system grounding for off grid premises

Status
Not open for further replies.
Consider a structure not connected to a utility supply. Say it has a remote mounted PV array. In the structure batteries, charge controller, and inverter supplying a standard split phase panel and loads. Considering just theory and not code, what if any benefit does grounding of the split phase system provide? Please provide specific scenarios and show using electrical theory why there was a benefit to the system being grounded (if that is your position) :)
 

GoldDigger

Moderator
Staff member
Location
Placerville, CA, USA
Occupation
Retired PV System Designer
Consider a structure not connected to a utility supply. Say it has a remote mounted PV array. In the structure batteries, charge controller, and inverter supplying a standard split phase panel and loads. Considering just theory and not code, what if any benefit does grounding of the split phase system provide? Please provide specific scenarios and show using electrical theory why there was a benefit to the system being grounded (if that is your position) :)
Depending on the inverter you use, you may not be able to actually float the AC output neutral. It may be tied to the DC negative and chassis ground. In that case, unless you are willing to run an ungrounded DC system too you should explicitly ground the neutral at the panel (or rely on the ground-neutral bond at the inverter.

If you do not ground the AC neutral, I would recommend that you include a ground detector on the AC side. Otherwise a single fault could lead up to a situation where a second fault could cause serious problems.
 
Depending on the inverter you use, you may not be able to actually float the AC output neutral. It may be tied to the DC negative and chassis ground. In that case, unless you are willing to run an ungrounded DC system too you should explicitly ground the neutral at the panel (or rely on the ground-neutral bond at the inverter.

If you do not ground the AC neutral, I would recommend that you include a ground detector on the AC side. Otherwise a single fault could lead up to a situation where a second fault could cause serious problems.

It appears the magnum Ms series are isolated and have no internal system bond. I think Outback FX are, as you said, tied to batt negative (not sure if non isolated topology or not - doesnt really matter for this discussion).

So say we have two scenarios:

1) "neutral" of AC system is bonded to the equipment grounding system. Whether it is done at the inverter factory or we do it I dont think matters for this discussion and scenario. Equipment grounding system as normal everything interconnected/bonded. Say we dont (intentionally) connect to a grounding electrode. Also say that system has not incidentally become grounded by say a gas line or water line, etc. Now this would be bad because we would have the EGC system floating with one of the supply conductors and developing some degree of capacitive coupling and potential to earth which could be say a concrete basement slab we are walking on correct?

2) Ac system is ungrounded. We want to connect EGC system to earth to avoid same/similar problem in scenario 1 above (but probably less serious) correct? Now we have your basic ungrounded system but say who cares if there is a fault, it becomes a grounded system so what? Ill check it now and then and if I miss it and there is a second fault, inverter will shut down from overload and I fix the problem?

I know the laws of physics dont change when you are not utility connected, just want to take some of the issues that arise from a utility connection out of the picture to make sure I have this down completely. :)
 

GoldDigger

Moderator
Staff member
Location
Placerville, CA, USA
Occupation
Retired PV System Designer
It appears the magnum Ms series are isolated and have no internal system bond. I think Outback FX are, as you said, tied to batt negative (not sure if non isolated topology or not - doesnt really matter for this discussion).

So say we have two scenarios:

1) "neutral" of AC system is bonded to the equipment grounding system. Whether it is done at the inverter factory or we do it I dont think matters for this discussion and scenario. Equipment grounding system as normal everything interconnected/bonded. Say we dont (intentionally) connect to a grounding electrode. Also say that system has not incidentally become grounded by say a gas line or water line, etc. Now this would be bad because we would have the EGC system floating with one of the supply conductors and developing some degree of capacitive coupling and potential to earth which could be say a concrete basement slab we are walking on correct?

2) Ac system is ungrounded. We want to connect EGC system to earth to avoid same/similar problem in scenario 1 above (but probably less serious) correct? Now we have your basic ungrounded system but say who cares if there is a fault, it becomes a grounded system so what? Ill check it now and then and if I miss it and there is a second fault, inverter will shut down from overload and I fix the problem?

I know the laws of physics dont change when you are not utility connected, just want to take some of the issues that arise from a utility connection out of the picture to make sure I have this down completely. :)

1. Whatever else you do or do not do, the EGC must be grounded, whether it is bonded to the neutral or not. This is to provide an equipotential bond between exposed metal surfaces of equipment and "earth ground" which may appear on a variety of places from water pipes to concrete floors. You can float the power conductors, but you must never float the EGC. In the case of an RV insulated by rubber tires and with no external connections, not even water, the EGC will be bonded to the vehicle chassis, just like the negative of the 12V electrical system. (Even if you use a different floating battery bank for house loads.)

2. Whether you bond the AC neutral or just check it periodically for a ground fault condition probably has little effect on the safety of the system. NEC requires that any power source which can be configured with less than 150V from all ungrounded conductors to ground be grounded. You can decide whether you think the NEC applies to an off grid house, but the majority will say that it does if it has be adopted by the jurisdiction the house is located in. There may not, however, be any inspection requirements. :)
 
1. Whatever else you do or do not do, the EGC must be grounded, whether it is bonded to the neutral or not. This is to provide an equipotential bond between exposed metal surfaces of equipment and "earth ground" which may appear on a variety of places from water pipes to concrete floors. You can float the power conductors, but you must never float the EGC. In the case of an RV insulated by rubber tires and with no external connections, not even water, the EGC will be bonded to the vehicle chassis, just like the negative of the 12V electrical system. (Even if you use a different floating battery bank for house loads.)

2. Whether you bond the AC neutral or just check it periodically for a ground fault condition probably has little effect on the safety of the system. NEC requires that any power source which can be configured with less than 150V from all ungrounded conductors to ground be grounded. You can decide whether you think the NEC applies to an off grid house, but the majority will say that it does if it has be adopted by the jurisdiction the house is located in. There may not, however, be any inspection requirements. :)

I like the ungrounded system. Perhaps the most common cited disadvantage of it is that it requires action to fix a fault - not automatic like a grounded system which is a good for joe and suzy homeowner. However, as I propositioned before, who cares if there is a first fault? I have anyways been used to thinking in terms of the quintessential example of a facility using an ungrounded system where it is imperative that power is not lost and the importance of the "get out of fault free" card. But what about for a "normal" structure? The problem I see in practice is not really a problem with the ungrounded system itself but our existing infrastructure: If we had different first faults at different structures that were served from the same transformer well that would just be a disaster. But if every structure had its own transformer, I think I prefer the ungrounded system. How does the rest of the world do it? I assume each structure needs its own transformer - no double dipping for the utility - and how are first faults handled?
 

GoldDigger

Moderator
Staff member
Location
Placerville, CA, USA
Occupation
Retired PV System Designer
Just for grins, I recall mention of a study done in a country (Norway?) where ungrounded, one service per transformer, was standard.
Something like 50% of the sites checked had a first fault of long standing, which removed all of the purported safety advantages of the ungrounded system.
 

dereckbc

Moderator
Staff member
Location
Plano, TX
Preferable one would not want to have the DC side grounded, but is nearly impossible to implement because most charge controllers and inverters have the negative polarity bonded to the chassis.
In my opinion is a huge mistake manufactures make. Doing so forces normal DC load currents to flow on bonding conductors, and in some cases incompatible with solar panels that have positive bonded to the frames of the panel.
 

Tony S

Senior Member
I like the ungrounded system. Perhaps the most common cited disadvantage of it is that it requires action to fix a fault - not automatic like a grounded system which is a good for joe and suzy homeowner. However, as I propositioned before, who cares if there is a first fault? I have anyways been used to thinking in terms of the quintessential example of a facility using an ungrounded system where it is imperative that power is not lost and the importance of the "get out of fault free" card. But what about for a "normal" structure? The problem I see in practice is not really a problem with the ungrounded system itself but our existing infrastructure: If we had different first faults at different structures that were served from the same transformer well that would just be a disaster. But if every structure had its own transformer, I think I prefer the ungrounded system. How does the rest of the world do it? I assume each structure needs its own transformer - no double dipping for the utility - and how are first faults handled?

I?ve only worked on four ungrounded systems, the smallest being 2.6MVA. If you think I?d hang about for the 2nd fault you?d be sadly mistaken. Ground fault detection shut the system down.

The reason for the system being ungrounded? We used water cooled cables on the 660V side.
 
I?ve only worked on four ungrounded systems, the smallest being 2.6MVA. If you think I?d hang about for the 2nd fault you?d be sadly mistaken. Ground fault detection shut the system down.

The reason for the system being ungrounded? We used water cooled cables on the 660V side.

But you are talking about a very large system probably with high voltages where you wouldnt want to be around ANY fault. Everything else being equal, there is nothing different about a 2nd fault on an ungrounded system than a first fault on a corner grounded system or a line to line fault on a center point Y grounded or center tap delta grounded system.
 
Status
Not open for further replies.
Top