I have a Goodwe  GW5048D-ES inverter. During load shedding some of my LED lights on the back up output of the inverter very dimly glow even when not switched on. I never thought much about it but with load shedding now being much more frequent I want to get to the bottom of this. I measured +-80V between neutral and earth which indicates that the neutral is floating when the grid is down. Could this be the reason why the LED's glow? I also just received an answer from David Havenga (Goodwe rep Southern Africa) confirming that the ES inverters do indeed not have a neutral/earth bonding relay on the back up output ☹️. He suggested to permanently bond neutral and earth on the back up output. But this is as far as I know illegal in South Africa if the earth supply is provided from the grid. What do the experts here suggest I do?

I had the same issue with another inverter brand.  The electrician then did the bond on neutral and earth (apparently this is common practice). According to the electrician it is illegal to have higher than 20v(If I can remember correctly) on the neutral. ( I am not an installer or electrician just sharing what happened during my installation)

Have a look at this Goodwe installation video :  (take note at 03:10)

 

2 hours ago, DennisD said:

The electrician then did the bond on neutral and earth (apparently this is common practice).

SANS actually requires a bonding switch.

 

And you're not supposed to have more than one TN bond, so even though sparkies will sometimes just bond the output of the inverter (as a kind of lesser-evil approach), it's not really right.

So yeah... I don't know these inverters well enough, and I really don't want to speak ill of something I don't know that well, but on the face of it, it actually seems the Goodwe is not SANS compliant, at least not on its own (some external interlocking bonding device is required). Which is a weird situation, since it is NRS097 compliant and on the Cape Town list...

3 hours ago, DennisD said:

take note at 03:10

That suggests tying the neutrals together. If you do that, then you get a TN bond via the existing bond at your earth spike (or at the transformer for TN-S). However, SANS says you  are not allowed to rely on this bond in this manner.

To illustrate: In a single phase TN-S setup, with a double pole isolator in the distribution board, opening the isolator will disconnect the neutral along with the bond that you were relying on.

As a further illustration, if an overload trips your main switch, and the inverter takes over for the backup loads, again you have lost the bond.

23 minutes ago, plonkster said:

it actually seems the Goodwe is not SANS compliant, at least not on its own (some external interlocking bonding device is required). Which is a weird situation, since it is NRS097 compliant and on the Cape Town list...

I agree. That is why I was so surprised what David told me. But we have to be careful with our judgement because it is at the moment hearsay. Maybe David got it wrong. Maybe my inverter is faulty. Can other owners with Goodwe ES inverters (without a permanent external earth/neutral bond on the back up output) maybe also measure the voltage between neutral and earth on the back up output when the inverter is in island mode and advice. @Bobster?

Haha. I have this problem all over the place and it predates the Goodwe. 

EG i have an LED light in the foyer. I turn on the outside lights (they're on the same 2 lever light switch) and the foyer light starts glowing. 

I gave up with an LED light in our garage because it always glowed.

But it's not throughout the house. I have other LED lights that behave perfectly. 

Where would i take the measurement you are asking for? 

This may be useful... https://lamphq.com/led-lights-glow/

I have also noticed at our office where we have led tube lights installed that some of the lights had a dim glow even if switched off - those lights were not connected properly (live - live, neutral - neutral).  They were instead connected live-neutral and neutral to live - as soon as that was fixed on the light connection side the lights worked correctly.

23 hours ago, Bobster said:

Where would i take the measurement you are asking for?

between any neutral wire on the "back up" section of the DB and earth. But the inverter has to be in island mode (not connected to the grid or grid is down).

17 hours ago, Fuenkli said:

between any neutral wire on the "back up" section of the DB and earth. But the inverter has to be in island mode (not connected to the grid or grid is down).

Basically, take any 3-pin plug top, pull it out slightly, then carefully stick the the multimeter in behind the plug and push the probe against the neutral and earth pins (this sort of measurement should preferably not be done if the cover plate is metal... I was probably about 12 years old when I learned that... don't ask).

If the system is bonded, you will get a zero reading. If not, it tends to float up to almost halfway, so typical values are between 90V and 110V.

5 minutes ago, plonkster said:

(this sort of measurement should preferably not be done if the cover plate is metal... I was probably about 12 years old when I learned that... don't ask)

hint, 2 strip wires and......... hehehehehhehe!! small, i spotted my father doing this, at 8 year old I was clever to put a 12v thing in to 220v ,by stripping the wires and doing this trick, lets say, it was good bang and sparks. hehehehehe and that's how learned. 

11 minutes ago, Gerlach said:

spotted my father doing this

Likewise. My father would check the voltage by doing this trick. So we were out in the Free State visiting family, and I god a multimeter for Christmas. I proceeded to show my cousin the trick. When we were done, there was a black mark next to the live pin and the little silver screw holding the cover plate on was no longer silver.

We never told a soul. The multimeter probe survived too, but the mark was there to remind me for all of high school.

1 minute ago, plonkster said:

Likewise. My father would check the voltage by doing this trick. So we were out in the Free State visiting family, and I god a multimeter for Christmas. I proceeded to show my cousin the trick. When we were done, there was a black mark next to the live pin and the little silver screw holding the cover plate on was no longer silver.

We never told a soul. The multimeter probe survived too, but the mark was there to remind me for all of high school.

Hahahahahaha!!!! the stuff we did when we where small. I bought my first analog multimeter when we stayed in Prt, at the Pick and Pay Hypermarker. We move back to cape town in 1998 so it was before that time, and still got it and still working. Looks like this:

53 minutes ago, plonkster said:

...this sort of measurement should preferably not be done if the cover plate is metal... I was probably about 12 years old when I learned that... don't ask).

This reminds me - do check that your probes are not plugged into the mA/A/current-measuring input on the Multimeter! This will do a bit more damage than leaving a black mark...! 🤯

Using the Plonkster method I get 104v between ground and neutral. Not quite half way as L to N gives exactly 230

2 hours ago, Bobster said:

Using the Plonkster method I get 104v between ground and neutral. Not quite half way as L to N gives exactly 230

thanks for measuring. So this basically confirms that the ES Goodwe inverters do not bond neutral to earth on the back up output. The way I see it is that we have 4 options to address this problem. 3 of them are probably illegal.

1. do nothing. What is the risk?

2. Permanently bond neutral to earth on the back up output. If I remember correctly this is illegal because it creates a second neutral earth bond

3. Connect the inverter in and out neutrals. Probably also illegal because you loose the earth connection when the grid is down or the main switch is switched off.

4. Install an external bonding relay. I am not sure how one would do this and how much this would cost. 

Edited February 8, 20205 yr by Fuenkli

Now the power is back I get

L to E - zero

N to E - 220

L to N - 220

So maybe I'm not the best test case 

22 minutes ago, Bobster said:

L to E - zero

N to E - 220

to me this looks like you have mixed up live and neutral. Did you measure in the DB board or on a plug? Plugs sometimes are not wired correctly.

2 hours ago, Fuenkli said:

to me this looks like you have mixed up live and neutral. Did you measure in the DB board or on a plug? Plugs sometimes are not wired correctly.

Same plug in the same socket as the first test. Plug is just a plug, nothing connected. I made sure of the pin orientation first.

In a multisocket extension lead with the original molded plug on it. 

But yes... Not exactly measuring at the DB

12 hours ago, Bobster said:

I made sure of the pin orientation first.

if you are correct then you have 220V on neutral 😱. I don't understand how this is possible because in normal operation there is a dead short between neutral and earth. 

Edited February 9, 20205 yr by Fuenkli

I'll check again, and elminate the extension strip. 

It is fine. Turned out to be a cross wired extension cord. 

Thanks. I learned a bit through all of this and googling the difference between live and neutral 

34 minutes ago, Bobster said:

It is fine. Turned out to be a cross wired extension cord.

thanks for letting us know. Unfortunately we still have the floating neutral problem with the Goodwe. What are you going to do? As a short term (illegal) solution I am going to permanently bond neutral to earth on the back up output. I am not comfortable to have a potentially lethal voltage on my neutrals. As a long-term solution I will add an external neutral to earth bonding relay. 

On 2020/02/08 at 1:54 PM, Bobster said:

Using the Plonkster method I get 104v between ground and neutral. Not quite half way as L to N gives exactly 230

So because in thus test the L and N were cross wired in an extension cord, the 106 would have been earth to LIVE and so I presume 124 N to E

1 hour ago, Bobster said:

so I presume 124 N to E

correct. This is very much in line with what I have measured on my side as well. Our back up output neutral is floating ☹️. I am still wondering how dangerous this really is? I assume the RCD would prevent a dangerous current from flowing to earth.