Infini Plus 3Ph 10KW Neutral wire bond/Join with Grid Neutral?

[LionKing]

I'm from Pakistan and we have 220V 3Ph connection with grid at our house (common in this region), however, we don't have any 3P load. Due to unstable grid, I'm getting a hybrid Infini plus 3P inverter installed. Our plan is to have only critical load (lights, ceiling fans, one AC, fridge/freezer and a TV)  on inverter and leave rest on grid. We're also planning on supplying excess power to Grid (Net Metering).

The issue is while we left provision in home wiring for a Live solar wire in each outlet i.e. one Live coming from Grid and one from Inverter, there is only one common Neutral going into each room. This usually works with UPS because they come online while grid is off. In this case, Inverter would be supplying power while Grid is on as well. 

One solar installer said that Inverter Neutral Out needs to be kept separate from Grid Neutral, the other guy said that its ok to bond them together and that they do bond Neutrals on 3Ph Generators as well and its never an issue. If we're able to bond the neutral that'll simplify the installation a lot since we can only have select outlets on solar - by just swapping the Live grid wire with Live solar. But I don't wanna take a risk. 

Please advise.

Edited June 12, 2018 by LionKing

[Chris Hobson]

You should have only two wires to each supply point as what you describe is 3P but your appliances are 1P.

So you should have 3P + N coming to your distribution board. Non critical loads (1P + N) and the inverter (3P +N) should all be supplied with grid power . Your critical loads should then be split up on the 3 phases of the inverter (1P + N). Your critical loads will receive grid power via the inverter. Having grid power and inverter power at power point is asking for trouble.

Your inverter will seamlessly switch from grid to solar and the only way you will be aware there is a power failure is your non critical loads will be off or you will be able to see on the inverter LCD.

[LionKing]

Thanks. Im still unsure though whether its ok to combine grid and inverter-out neutral? Since our house is wired for traditional UPS systems in which neutral was common between UPS and grid.

[___]

1 hour ago, LionKing said:

One solar installer said that Inverter Neutral Out needs to be kept separate from Grid Neutral, the other guy said that its ok to bond them together and that they do bond Neutrals on 3Ph Generators as well and its never an issue. If we're able to bond the neutral that'll simplify the installation a lot since we can only have select outlets on solar - by just swapping the Live grid wire with Live solar. But I don't wanna take a risk. 

There are two questions here. The one is whether it is legal to share the Neutral. The other is whether equipment would be damaged.

To the first question, in South Africa it is not legal to share the Neutral. I don't know about other countries, but there is usually a good reason behind the illegality of something.

To the second, at least one forum member here lost a UPS because the Neutral wasn't properly switched.

If I think about it a bit more, one good reason would be that the neutral wire will carry up to double the normal current if you share it. While the grid is on, your UPS-line is in phase with the non-UPS line and the neutral wire is now a shared neutral that has to carry all the return current for the other two. Even if you size the wires to account for this, you have the danger of a lost neutral.

Consider the worst case scenario where you lose only the neutral, and the inverter switches to backup mode. Now you have two live lines with various appliances across them and nothing in the middle holding it down. This could result in peak voltages as high as 800V across the two. With a suitably low impedance appliance on the grid side you could potentially channel the grid in a circle into the AC side of the inverter.

I'd say don't do it. Just save yourself the worry and don't do it.

[pilotfish]

5 hours ago, LionKing said:

The issue is while we left provision in home wiring for a Live solar wire in each outlet i.e. one Live coming from Grid and one from Inverter, there is only one common Neutral going into each room. This usually works with UPS because they come online while grid is off. In this case, Inverter would be supplying power while Grid is on as well. 

I see smoke on the horizon

• You should only have 1x live and 1x neutral (and 1 by earth) to each outlet, these are wired back to your Inverter DB.
• The inverter DB is supplied by your inverter.
• The inverter will send grid power to this DB when in Grid mode, or Inverter power when in battery mode.

Good luck!

[Guest]

3 hours ago, plonkster said:

... one forum member here lost a UPS because the Neutral wasn't properly switched.

Jip. That's me. Inverter being a online inverter, meaning the loads are ALWAYS on batteries, had it's charger blown to smithereens, literally.

EDIT: Online saved my / Electricians bacon, for if any other type of UPS, ALL the electronics would have gone to the scrapheap. And there was a LOT of sensitive stuff on them circuits.

So keep the inverter neutral FAR FAR (... Galaxy FAR) away from Eskom's neutral.

Like on a changeover switch for simple lights, where the electrician did not listen (understand), you better switch live AND neutral, not just live, as they are used to do with just 1 source like say Eskom.

Anyone brave enough to share that neutral between Eskom and inverter, will make me lose my no 1 position on the "Do not Do That or Set the Smoke Free!" board. 

Edited June 8, 2018 by Guest

[LionKing]

Appreciate your feedback guys. I respect your opinions but locally I'm getting different ones. The solar guy and electrician bonded Grid and inverter-out Neutral wires, they were insistent that its very common here to do it this way and they've never had any issues. Solar guy said they do this even with Net metered setups and its fine. He used to work with some big Solar company in Saudi Arabia - although I doubt they bonded Grid & Inverter Neutrals.

My uncle is an Engineer at my Grid/Utility company (Wapda). I spoke with him and he also said that there shouldn't be an issue because the Neutral is grounded in the Transformer on each street and it doesn't go back to the Grid, so he doesn't see why there would be an issue. 

Its been 3 days and due to unstable grid...we've had grid outage several times (while on solar - daytime, and also on battery - nighttime). Couple of times there were 2 phases with low voltage - the inverter didn't sync with Grid until the phase issue was resolved.

In terms of load, we've run heavy loads (AC, 2 Irons etc) on Solar and ACs on Grid connection and everything seems to be working fine. 

Can you guys share some technical document/paper that talks about Not bonding Neutral wires on Transformerless (TL) inverters? It looks like Voltronic doesn't provide technical assistance and the importer here is about as non-technical as it gets.

[pilotfish]

3 minutes ago, LionKing said:

The solar guy and electrician bonded Grid and inverter-out Neutral wires,

As an electrical contractor I would have no problem with bonding [neutral in/neutral out] at the inverter from a safety perspective. This will ensure that that your neutral is always bonded to earth potential as it should be and as it is by grid provider.

Some municipalities may have some B/S requirement about isolating the alternative energy source from the grid, but this actually compromises safety in my opinion and there is absolutely no good reason for this that I can think of.

[___]

11 minutes ago, LionKing said:

My uncle is an Engineer at my Grid/Utility company (Wapda). I spoke with him and he also said that there shouldn't be an issue because the Neutral is grounded in the Transformer on each street and it doesn't go back to the Grid, so he doesn't see why there would be an issue. 

I'm not an electrical contractor, so I would attach more weight to @pilotfish above. The reason for not allowing a second bond is that it allows working current to flow on an earth wire. The rule (at least here) says earth wire must only carry fault current.

Nevertheless, there are systems that bond earth/neutral multiple times for extra safety (it's called PME, Protective Multiple Earth), so if it was done once and once only at the output of the inverter, it's probably fine, as long as it complies with local regulations.

Why earth wires are only allowed to carry fault currents is something I do not know. Yet. The only reason I know if is the risk of a broken Neutral. Current would simply flow on the earth wire between the two bonding points and the fault would be masked. That's the only reason I know of.

This issue is different to the other question of a shared neutral (and two live wires which may be active at the same time. I still think that is a bad idea.

[pilotfish]

1 minute ago, plonkster said:

The reason for not allowing a second bond is that it allows working current to flow on an earth wire. The rule (at least here) says earth wire must only carry fault current.

Quite right, but that is not what I was talking about although my use of the word BOND may have confused the issue. I meant JOIN (rather than bond) an un-switched grid neutral to inverter neutral.

Grid neutral is at earth potential, if you connect grid neutral to inverter neutral then your inverter voltage will always be referenced to earth as 0v. Without this your voltages can float around, often with "earth" being the midpoint between live and neutral, so you will measure [N/E/L =115v 0v 115v] whereas you should measure [N/E/L = 0V 0V 230V]

Hope that makes sense.

[LionKing]

16 hours ago, pilotfish said:

As an electrical contractor I would have no problem with bonding [neutral in/neutral out] at the inverter from a safety perspective. This will ensure that that your neutral is always bonded to earth potential as it should be and as it is by grid provider.

Some municipalities may have some B/S requirement about isolating the alternative energy source from the grid, but this actually compromises safety in my opinion and there is absolutely no good reason for this that I can think of.

Thanks...that puts my mind to ease. We actually don't have any Neutral bonded/joined to Earth/Ground wire. For some reason it's not done here, so at the moment I only have Neutral-In & Neutral-Out joined together. 

Another trend here is Not to bond earth with metal enclosures. I just noticed that while Ground wires are appropriately connected to the inverter. The inverter chassis also has a ground wire connection, which is not being used. Should I connect it with the common Ground wire?
 

16 hours ago, plonkster said:

I'm not an electrical contractor, so I would attach more weight to @pilotfish above. The reason for not allowing a second bond is that it allows working current to flow on an earth wire. The rule (at least here) says earth wire must only carry fault current.

Nevertheless, there are systems that bond earth/neutral multiple times for extra safety (it's called PME, Protective Multiple Earth), so if it was done once and once only at the output of the inverter, it's probably fine, as long as it complies with local regulations.

Why earth wires are only allowed to carry fault currents is something I do not know. Yet. The only reason I know if is the risk of a broken Neutral. Current would simply flow on the earth wire between the two bonding points and the fault would be masked. That's the only reason I know of.

This issue is different to the other question of a shared neutral (and two live wires which may be active at the same time. I still think that is a bad idea.

I think I created confusion by using the word bond - typically its used when referring to Neutral-Ground bonding at the main electrical panel and the advise is Not to bond these at sub-panels to avoid circular currents. Basically, I was just talking about joining/combining Neutral-In and Neutral-Out.

16 hours ago, pilotfish said:

Grid neutral is at earth potential, if you connect grid neutral to inverter neutral then your inverter voltage will always be referenced to earth as 0v. Without this your voltages can float around, often with "earth" being the midpoint between live and neutral, so you will measure [N/E/L =115v 0v 115v] whereas you should measure [N/E/L = 0V 0V 230V]

Hope that makes sense.

Thanks for the explanation...that makes sense. Since grid neutral is grounded at transformer (should be at Earth Potential), I don't need to bond Grid Neutral with my home's Grounding wire at the main Panel? Basically, right now I have Inverter Neutral-In and Out joined, however, Grid neutral is Not joined with Home Earth/Ground wire at the main panel.

[pilotfish]

2 hours ago, LionKing said:

Basically, right now I have Inverter Neutral-In and Out joined, however, Grid neutral is Not joined with Home Earth/Ground wire at the main panel.

That sounds about right to me, don't change anything.

[___]

4 hours ago, LionKing said:

Basically, I was just talking about joining/combining Neutral-In and Neutral-Out.

I recall a mybroadband thread where someone listed all the options and this one came out as the best one in a situation where the inverter doesn't have a bonding relay (which at that point it didn't). I think it might have been the same @Gnome here that was also involved in that thread (linking to the middle of it here) three years ago.

I think the trouble was that the transfer switch in that inverter (which is not an Infini, it is the cheaper brother) does not disconnect the inverter's neutral when in grid mode, only the live, so you could not install a bond internally inside the inverter as it would remain in place while on grid (breaking the no-bonding-again rule).

As I recall the best solution at the time was to bridge neutral on the input/output side so that while in inverter mode earth and neutral remains bonded via the existing bond at the transformer (assuming TN-S) or where it enters the premises (assuming TN-C-S). Should work fine as long as you don't lose either earth or neutral :-)

[Gnome]

 

Quote

6.12.4 Earthing of the neutral of combined sources

When an installation that has a common neutral is supplied from a combination of transformers and generators located near one another, the neutral terminal of each of these items shall be connected to a single neutral bar. This neutral bar shall be the only point at which the neutral of the installation is connected to the consumer's earth terminal except as in the case in 7.12.3.1.3.

 

Quote

7.12.3 Earthing requirements and earth leakage protection

7.12.3.1 Neutral bar earthing

7.12.3.1.1 Protection in accordance with the requirements of 6.7 shall be provided for the electrical installation in such a manner as to ensure correct operation of the protection devices, irrespective of the source of supply or combination of sources of supply. Operation of the protection devices shall not rely upon the connection to the earthed point of the main supply when the generator is operated as a switched alternative to the main supply.

Where there is no existing earth electrode installed in the electrical installation, a suitable earth electrode may be installed in accordance with SANS 10199. When installed, the electrode shall be bonded to the consumer's earth terminal and to the earthing point on the generating set by a conductor of at least half the cross-section of that of the phase conductor, but not less than 6 mm2 copper, or equivalent. This also applies to a singlephase supply (see also 5.1.3.1).

NOTE: In a TN system earthing of electricity supply, an earth electrode is normally not required in an electrical installation.

7.12.3.1.2 In an installation that is supplied from a combination of transformers and alternative supplies located near to each other, including an alternative supply (supplies), the neutral points of each of these items shall be connected to a single earthed neutral bar (see annex S). This earthed neutral bar shall be the only point at which the neutral of the installation is earthed. Any earth leakage device shall be positioned in such a way as to avoid incorrect operation due to the existence of any parallel neutral/earth path.

7.12.3.1.3 Where alternative supplies are installed remotely from the installation, or from one another, and where it is not possible to make use of a single neutral bar which is earthed, the neutral of each unit shall be earthed at the unit and these points shall be bonded to the consumer's earth terminal (see 6.12.4). The supply from each unit which supplies the installation or part of the installation, shall be switched by means of a switch that breaks all live conductors operating substantially together (see annex S), to disconnect the earthed neutral point from the installation neutral when the alternative supply is not connected (see also 6.1.6).

7.12.3.1.4 Where only part of an installation is switched to the alternative supply in the same distribution board, the neutral bar shall be split (see figures S.2 and S.3 in annex S).

7.12.3.2 Additional requirements for installations that incorporate static uninterruptible power systems (UPS) equipment

7.12.3.2.1 The UPS shall be fitted with overcurrent protection devices designed to disconnect the output of the UPS in the case of overcurrent or
earth faults that occur in the part of the installation supplied by the UPS. The earth loop impedance, including the internal impedance of the UPS, shall comply with the requirements of 8.7.5.

7.12.3.2.2 Where a common neutral and a bypass switch are used, the part of the installation supplied by the UPS shall be provided with earth leakage protection as required in 6.7.5. (See figure S.3.)(See also 6.7.5.5(a) for exclusion relating to safety supplies.) 

 

[Gnome]

Gist of that is as follows:

Your earth and neutral should be bonded in one place, usually right in front of your property. If this is not possible then it should be bonded at the device output (Rather than floating). Floating is the worst possible case. (More on that below).

I couldn't really find info on where we stand with earthing relays, because realistically and earthing relay could fail creating a dangerous situation. Relays have two failures modes:

1) The relay no longer switches contacts

2) The contacts are welded together

Problem case 1 can be solved by having the bonding active even if the relay is not powered (that is how I have mine setup). Problem case 2 is can only be solved by not allowing the alternative supply turn on unless the ground is bonded. This is much harder to achieve but I could certainly whip up a circuit very quickly that does this.

Floating case:

If your ground floats some common household devices could actually become very dangerous. For example switch mode power supplies and electric motors often have class Y capacitors between live and ground to allow noise suppression (very common on any quality product). Because ground is floating it'll shock you (probably not kill but it is really unpleasant and I argue someone with a weak heart could end up worse off). Furthermore if any metal surface becomes live, it will no longer trip the power leading to some fun and shocking situations.

Edited June 12, 2018 by Gnome

[Gnome]

2 hours ago, plonkster said:

I recall a mybroadband thread where someone listed all the options and this one came out as the best one in a situation where the inverter doesn't have a bonding relay (which at that point it didn't). I think it might have been the same @Gnome here that was also involved in that thread (linking to the middle of it here) three years ago.

Looks like I was wrong there. I assumed bonding is illegal. But guess SANS says rather bond than don't ¯\_(ツ)_/¯

In my defense, I've browsed many electrical forums locally and nobody really is sure what is the best way. Because bonding neutral/earth can as I said there be dangerous also.

[___]

22 minutes ago, Gnome said:

class Y capacitors between live and ground

Yup, they will pass the AC. These caps are likely in the low nanofarad range, so will pass extremely small amounts of energy, but as you say it is not pleasant.

31 minutes ago, Gnome said:

Operation of the protection devices shall not rely upon the connection to the earthed point of the main supply when the generator is operated as a switched alternative to the main supply

If I read that correctly, it means my RCD must still work correctly if someone chops off my cable to the transformer, which includes the earth wire since it is TN-S. Mine isn't exactly set up to deal with that...

[Youda]

The technical reason behind issues with bonding N-in and N-out together is based on the fact that most inverters are using full H-Bridge architecture in order to create alternating current from the DC. Now, if the inverter contains an AC battery charger (many PV inverters do), then one of battery poles might have been internally connected with N-in through that charger circuit. If that's the case, then bonding N-in and N-out together effectivelly bypasses some of the IGBT's as seen on the attached picture. Result is a damage, of course. Since vendors are aware of this, they generally do not recommend bonding.

On the other hand, many inverters have the internal architecture that solves the above issue and bonding N-in and N-out is pretty safe with them:

- For example, bonding is OK for InfiniSolar 10K 3Phase and for InfiniSolar Plus 5k also.

- For InfiniSolar Plus 3K and InfiniSolar Super 4K it's not safe and might cause a permanent damage, depending on the harware revision. Same is true for some of the Axpert models and HW revisions.

Please note, that if you will bond N-in and N-out together, then you MUST NOT ground any pole of the battery. Same with PV input - you MUST NOT ground negative nor positive pole. The reason behind this is, that regardless of having TN-S or TN-C grid, the N has a direct connection with the ground somewhere (at the grid transformer or at house DB mostly). That info should be in the installation manual of the inverter, but the reality is that it's missing most of the time. There are some inverters that do allow you to connect + or - pole of the PV to the ground, but these are made specifically for use with GICS/Amorphous PV modules which require earthing for longevity.

BTW: Speaking of diesel generators, it's safe to bond one of the poles with N-grid, because there's no AC-powered charger involved.

[___]

10 minutes ago, Youda said:

Since vendors are aware of this, they generally do not recommend bonding.

It depends on the make, literally. Some can be bonded safely. Most cheap modified sine-wave things cannot.

Whether you can ground the battery and/or PV input also depends. With some MPPTs the current sensor is in the negative line, so grounding both negatives bypasses the current sensor and can potentially damage the MPPT. In general you should not ground the negative of anything (except maybe in a vehicle, and then only the battery).

Inverters with a fully isolated boost stage can be grounded both battery and AC neutral side. But why would you want to find out? :-)

[Gnome]

2 hours ago, Youda said:

The technical reason behind issues with bonding N-in and N-out together is based on the fact that most inverters are using full H-Bridge architecture in order to create alternating current from the DC. Now, if the inverter contains an AC battery charger (many PV inverters do), then one of battery poles might have been internally connected with N-in through that charger circuit. If that's the case, then bonding N-in and N-out together effectivelly bypasses some of the IGBT's as seen on the attached picture. Result is a damage, of course. Since vendors are aware of this, they generally do not recommend bonding.

On the other hand, many inverters have the internal architecture that solves the above issue and bonding N-in and N-out is pretty safe with them:

- For example, bonding is OK for InfiniSolar 10K 3Phase and for InfiniSolar Plus 5k also.

- For InfiniSolar Plus 3K and InfiniSolar Super 4K it's not safe and might cause a permanent damage, depending on the harware revision. Same is true for some of the Axpert models and HW revisions.

Please note, that if you will bond N-in and N-out together, then you MUST NOT ground any pole of the battery. Same with PV input - you MUST NOT ground negative nor positive pole. The reason behind this is, that regardless of having TN-S or TN-C grid, the N has a direct connection with the ground somewhere (at the grid transformer or at house DB mostly). That info should be in the installation manual of the inverter, but the reality is that it's missing most of the time. There are some inverters that do allow you to connect + or - pole of the PV to the ground, but these are made specifically for use with GICS/Amorphous PV modules which require earthing for longevity.

BTW: Speaking of diesel generators, it's safe to bond one of the poles with N-grid, because there's no AC-powered charger involved.

I don't really see how this is possible.

In inverter mode it is boosting the voltage to 400v DC, rectifying then run it through an H-bridge.

The boost converter is being driven through a transformer so I don't really see how the battery is connected to the circuit.

Your diagram only contains the inverter stage but I don't understand how you see that being connected to ground?

Are you saying that the buck converter is somehow always creating a path to ground?

How?

You can't just feed the 240v AC back into the battery, you either need a buck converter or you need to feed it back through that boost converter transformer. Are you saying that buck converter is not galvanically isolated? (I can't imagine that being legal)

Edited June 21, 2018 by Gnome

[___]

13 minutes ago, Gnome said:

Are you saying that the buck converter is somehow always creating a path to ground?

It depends entirely on the topology used, and whether the designer thought it best to reference all things to a common place. The long and short of it is don't make assumptions :-)

[Youda]

Gnome,

Path to the ground - yes, and it's not just the buck converter. There's a lot of other circuits in the inverter (just like plonkster mentioned in his previous post) that may cause troubles when you ground them (or connect to N). Speaking of transformers, high-efficiency inverters tend to avoid them as much as possible, because they are hindering the efficiency and rising the cost of device. Therefore, without knowing what's that particural inverter's topology (or architecture), you cannot be sure that there's galvanic isolation between the components.

Two main topologies ale Low Frequency with Transformers and High Frequency transformerless. For a shame, both can be manufactured as non-isolated, so this won't help you much.

Even in the devices where there's galvanic isolation between the battery and the high-voltage bus, there's no galvanic isolation between the grid-in and inverter-out. Just imagine how big chunk of metal is a 10KW 3P transformer....

So, whenever you want to bond N-in and N-out, you have to check the device capabilities first. That's why some PV installers say that it's okay, while others are strongly against.

@plonkster, by the way, how about Victron and neutrals bonding?

[___]

2 hours ago, Youda said:

High Frequency transformerless

I hate this term. They still have a transformer, they can just get away with a much smaller component and use a different material (eg Ferrite, so no eddy current issues in the core) because of the higher frequency.

Basically, if you study the various conversion topologies, there are a few low-power designs that use only capacitors (also called charge pumps), but they are generally good for generating a high voltage supply to drive FETs at most. For anything that can do any level of power, you'll want to use some kind of magnetic component, some kind of inductor. Now there are numerous such converters, buck, boost, sepic, cuk, and they all use some kind of magnetic component, and almost always there is an equivalent isolated version that essentially uses a secondary winding on the inductor: aka a Transformer.

So anyone talking about it being "transformerless" must be in marketing. They are also the idiots who think the inverters with bigger transformers (MLT, Victron) are old tech... the truth is these inverters use toroidal transformers for better efficiency.

2 hours ago, Youda said:

@plonkster, by the way, how about Victron and neutrals bonding?

Zero problems. Designed in in fact. It has a bonding relay. When the power goes out, it bonds the neutral to the case of the inverter. All you have to do is ground the inverter case to a suitable earth. For example, on a boat/barge/yacht it would be grounded to the hull while in inverter mode, but it would use the shore earth when on the grid. When you install it in the city, you simply earth the case of the inverter and you're done.

[Guest]

29 minutes ago, plonkster said:

When you install it in the city, you simply earth the case of the inverter and you're done

If one where to connect the Victron inverter casing to the earth in the DB board, would that be acceptable?

[___]

1 minute ago, The Terrible Triplett said:

If one where to connect the Victron inverter casing to the earth in the DB board, would that be acceptable?

I'm not sure, but that is what I have done. There are some that say all earths must be equipotential and connected together. If your grid connection is TN-S and you're not allowed to make a second earth at the premises (aka PME), you really have no choice. And my excuse is that a qualified electrician installed it that way.

There is however something in the rules somewhere that says it should remain safe, or more specifically that things such as your RCD must continue to work correctly, even if you lose the earth from the transformer. You could do this by having a good local earth spike and connecting the case to that. That means that while in invert mode you always have your own local earth.

So my gut feeling is that both ways are acceptable, but since I'm not an electrician I cannot say precisely. Both of them are certainly safe, the former just has the slight risk of a broken earth.

With a broken earth you still have RCD protection, it's just not as sensitive because the earth fault loop impedance is high. You essentially have an IT-system. It is possible that it actually complies with the rule that things should continue to work. I don't know :-)