Adding a back up 8k Generator, want to get the N-G situation right

[OffGridForGood]

I show a bonding screw, with the implication being that it is there and can be removed, but I have no confirmation of that on this inverter.

yes four bonding screws, two on each inverter, one on each half.

 

[Zwy]

With the way the Reliance handles the N-Grid to N-Offgrid, correct.
Re-do the test:
Inverter ON, battery bank ON,
Disconnected AC-IN (all four terminals, G L1 L2 N)
Turn off Loads (off grid panel) this includes the power supplied to the Reliance.
Disconnect Output N
Test Input-N terminal to Output-N terminal = 0 ohms

Don't like that. Next I would connect a 120V load to inverter, reconnect output N and leave AC input disconnected. Turn on inverter and measure for any voltage at AC Input N terminal.

The Reliance transfer switch makes it difficult to determine if N is pass thru or switched due to Neutrals are in main panel. That is why you just put a single 120V load on the inverter output without any connection to main panel.

 

[Zwy]

That is generally refered to as 'stacked' inverters.

Correct.. I mistyped. It should have said the neutrals are connected in pass-through mode.

I thought maybe you misread it, just wanted to make we are all on the same page.

I do not trust any resistance/continuity measurement while the unit is turned on. The voltage present on the circuit can confuse the meter.
That is why I am looking at voltage if the system is on.

True, but if it reads no continuity it would be a simple test and he can move on. I'm not certain this switches N.

 

[FilterGuy]

I'm not certain this switches N.

It must switch neutral.

• When the unit is powered off, the neutrals are not connected indicating that while on battery the neutrals are not connected.
• Since it is not an LF inverter, there is no output transformer that creates neutral.
• When in passthrough mode, the voltage between the neutrals is almost zero, indicating that in passthrough mode the neutrals are connected.

I guess anything is possible, but everything points to it switching neutral and doing dynamic bonding.

 

[FilterGuy]

OK, after a few false starts, I think I am getting a pretty good picture of how this inverter is put together. The best way to think of it is that they jammed two single-phase, separate-neutral, dynamic bonding inverters into one package. They even duplicated the MPPT part. Here is the most likely model. (I am feeling pretty good that this is fairly accurate)



Note that both sides switch hot and neutral as well as dynamic bonding.

One interesting thing is that it ships in a configuration that gives two N-G bonds when in battery mode. The objectionable current would be entirely contained within the metal box of the inverter so it is not a big deal.... but it is not ideal either.


Plugging it into the system layout we get this:



Ironically, we have come full circle and are back to a common neutral layout with multiple N-G bonds.

I have not heard a statement from MPP as to whether common neutral is supported on this inverter, but I would be willing to bet a hefty sum that they would say "NO". However, the same quandary exists: Why would they say no? Here is what I said before and it still holds true

It seems to me that the input and output neutrals are tied together no matter what.

• When in pass-through mode, the two neutrals are tied together using the internal relay of the internal bypass switch, so that is not an issue.
• In battery mode, the output neutral gets tied to ground and the input neutral is already tied to ground so they are tied together through the ground wire.

Because of this, I am struggling to figure out how a common neutral layout is a problem. That leaves us with three options.

1) Assume they don't know what they are talking about and do it anyway.
2) Push back on the company till we get some clarity. (I have tried and given up)
3) Follow their instructions.

I am not really willing to say the manufacturer is wrong, but the first option is not as preposterous as it sounds. It could be that we are hearing from poorly trained support folk that has not taken the question back to engineering. In this scenario, the safe answer for the support staff is "NO". That way they can not be blamed if it is a problem. (This kind of thing has happened to me before, particularly when working with the US support staff for a Chinese company.)

So..... What should @OffGridForGood do? I can't answer that, but here are my thoughts on what I would do.
1) I would not build a new system that violates the manufactures instructions in this way. However.....
2) If I inherited a system that was already built with these inverters and a common neutral I would struggle with deciding what to do with it. I would probably leave it alone..... but worry about it because I can't be sure.

As far as the multiple N-G bonds go, I would remove all four of the bondings screws (As a reminder, this voids the warranty)

Note that in the system diagram above, I changed the grounding to avoid ground loops.

 

[OffGridForGood]

As far as the multiple N-G bonds go, I would remove all four of the bondings screws (As a reminder, this voids the warranty)

Note that in the system diagram above, I changed the grounding to avoid ground loops.

In summary:
Remove the four internal N-G screws from the two MPP's
Remove G from 'off-grid panel to Reliance (ie cap this end), but leave the G to the reliance from Utility panel.
Remove G-N from Generator.
Did I miss anything?

 

[FilterGuy]

Remove the four internal N-G screws from the two MPP's

Correct

Remove G from 'off-grid panel to Reliance (ie cap this end), but leave the G to the reliance from Utility panel.

If I understand the existing grounding layout, the following three ground connections need to be severed by disconnecting and capping one end with a wire nut.

* inverter 2 -> Off-grid load box (Disconnect and cap one end)
* Off-grid load box -> reliance (Disconnect and cap one end)
* Time switch -> off-grid load box (This one can/should be removed entirely.)

It does not matter which end is capped.

Note: the reason to leave the ground wire but disconnect one end is due to the NEC stating that the ground must be run with the Non-grounded conductor (The non-grounded conductor is the hot wire. The Grounded conductor is the neutral wire). This NEC rule is an excellent rule for 99% of the cases, but in cases like this, it can cause unwanted loops in the grounding circuit. Leaving the wire in place but only connecting one end is my workaround to the rule.

Also, note that everything is still grounded, but there is only one path from any place on the grounding circuit back to the system grounding electrode(S) (There are no loops)

Remove G-N from Generator.

Correct, but note that the MTS for the GenSet is NOT switching neutral. If you wire the MTS to switch neutral, the N-G jumper in the generator should remain in place.

 

[OffGridForGood]

If I understand the existing grounding layout, the following three ground connections need to be severed by disconnecting and capping one end with a wire nut.

* inverter 2 -> Off-grid load box (Disconnect and cap one end)

Agree

* Off-grid load box -> reliance (Disconnect and cap one end)

Agree

* Time switch -> off-grid load box (This one can/should be removed entirely.)

Should tie only to the metal timer box case itself - yes? Not to anything else.

It does not matter which end is capped.

Note: the reason to leave the ground wire but disconnect one end is due to the NEC stating that the ground must be run with the Non-grounded conductor (The non-grounded conductor is the hot wire. The Grounded conductor is the neutral wire). This NEC rule is an excellent rule for 99% of the cases, but in cases like this, it can cause unwanted loops in the grounding circuit. Leaving the wire in place but only connecting one end is my workaround to the rule.

Understood, meets with the code's intent, and protects the cable itself, if say some day an excavator pulls up the line out in the yard for example. Cap removes the loop. The Reliance is grounded only to the main panel. Makes sense.

Also, note that everything is still grounded, but there is only one path from any place on the grounding circuit back to the system grounding electrode(S) (There are no loops)

To Test This set up:
I can make the changes discussed and turn the system back on, confirm everything 'runs' and no faults come up on the MPP's.
Then what can I do to confirm we caught all the loops, and (perhaps more importantly) didn't leave anything ungrounded in error?

 

[FilterGuy]

BTW: For anyone following along. These are my indication of the bonding screws in the inverter



If the screws are removed, I would replace the image with a gap where the screw should be:


Note that this makes it impossible for the bonding relay to connect to ground.

If @OffGridForGood makes the changes he listed, it would look like this:

 

[FilterGuy]

Should tie only to the metal timer box case itself - yes? Not to anything else.

I tried to look up the timer but there are too many models.

If the case is metal, it needs to be grounded.
Most of the Intermatics I saw had a case that is opened up for wiring it but there is no ground bus. If this is what you have, the best way to deal with the ground is with a wire nut or a wago connector to create a pigtail with the following:

1) ground wire from the main breaker box
2) Ground wire to ground timer (if needed)
3) Ground wire to go to inverters (or to MTS when the genset is installed)

To Test This set up:
I can make the changes discussed and turn the system back on, confirm everything 'runs' and no faults come up on the MPP's.
Then what can I do to confirm we caught all the loops, and (perhaps more importantly) didn't leave anything ungrounded in error?

Good Questions.

Then what can I do to confirm we caught all the loops,

That's a difficult one. It is hard to identify loops through measurements

and (perhaps more importantly) didn't leave anything ungrounded in error?

That is easier,
1) Before powering anything on, check for continuity from the ground at the main breaker box to each metal box. With the distributed layout you have that will be difficult for the shed. For the shed, at least check to make sure there is continuity back to the timer box.

2) Once powered on but no loads, , check for voltage between each metal case and earth ground. There may be a few volts (probably under 4 or 5 volts), but if there are no loads, they should all read about the same.
.

 

[Zwy]

It must switch neutral.

• When the unit is powered off, the neutrals are not connected indicating that while on battery the neutrals are not connected.
• Since it is not an LF inverter, there is no output transformer that creates neutral.
• When in passthrough mode, the voltage between the neutrals is almost zero, indicating that in passthrough mode the neutrals are connected.

I guess anything is possible, but everything points to it switching neutral and doing dynamic bonding.

I'd rather test and determine for certain it actually switches.

 

[OffGridForGood]

Both MPP's all ground screws removed x4 , this was the hardest part of the process, they don't make it easy to see what you are doing I will tell you if your following along and considering removing this screw You Need A 4" or longer Magnetic #2 Phillips screwdriver (I had one long enough luckily). I can only imagine all hell breaking loose if I dropped one of these grounding screws down into the electronics. (if you DO drop the screw be sure to go find it before you turn anything back on!
I capped off the ground at the Reliance coming from the shop (solar) side. Left the ground to the Reliance that comes from the subpanel (which in turn comes from the main utility panel.
In the shop, disconnected one MPP-G-out to the solar off grid panel, this leaves the two MPP's grounded to each other, but only the master has ground to the solar loads panel.
I didn't see your comments on the timer ground until I got back from actually doing the changes, it turns out in my timer there is a grounding lug inside behing the control circuit so now this metal case is grounded.
Started everything up (yeah I held my breath for a few moments there) all started normally.
I checked to be sure all the cases are grounded, and they are.
The only way I could see to check for any missing loops, was to physically follow the wires, not high tech, but it all makes sense. Everything is running and seems to be fine.

 

[FilterGuy]

I'd rather test and determine for certain it actually switches.

OK. But I believe the tests that have already been done are sufficient for me to conclude they switch neutral

If it never switched neutral, then either the neutral-in and neutral-out are always connected or the inverters 'create' a new neutral in passthrough mode. Since we know that when powered off the neutral is not connected, we know it is not always connected. I guess someone could come up with an HF inverter design that creates neutral during passthrough, but that would be overly complicated and unnecessarily expensive. (Keep in mind that these are value-priced inverters) Furthermore, all indications are that these inverters are two single-phase dynamic bonding inverters that are jammed into one box. The designs for the MPP single-phase dynamic bonding inverters all do Neutral switching.

 

[OffGridForGood]

Oh and now I can (finally) connect my Generator-in to the MTS knowing the N-G bond(s) are sorted out!
That took a bit longer than I was expecting, but glad to have had all the help going through the systems and sorting out the logic.
I have been wondering about the whole set up ever since I installed the Reliance, now I know it is right.
Edit
Follow-up:
Installed the MTS and set up a test with the new 8kW genset;
Utility is on position "1" on the MTS
OFF is the centre-isolated position
Generator is on position "2"

I made up clear lables with a standard label maker. I started up the generator, let it warm up a few minutes, and plugged it into the new exterior Generator plug-in (30A 240vAC). Back inside the shop, I put the inverters into Utility First mode, with the MTS in position-1 (Utility) and let the inverters switch to pass-through, which also starts battery charging.
Then moved the MTS to the centre OFF position, (like a utiltity failure) the inverters switch instantly to battery/inverting mode (even my lap top stays on).
Finally, switched the MTS to position-2 (Generator) and the Inverters automatically switched over to pass-through again, but this time it is the generator supplying the energy. The generator display shows 29A output, Tested all the terminals and everything is correct. Putting a meter on the generator frame and shows continuous with Shop electrical grounds, all good, Neutrals are tied at the MTS all working as designed. Thanks for the help Filter Guy/Hedges and others that commented. My On-Site power system now has 4-layers of redundancy, PV/Battery/Utility/Generator. Feeling pretty bullet-proof on electrical supply.
The 30Ac property supports both my home and business, and we have a 'homestead' set up that produces most of our essential food (eggs, poultry, rabbits, pork, veggies) and we trade with some neighbours for things we don't produce ourshelves (honey, milk and beef). The electrical system completes the loop on our self-reliance for all things we consideer essential. Took 21-years to get to this point, but I wouldn't change a thing if I did it all again, the parts that didn't work perfectly the first time were no-less important learning lessons.
Next up: EV's...