metal flashlight off a power line. Flashlight continues to operate, but if any part of the flashlight comes in contact with the ground, there will be a fault.

Again, what's your point?

 

[jaggedben]

In a single ground fault in an ungrounded array, no current should flow through ground. ...

Seems to me there's a return path through the grounded AC conductor and the loads or transformer, as well as the other side of the inverter. But I don't know what really happens through the power electronics of the inverter.

What I am fairly certain of is that utility current does not fault to ground against the polarity of the PV array. The current has to flow through the inverter or PV panels or both. That's not a short circuit fault to ground.

 

[Zee]

The terms non-isolated vs isolated are better description than grounded vs not. A static converter is somewhat like an auto-transformer. You can't really "unground" a grounded system with a static converter. Attached picture gives you a general idea of how its wired up in a non-isolated inverter.

For example take a corner grounded 480v source and feed it into a tapped single winding transformer tapped like:
0-240-360-480v.

You can build a platform with a steel plate on a bunch of 5 gallon buckets and feed it 120-N-120 power from it by bringing wires from 240, 360 and 480 and designating 360 as the mid point and bond the 360 to the metal plate. You'll have 120-N-120 power on the platform, but any part on the platform can fault to the real ground.

An isolated system is separated mechanically(motor gen sets) or magnetically (transformers) and do not fault to the real ground and the ground point do not relate to the primary side.



You could hang a metal flashlight off a power line. Flashlight continues to operate, but if any part of the flashlight comes in contact with the ground, there will be a fault.



Just like your GFCI. If there's a short between neutral and ground in the extension cord plugged into the GFCI, part of the current will not return by neutral and trip it.

I do not understand the technical nature of your reply. It seems you know more than i do, substantially, about various electrical systems.
Also, is it possible that your technical reply doesn't address the fact i am pointing out?
Namely, do you have V from either pos. or neg. to ground, in an ungrounded PV system?

 

[Zee]

In a typical fused ground fault detection system on a grounded array, the system is blind to ground faults on the grounded conductor. This scenario has caused fires before.

excellent point. an additional reason ungr. pv is safer.

 

[Zee]

I believe that when the inverter is turned on there is a potential of both conductors to ground, but only because the AC neutral is grounded. And since the completion of the circuit would flow through either loads or the utility side transformer, it's not a short circuit situation.

At least, that's the best I can dope out on the back of an envelope.

The key is that the inverter connects the DC side in an opposing direction to that of the utility voltage. So even though there's no transformer separation the DC conductors really can't fault to ground with AC voltage, because the PV is pushing against that. Besides which, the GFDI mechanisms will shut it down real fast.

It seems you explained it well. However, you have reached the limits of my electrical knowledge. Can i go somewhere to find out? :thumbsup:
Namely: I REALLY REALLY want to know what happens when i lick my fingers and cut off an MC connector, strip the insulation and crimp a connector.... on an ungrounded array. I am serious. I had to do this just yesterday.
(OK, technically it was a grounded array that i was working on. In this case i had "ungrounded it" ....I shut off the dc disconnect (PV+) and isolated the PV- from ground, while working.
I did not do the licking fingers part.:dunce:

 

[jaggedben]

It seems you explained it well. However, you have reached the limits of my electrical knowledge. Can i go somewhere to find out? :thumbsup:
Namely: I REALLY REALLY want to know what happens when i lick my fingers and cut off an MC connector, strip the insulation and crimp a connector.... on an ungrounded array. I am serious. I had to do this just yesterday.
(OK, technically it was a grounded array that i was working on. In this case i had "ungrounded it" ....I shut off the dc disconnect (PV+) and isolated the PV- from ground, while working.
I did not do the licking fingers part.:dunce:

I think the key point about doing this on an ungrounded system is that in case a person doesn't know that one should disconnect all relevant circuits from the inverter, or they don't know how, or they forget, the task is quite a bit safer. But put the connector back on the positive lead before you strip the negative one! And test the connector to ground before you cut it off! (in case of a ground fault) And watch out for systems with combiners.

But as far as the discussion we were having above, that's really about when the inverter is operating, not when it's disconnected and shut down and the DC is isolated.