How does your inverter deal with ground.

[pollenface]

I think grounding is necessary if you're using RCDs

 

[FilterGuy]

I think grounding is necessary if you're using RCDs

There are a few scenarios where an RCD/GFCI would still trip without an earth ground....They all center on power going from the hot wire to the equipment grounding system (The 'ground wire') that is not connected to earth ground. However, it is also true that there are a lot of cases that are not covered if there is no earth ground.

 

[bgflyguy]

Anyone do the testing for any of the GroWatt inverters yet?

 

[ChrisG]

Anyone do the testing for any of the GroWatt inverters yet?

Following this thread myself with my GroWattSPF3000. I don't have the GW grounded directly, don't think there is an option/lug for it. However, I do have it connected to grid via AC IN with Hot/Neutral/Ground. Testing output from device with an open ground plug tester, everything seems to be fine (according to $10 tester from Home Depot) in both direct grid power mode and inverter mode to AC OUT.

 

[SolarShed]

My 1000W inverter is in my shed, so to ground it, I have to create a ground by burying some pipe in the earth (earth ground) and have proper wire (14AWG) BUT I'm curious about something, I've only ever plugged in appliances/devices the do NOT have a ground prong, i.e. only 2 prong...so would having a ground even benefit this situation? I'm on the fence as to where I should make a ground.

If I decide to ground (or if it recommended) what would a proper earth ground be in this case? a 6ft copper pipe (hollow or solid? and what diameter)...thanks!

 

[Trout]

View attachment 33722

From what I’m seeing, I should take the grounding lug to my earth ground . Thanks for all the time and effort you put into this forum. Your my go to guy for straight information

 

[Delmar]

Same problem here, just got my new AIMS 48v 6000w inverter last night, hooked it up, saw the open ground. So I thought, oh I'll just bond the ground-neutral externally. Emailed Aims to make sure that's okay, and good thing I did because I got this reply:

"You CAN NOT bond the neutral and ground on this inverter as it is a high frequency style inverter and creates 60V on the neutral and 60V on the hot line. Bonding these will force the power to the hotline and cause damage to the inverter that is not covered under warranty."

So now I have to waste hundreds of $$$ on shipping and "restocking fees" to return this garbage, and get something with a real 120 hot wire, and that can be bonded properly.

Another popular inverter (only because of the cheap price) is the high-frequency Reliable WZRELB. Most will measure a floating 60V hot-ground and 60V neutral-ground, as was tested on my 48V 2500W.

As a test I back-fed my house breaker panel by terminating all three leads, effectively bonding the inverter neutral and grounding the case. However I temporarily connected a neutral inline 5A fuse in the event of a 60V short. Worked just fine, no smoke and did not blow the 5A fuse after loading to around 4.5A. Since then I have pushed the WZRELB hard multiple times, especially during the great Houston freeze where it ran my furnace to keep the pipes from bursting. No problems so far with the grounding.

I always connect the inverter to the breaker panel first before powering it on. Maybe AIMS felt the connection sequence would be a liability.

 

[Hedges]

Another popular inverter (only because of the cheap price) is the high-frequency Reliable WZRELB. Most will measure a floating 60V hot-ground and 60V neutral-ground, as was tested on my 48V 2500W.

Do you think AC output is galvanically isolated from battery, or not?
Measure AC voltage from ground to battery negative terminal, and to positive terminal. I'm thinking depending on how it is designed, it might drive one of those to 60Vrms (and the other higher due to DC riding on top of that.)

 

[Riderice]

Maybe AIMS felt the connection sequence would be a liability.

Yeah, could be. What I didn't mention in my original comment (or to AIMS) was that I actually *did* bond its ground and neutral and ran it for a while. It seemed fine. But after I getting their dire earnings against it, I decided I didn't want to deal with it blowing up at some later point and them not covering the warranty.

 

[marivesel]

Hi, folks.

I've measured my solar inverter VM II PLUS 5500W and this is the typical installation at home. (simplified... really)
Does everything seem normal at first glance? I'm confused about the N-G bond at the inverter (and why the bond "oscillates" between AC-IN and AC-OUT terminals, regarding the BYPASS or INVERT mode?) . This way it's making a second one. (I already have one bond at mains enclosure)
After discussing this installation, I will have one more important question for different scenario, if I may ask.

The yellow lines indicate the internal physical connections in the inverter itself, between the terminals.

 

[Hedges]

The way that switches and bonds neutral doesn't look correct to me. I will send neutral return current through ground, and other loads in the house may send their neutral return current through the inverter's bond and ground.

Ground wire from toaster should be continuous through to mains AC enclosure, with no switch interrupting it. (or if 2-wire, no ground connection.)
I don't like appliance chassis connected to N. I've seen that in clothes dryers, and rewired it (with 4-prong plug instead of 3-prong.)
There should be a ground rod connecting to ground bar at mains AC enclosure.

I think the inverter should only have neutral-ground bond on the right (for load), not on the left.
I think that bond should be closed with L and N switches are open. (not closed when switches are closed.)

Switching and bonding neutral like that is used for mobile devices (RV, boat) but for hard-wired, at least in the U.S., it can leave N (and ground) always connected to mains AC enclosure, never bond neutral to ground in the inverter, only open L.

 

[Riderice]

why the bond "oscillates" between AC-IN and AC-OUT terminals

The second diagram makes sense to me: during inverter mode (power from battery), your device automatically completely disconnects from all house AC sources, for obvious reasons (although it's weird that the cartoon doesn't show a battery anywhere).

The first diagram looks wrong, it almost looks like a ground loop between the AC mains and inverter internals...? As @Hedges mentioned, it would be better if the toaster ground travelled separately all the way thru inverter back to the AC mains, where it's bonded to N.

 

[Hedges]

"Old enclosure"

By "Old", do you mean that used to be the main breaker enclosure, fed by the grid? But now used as a sub-panel after a new main panel was installed?
That would explain why it used one busbar for both neutral and ground. But as a sub-panel it would require an isolated neutral bar. Either add
an isolated neutral bar or replace the panel with a new panel that does one.

Panels commonly available today have bus for L1 and L2, that breakers plug onto. They have an isolated neutral. They have an optional screw to bond neutral bar to enclosure, if to be used as a main panel with neutral-ground bond. Optional (or included) separate busbar not isolated, connected to box, is available.

 

[FilterGuy]

Is that image from the manual?

That is a very strange set-up and is not proper.

* In passthrough mode there is a NG-bond that is addition to the one on the grid.... meaning there are two N-G bonds..... Not good
* In battery mode, there is also on NG bon on the input (creating a 2nd bond) and there is no bond on the output.... double not good.

Could you provide a link to the manual?

 

[timselectric]

Nope, that's not safe at all.

 

[marivesel]

Ok... let me try to explain a little bit more:
1. The image is drawn by me and is really over-simplified, I just wanted to have illustration of "complete circuit", but really just wanted to focus on the inverter terminals.
2. I live in Bulgaria and here our old houses have "two wires" installations from the MAINS (L and N, just one 230V phase), and that is why the drawing between inverter (near the MAINS) and the "OLD ENCLOSURE" (in the house) is indicating only two wires (I need to reroute another grounding wire.. I get it, but it's not subject of my question/problem). Also, there is a grounding rod before the MAINS enclosure, directly connected to the N (in the enclosure provided by the electricity distribution company)
3. From now on, lets assume that my MAINS AC is 3-wire and also my house enclosure is also 3-wire and everything is "perfect" outside of the inverter.
4. Lets just focus on the inverter terminals and the N-G bonding's in it, which worries me and are the general subject of my posting.

Here is a drawn by me illustration of the method which I used to conclude the internal bonding of the terminals on the inverter.

Please note - in this scenario I have connected wires L and N from the mains ONLY (without grounding, so it won't "bond" on the multimeter from the mains)

 

[timselectric]

Ok... let me try to explain a little bit more:
1. The image is drawn by me and is really over-simplified, I just wanted to have illustration of "complete circuit", but really just wanted to focus on the inverter terminals.
2. I live in Bulgaria and here our old houses have "two wires" installations from the MAINS (L and N, just one 230V phase), and that is why the drawing between inverter (near the MAINS) and the "OLD ENCLOSURE" (in the house) is indicating only two wires (I need to reroute another grounding wire.. I get it, but it's not subject of my question/problem). Also, there is a grounding rod before the MAINS enclosure, directly connected to the N (in the enclosure provided by the electricity distribution company)
3. From now on, lets assume that my MAINS AC is 3-wire and also my house enclosure is also 3-wire and everything is "perfect" outside of the inverter.
4. Lets just focus on the inverter terminals and the N-G bonding's in it, which worries me and are the general subject of my posting.

Here is a drawn by me illustration of the method which I used to conclude the internal bonding of the terminals on the inverter.

Please note - in this scenario I have connected wires L and N from the mains ONLY (without grounding, so it won't "bond" on the multimeter from the mains)


View attachment 86865

Are you sure that those measurements are in the correct places. N/G bonding should take place in battery mode and not in bypass mode. According to your picture, it's the opposite.

 

[marivesel]

Are you sure that those measurements are in the correct places. N/G bonding should take place in battery mode and not in bypass mode. According to your picture, it's the opposite.

Yes, I'm pretty as I double checked, since it seemed strange for me.
Then I decided to write in the forum if this is normal, or my measurement method is not correct maybe?

 

[timselectric]

I don't see anything wrong with the method, as shown. But, the readings are backwards from what they should be.

 

[Zil]

Folks, When setting up an inverter, one of the more important safety things to get correct is the grounding and the neutral-Ground bond.

• All of the inverters have a grounding lug
• All of the inverters have a ground connection on the AC out.
• Some inverters have an AC in and when they do they have a ground connection on the input.

Sadly, the information provided in many manuals is nearly non-existent when it comes to how it handles ground internally.

Are the two (or three) grounds tied together?
Is there a neutral-ground bond?
If there is an AC-IN, does it change the neutral-ground bond when getting power from the AC-in?

I have started to try and collect this type of information for various inverters. Once I get information on several of them I will put it all into a resource and post it.

How does your inverter handle internal ground? If you send me the make/model of your inverter, I will add it to the resource.

The next 3 posts are examples of what I envision for each page of the resource.

This can be a great resource. We would like an index that links to the inverter information without scrolling many many many posts.

 

[FilterGuy]

This can be a great resource. We would like an index that links to the inverter information without scrolling many many many posts.

I have not updated it in a while but.....

Grounding details for specific make/model of inverters

To get to the paper, click on the orange button at the top of the screen. It is important to understand how an inverter handles grounding in order to set it up correctly. Unfortunately, many inverters do not provide adequate documentation on...

diysolarforum.com

 

[tssd]

EPEVER UP3000-HM10022
Hybrid Inverter

Just Been through this exact same senario with Epever, initially contacted them about their output Neutral Ground /Earth bonding and after confirming its suitability i found out that infact when supplied by a Mains TNC-S supply( Neutral Earthing bond at the Main switchboard ) operating in Bypass Mode the earthing system configuration passes through to the output correctly', but when placed in inverter mode the inverter output creates an IT system /isolated supply -unbonded system , therefor in Australia the outputs downstream RCD protection would not be configured correctly for correct operation as its looking for a TNC-S supply

In simple terms the output should be neutral earth bonded but cannot be because when in bypass it already is and then placing a second neutral earth bond at the output places this in parallel to the first, should the inverter supply be by a RCD / GFCI circuit then nuisance tripping will occur whilst protective earth would share load current with the neutral that's not permitted

So moral to this story , understand your local standards and the accepted earthing system to be employed , because there are a multitude of earthing systems and the language varies by country , ie GFCI,RCD, Neutral Earth Bonded, MEN system , TNC, TNCS , IT , Earthing Conductor , Grounding Conductor, being very clear with your inverter supplier of what you require in the manuals i have read, this topic is definitely not clear when you start dealing with Clone type inverters and Languages or in Australia we say the Lingo

The solution to the above is now not that easy , an inverter with a configurable earthing system is what's required or a system as shown in filter guys inverter drawings where the output is automatically reconfigured, please look at the variances and i congratulate the Filter Guy for trying to clarify and hope i have added some further incite

I have the EPEVER UP5000 M3842, and it has the same lack of Neutral grounding.
I currently use it with a Neutral ground link on the Inverter output, so the RCD (GFI) will work.
This will cause a trip on the Utility RCD if I switch to bypass.

We have TT Earthing so the NG Bond is at the utility pole transformer.

I did have an Iconica 3000W hybrid inverter for a short while, and this did have a switchable neutral ground bound, but the solar charger had a short on it, so that went back pretty sharpish.

Shame the EPEver doesn't have the option, as it's pretty good apart from that (oh and the fan seems to run fast all the time now- another fault these seem to get)

 

[Kornbread]

How did I miss this thread?

Growatt spf6000t dvm us and its brethren ... we had a long discussion about this inverter and its lack of proper documentation in regards to n-g bonding.

 

[Hartswood]

When Deye, Sunsynk inverters disconnect from grid and the signal islanding mode has been selected an output is sent to two contacts to which a user supplied relay can be connected. Neutral and ground wires are connected to the relay and a ground neutral bond is created when activated. When grid restored the bond is broken by the relay. Here are a couple of Sunsynk videos describing the set up:

You think with this feature they would have built the relay into the chassis. I am about to buy the Sunsynk 8K inverter and intend using signal islanding mode. Keith, the CEO, makes the comment some RCDs can be too quick for the relay, causing a problem. Elsewhere Sunsynk recommend using a fast RCD to protect from overcurrent. I'm concerned about nuisance RCD triggers. Keith talks about this here:

Here's another video where Keith gives a very clear overview of RCDs and earth neutral bonding:

 

[burgerking]

re Deye, Sol-Ark, SunSynk

Terms:
Inverter Module refers to the electronic parts that converts DC to AC power.
Inverter (Deye, Sol-Ark, SunSynk) refers to the whole product.
GRID = refers to terminal and wires connected to it.
LOAD = refers to terminal and wires connected to it.
Grid = refers to the power company

These Inverters have internal Bypass (or Passthrough) relays that connects Inverter Module, LOAD and GRID together during normal operation.

And disconnects GRID (from LOAD) during a utility outage. LOAD is now powered solely by Inverter Module.

When GRID is disconnected, the two LOAD lines are now floating and Inverter becomes an ungrounded power supply.
These two lines could be safely at ±120V from GND, but could drift to >1000v from GND.
When it does, could discharge to the grounded chasis through some other means causing damage to Inverter, to appliance, to occupant touching appliances.

It is recommended that during an outage (or when Inverter is used in UPS mode), it is to bond one of the LOAD wires (preferably the one labeled as N or L2) to GND, so one of the wire would be 0V, and the other would then be 240V never beyond. Now the Inverter is a ground Power Supply and neither wires have voltage potential that exceeds dielectric strength of the electronic parts, wire insulation, etc...

Watch video by SunSynk on bonding during a power outage using ATS240V signal.
(This signal does not seem available on the 12K, 15K inverter versions. Maybe these models come with a different bonding method?)


When Utility returns, Inverter:
1. waits for "Reconnection Time", then
2. deactivates ATS240V signal and connects GRID with LOAD
Everything reverts back to normal.


HERE IS THE PROBLEM
If your Inverter was wired as suggested by Keith's video, there is this chance that the external relay is much slower than the internal relay, and when internal relay connects GRID with LOAD, one of LOAD terminal is still connected to GND causing a short. DO NOTE: the Internal Relays are being powered (relying on the power of electricity to move the arm) while the external relay is being unpowered (relying on the strength of the spring to move the arm)

QUESTION:
Can we use a resistive load, say a 100w 1kΩ (replace the BLUE WIRE on the last diagram with a resistor) to complete the bonding of this external relay to GND? This way, if the relay is slower than the internal relay, it will not cause a short.

Unfortunately, this is a weak grounding.

Maybe this is why the bigger Deye inverter no longer have the ATS240V signal? Maybe the grounding is built-in already?