[RE-wrenches] N-G bond for supply side connection Midnite Power

Wed Oct 29 06:41:16 PDT 2025

William this is really good information, thanks for sharing the technical
perspective.

I have two sites with two of these inverters and both of them are
exhibiting very similar behavior. On Friday, I was out at the other one to
wrap up a few things. While I was out there I verified some of the things
that Midnite tech support suggested like making sure that the AC wires to
each inverter's grid in and cables are the same length and also from the
battery bus bar to each inverter are the same length. This was actually a
bit surprising to me, considering with lead acid it is more important to
have the battery cables from the bus bar to the batteries be the exact same
length and the cables to the inverters did not matter as much.

Based on what we were talking about last week, I removed the N-G bond in
the disconnect for the solar, but I have not yet gone looking for other N-G
connections in the way that Jason suggested. I have not gotten a CA-DC link
overvoltage error, but I have still seen some of the symptoms. It is not as
bad as the other site, but what I notice on the monitoring is that in the
late afternoon/early evening one inverter shows an unreasonable amount of
solar (with power going into the grid and into the battery), while the
other inverter shows low solar (with a ton of power coming out of the
battery going to the load and the grid), and the battery SOC drops very
quickly down to the discharge end SOC (on grid) setting of 85%.
Jason and William, do you know whether any faulty wiring with N-G connected
in loads on the non-backed up side will affect the system the same way? Or
is it mostly just important to check for any neutrals connected to ground
on the backed up loads side?

Jason, as to why a supply side connection - I was under the impression that
a Power Control System just effectively limits the amount of power that can
be sold to the grid. But on a standard 200A service, that limits you to
about 7.7kw of sell back. I would not want to limit the system that much
for a roughly 20kw system, even if more than the 7.7kw can go to powering
loads and battery charging. In the Midnite monitoring, it is pretty easy to
limit the grid sell back with the "maximum feed in grid power" in the Power
Control settings.

Cheers,
Dave

On Sun, Oct 26, 2025 at 8:51 AM William Bryce via RE-wrenches <
re-wrenches at lists.re-wrenches.org> wrote:

> All
>
> Most modern all-in-one (AIO) inverters are designed around a non-isolated
> high-voltage DC bus. This design is what allows manufacturers to advertise
> higher output numbers — for example, “15–18 kW” — while the inverter may
> only deliver 12 kW during nighttime operation when PV isn’t contributing
> (it is also cheaper to build). Without the shared non-isolated DC link,
> these power-boosted daytime ratings wouldn’t be possible.
>
> A DC-bus overvoltage fault almost always points to one of two issues:
>
>    1.
>
>    Improper system design
>    2.
>
>    Wiring and installation mistakes
>
> Off-Grid Example
>
> A common DIY mistake is pairing a large solar array (because glass is
> cheap) with a relatively small battery (not as cheap).
> When the battery is full and a large load suddenly turns off, the excess
> energy has nowhere to go. The inverter throttles back, but the stored
> energy can still push the DC-bus voltage up rapidly. If the inverter didn’t
> protect itself, the capacitors would be at risk — so it shuts down with an
> overvoltage error. This is also common when using a large retrofitted AC
> coupled array.
> Grid-Tie Example
>
> In grid-connected systems, DC-bus overvoltage is typically caused by wiring
> mistakes, some common examples are:
>
> Multiple neutral-to-ground bonds
>
> Reversed hot/neutral at outlets
>
> A single inverter is usually more forgiving. However, stacked inverters must
> share phase, neutral, and grounding relationships precisely. Any mis-wiring
> can create circulating currents. That circulating energy is not accounted
> for in the inverter’s internal power calculations, so the DC-bus voltage
> rises unexpectedly — and the inverter shuts down for protection.
>
> A DC Link Overvoltage error is usually not an inverter defect — it’s a
> system design or system wiring problem 99% of the time. Ensuring proper
> neutral/ground bonding, correct polarity, adequate battery capacity, and
> proper commissioning is critical to reliable AIO operation.
>
> This behavior is not unique to the MN AIO — all AIO inverters using a
> non-isolated DC-bus topology share the same fundamental protection
> limitations.
>
> Inverters are getting more complex to deal with the new rules and utility
> requirements, this causes reliability issues as we move (or pushed) to full
> integration with the grid.  Because you know we all want the utility to
> have full control over our systems!
>
> On Sat, Oct 25, 2025 at 9:51 AM Jay via RE-wrenches <
> re-wrenches at lists.re-wrenches.org> wrote:
>
>> I had an issue with a EG4 inverter and tech support was an insistent that
>> there was multiple ground neutral bonds in the system. Fully off grid
>> system with no generator.
>> It wasn’t the issue but it seems like these inverters are sensitive to
>> it.
>>
>> Is the inverter internally bonded, not that I’m aware of or what I’ve
>> measured.
>> And so I can’t figure out what it’s measuring between N/G to have a
>> potential issue.
>>
>> Maybe someone can explain?
>>
>> Jay
>>
>>
>>
>> On Oct 24, 2025, at 6:49 PM, Dave Tedeyan via RE-wrenches <
>> re-wrenches at lists.re-wrenches.org> wrote:
>>
>> 
>> Thanks all for your thoughts here. It seems that the general consensus so
>> far is that the N-G should be bonded in the solar main disconnecting means
>> when doing a supply side connection (even if it is not technically a
>> separate service). When talking with Midnite Power, we made it very clear
>> that we were talking about a N-G bond in the inverter combiner / solar main
>> disconnect, and not the backed up loads panel. And being that NY is still
>> on 2017 code (only going to 2020 next year), with the ambiguity in the code
>> I am probably okay to undo the N-G connection in the main solar disconnect.
>>
>> In this case specifically, we are backing up only a small subset of loads
>> in the house. The reason for the two inverters is because there is about
>> 24kw of solar. There is also a manual transfer switch to normally be able
>> to power the loads off the inverters, but to be able to switch to powering
>> the loads off the main panel in case of service needed on the inverters.
>> The neutrals in this system are all connected together, and I am not sure
>> if that could cause issues. It is probably difficult to explain all the
>> details in words, so I am copying part of the three line diagram here. The
>> second picture is the conductor and conduit schedule.
>>
>> Tom, I get it that the error sounds like it is on the DC side based on
>> the words, but when I talked to Midnite about it they were pretty clear
>> about it being due to multiple N-G bonds and they did not mention that it
>> could have anything to do with the DC side.
>>
>> Jason, that is a good idea as well to check the loads that got moved over
>> to see if there are any other inadvertent N-G bonds on the premises. This
>> also similar to what Midnite suggested, which was to lift the N-G bond in
>> the main panel and then see if they are still bonded somewhere else. It
>> seems like it can be a slow process to track that down with power cut to
>> the house, but that might be what I have to try if removing the N-G bond in
>> the solar main disconnect does not help.
>>
>> So, my main questions to anyone who has an opinion here are two fold:
>> 1. Do you bond the N-G in general in the main solar disconnect for a
>> supply side connection?
>> 2. Specifically with the Midnite AIO inverters, has anyone else had
>> issues with this "CA-DC link Overvoltage error" due to having a second N-G
>> bond in the house? (one in the main service panel, and one in the solar
>> main disconnect if it is a supply side connection). Or have you seen this
>> error and found that it was caused by something else?
>>
>> Cheers,
>> Dave
>>
>>
>>
>> On Thu, Oct 23, 2025 at 5:25 PM Jason Szumlanski via RE-wrenches <
>> re-wrenches at lists.re-wrenches.org> wrote:
>>
>>> I agree. It is not a separate service. It is a "tap" of the existing
>>> service (I use that term loosely here). The additional service disconnect
>>> for the supply side interconnection should have L1, L2, and N
>>> connected ahead of the premises main disconnect with no EGC. There should
>>> be a N-G bonding jumper in the PV service disconnect. Then there should be
>>> no N-G bonds downstream of both the premises main disconnect and the PV
>>> service disconnect.
>>>
>>> Basically, it's the same as two main disconnects running from a meter
>>> enclosure. Each of the service disconnects has a N-G bond required.
>>>
>>> Some AHJs (the crazy ones) require an EGC between the enclosure where
>>> the supply side connection is made and the PV service disconnect. This
>>> creates a parallel fault path, and should not be allowed, in my humble
>>> opinion.
>>>
>>> Now it sounds like you might be doing a partial home backup. The issue
>>> there is that sometimes the loads or enclosures you moved over the backup
>>> load panel have inadvertent N-G bonds. Sometimes it's as simple as someone
>>> who got creative and used a ground as a neutral in a switch box. I say
>>> simple, but that can be hard to track down.
>>>
>>> When Midnite said to remove it from the inverter distribution panel,
>>> they were probably talking about the protected loads panel.
>>>
>>> I think more information is needed.
>>>
>>>
>>>
>>> Jason Szumlanski
>>> Principal Solar Designer | Florida Solar Design Group
>>> NABCEP Certified Solar Professional (PVIP)
>>> Florida State Certified Solar Contractor CVC56956
>>> Florida Certified Electrical Contractor EC13013208
>>>
>>>
>>> On Thu, Oct 23, 2025 at 11:03 AM Tyrone Houck via RE-wrenches <
>>> re-wrenches at lists.re-wrenches.org> wrote:
>>>
>>>> My interpretation, and the way it was explained to me by both Ryan
>>>> Mayfield and Kyle Bolger(several years ago at this point, popular opinion
>>>> may have changed since). Is that it is not a seperate service.
>>>>  It is however required to have a ground-neutral bond( though in my
>>>> opinion pointless and redundant as there is a bond in the enclosure where
>>>> the tap is made). If it gives you any issue, ensure as close to equal
>>>> potential to the grounding electrode as possible(and make sure your neutral
>>>> is appropriately sized). I personally don't see the point but it is an NEC
>>>> requirement
>>>>
>>>>
>>>>
>>>>
>>>> On Thu, Oct 23, 2025, 6:08 AM Dave Tedeyan via RE-wrenches <
>>>> re-wrenches at lists.re-wrenches.org> wrote:
>>>>
>>>>> Hi All,
>>>>>
>>>>> For years I have understood that when doing a supply side connection,
>>>>> you are supposed to bond the neutral to ground in the solar service
>>>>> disconnect. I have a Midnite AIO system with two inverters (set up as a
>>>>> supply side connection) that recently was getting a CA-DC link Overvoltage
>>>>> error. When talking to Midnite, their first response was to make sure that
>>>>> there was only one N-G bond in the whole system. When I mentioned that the
>>>>> solar is a separate service and thus there is a N-G bond in both the main
>>>>> panel and the inverter distribution panel, they said to remove it from the
>>>>> inverter distribution panel. I want to make the system work right, but I
>>>>> also am hesitant to do something against how I have understood the NEC to
>>>>> be for a long time. I am curious if you all have thoughts on this or run
>>>>> into a similar situation.
>>>>>
>>>>> Cheers,
>>>>> Dave
>>>>>
>>>>> --
>>>>>
>>>>> [image: Logo] <https://www.sungineersolar.com/>
>>>>> Dave Tedeyan, P.E.
>>>>> Owner | Sungineer Solar
>>>>> p: he | him | his
>>>>> a: 1653 Slaterville Rd.
>>>>> <https://www.google.com/maps/search/1653+Slaterville+Rd.+%C2%A0+%7C%C2%A0Ithaca,+NY+14850?entry=gmail&source=g>
>>>>>
>>>>> <https://www.google.com/maps/search/1653+Slaterville+Rd.+%C2%A0+%7C%C2%A0Ithaca,+NY+14850?entry=gmail&source=g>| Ithaca,
>>>>> NY 14850
>>>>> <https://www.google.com/maps/search/1653+Slaterville+Rd.+%C2%A0+%7C%C2%A0Ithaca,+NY+14850?entry=gmail&source=g>
>>>>> w: www.sungineersolar.com <http://www.sungineersolar.com/>
>>>>> c: (607) 270-0370
>>>>> _______________________________________________
>>>>> List sponsored by Redwood Alliance
>>>>>
>>>>> Pay optional member dues here: http://re-wrenches.org
>>>>>
>>>>> List Address: RE-wrenches at lists.re-wrenches.org
>>>>>
>>>>> Change listserver email address & settings:
>>>>> http://lists.re-wrenches.org/options.cgi/re-wrenches-re-wrenches.org
>>>>>
>>>>> There are two list archives for searching. When one doesn't work, try
>>>>> the other:
>>>>> https://www.mail-archive.com/re-wrenches@lists.re-wrenches.org/
>>>>> http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org
>>>>>
>>>>> List rules & etiquette:
>>>>> http://www.re-wrenches.org/etiquette.htm
>>>>>
>>>>> Check out or update participant bios:
>>>>> http://www.members.re-wrenches.org
>>>>>
>>>>> _______________________________________________
>>>> List sponsored by Redwood Alliance
>>>>
>>>> Pay optional member dues here: http://re-wrenches.org
>>>>
>>>> List Address: RE-wrenches at lists.re-wrenches.org
>>>>
>>>> Change listserver email address & settings:
>>>> http://lists.re-wrenches.org/options.cgi/re-wrenches-re-wrenches.org
>>>>
>>>> There are two list archives for searching. When one doesn't work, try
>>>> the other:
>>>> https://www.mail-archive.com/re-wrenches@lists.re-wrenches.org/
>>>> http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org
>>>>
>>>> List rules & etiquette:
>>>> http://www.re-wrenches.org/etiquette.htm
>>>>
>>>> Check out or update participant bios:
>>>> http://www.members.re-wrenches.org
>>>>
>>>> _______________________________________________
>>> List sponsored by Redwood Alliance
>>>
>>> Pay optional member dues here: http://re-wrenches.org
>>>
>>> List Address: RE-wrenches at lists.re-wrenches.org
>>>
>>> Change listserver email address & settings:
>>> http://lists.re-wrenches.org/options.cgi/re-wrenches-re-wrenches.org
>>>
>>> There are two list archives for searching. When one doesn't work, try
>>> the other:
>>> https://www.mail-archive.com/re-wrenches@lists.re-wrenches.org/
>>> http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org
>>>
>>> List rules & etiquette:
>>> http://www.re-wrenches.org/etiquette.htm
>>>
>>> Check out or update participant bios:
>>> http://www.members.re-wrenches.org
>>>
>>> <PV3.0 as built.jpg>
>> <PV3.1 as built.jpg>
>> _______________________________________________
>> List sponsored by Redwood Alliance
>>
>> Pay optional member dues here: http://re-wrenches.org
>>
>> List Address: RE-wrenches at lists.re-wrenches.org
>>
>> Change listserver email address & settings:
>> http://lists.re-wrenches.org/options.cgi/re-wrenches-re-wrenches.org
>>
>> There are two list archives for searching. When one doesn't work, try the
>> other:
>> https://www.mail-archive.com/re-wrenches@lists.re-wrenches.org/
>> http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org
>>
>> List rules & etiquette:
>> http://www.re-wrenches.org/etiquette.htm
>>
>> Check out or update participant bios:
>> http://www.members.re-wrenches.org
>>
>> _______________________________________________
>> List sponsored by Redwood Alliance
>>
>> Pay optional member dues here: http://re-wrenches.org
>>
>> List Address: RE-wrenches at lists.re-wrenches.org
>>
>> Change listserver email address & settings:
>> http://lists.re-wrenches.org/options.cgi/re-wrenches-re-wrenches.org
>>
>> There are two list archives for searching. When one doesn't work, try the
>> other:
>> https://www.mail-archive.com/re-wrenches@lists.re-wrenches.org/
>> http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org
>>
>> List rules & etiquette:
>> http://www.re-wrenches.org/etiquette.htm
>>
>> Check out or update participant bios:
>> http://www.members.re-wrenches.org
>>
>> _______________________________________________
> List sponsored by Redwood Alliance
>
> Pay optional member dues here: http://re-wrenches.org
>
> List Address: RE-wrenches at lists.re-wrenches.org
>
> Change listserver email address & settings:
> http://lists.re-wrenches.org/options.cgi/re-wrenches-re-wrenches.org
>
> There are two list archives for searching. When one doesn't work, try the
> other:
> https://www.mail-archive.com/re-wrenches@lists.re-wrenches.org/
> http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org
>
> List rules & etiquette:
> http://www.re-wrenches.org/etiquette.htm
>
> Check out or update participant bios:
> http://www.members.re-wrenches.org
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org/attachments/20251029/d4a9f091/attachment.htm>