The partial schematic of the Axpert MAX family (7.2/8/10/11 ,OGx) below shows the relationship between GRID IN Neutral AC OUT Neutral and Earth. The blue routes are neutrals. Notice the relay 3 ( contacts RLY3A and RLY3B) .  There are 2 changeover contacts in parallel in order to share the load current requirements of the output neutral. 

The relay is energized by the DSP controller when the  machine is in LINE mode (e.g. Grid available and SUB/USB nominated as output source priority )

The wiper of each contact is connected to the load output neutral. The normally open contact is made when in line mode , and ensures the grid neutral is passed thru to the load neutral. When you have load shedding , grid falls away , machine is now in BATTERY MODE (islanding) ,  this relay drops and the normally close contact anchors the load neutral to earth as can be seen.

Please note that this internal bonding is NOT available on older machines , only generations 3 and 4 (eg MKSIII , MKSIV etc).

The arrangement is optimum as the changeover action of the relay is break before make. Such is its simplicity and gracefulness that my own MAX has an RCD grid bound as well as RCD  load bound . The grid bound RCD is not fuzzed by the changeover action of the relay when grid is restored.

 

Edited July 3, 20232 yr by BritishRacingGreen

7 minutes ago, BritishRacingGreen said:

The partial schematic of the Axpert MAX family (7.2/8/10/11 ,OGx) below shows the relationship between GRID IN Neutral AC OUT Neutral and Earth. The blue routes are neutrals. Notice the relay 3 ( contacts RLY3A and RLY3B) .  There are 2 changeover contacts in parallel in order to share the load current requirements of the output neutral. 

The relay is energized by the DSP controller when the  machine is in LINE mode (e.g. Grid available and SUB/USB nominated as output source priority )

The wiper of each contact is connected to the load output neutral. The normally open contact is made when in line mode , and ensures the grid neutral is passed thru to the load neutral. When you have load shedding , grid falls away , machine is now in BATTERY MODE (islanding) ,  this relay drops and the normally close contact anchors the load neutral to earth as can be seen.

Please note that this internal bonding is NOT available on older machines , only generations 3 and 4 (eg MKSIII , MKSIV etc).

The arrangement is optimum as the changeover action of the relay is break before make.

 

This approach is the same then as an external relay with AC signal to anchor the NC contacts between N and E in islanding mode.

13 minutes ago, CobusK said:

'Om te meet is om te weet'  

If you measure potential between the N and E bars on both non-essentials and essentials whilst the grid is on, it confirms bonding (150mV)  As soon as grid falls away and islanding takes place, potential voltage between N bar(Essentials) and E, climbs from mV to 70-80V, thus indicating 'no bonding'.  . 

Deye installation manual indicates AC signal to external relay to achieve bonding in islanding mode.

I presume that, based on the measurements, that permanent bonding on essential side, at the inverter, would serve the same function as an intermittent bonding when in islanding mode.

Thus, permanent bonding, or using AC signal for relayed bonding in islanding mode, depends on the interpretation of the regulation.  Both will achieve the same purpose.

On my installation i used an external relay to achieve bonding only in islanding mode.  

I can see why there excist so much confusion re dynamic versus hard bonding, and I am becoming reserved to insist on what I believe is correct.  There are conflicting  rules of engagement throughout the industry , and I want to say that with the South African conditions  the norm is now questioned.  Remember that grid availability is now suddenly an exception rather than rule, so some shortcomings are being amplified. 

Case in point  is a big 100kW site I have been involved in. The inverter has no notion of a grid neutral busbar and a load neutral busbar. There is only one, a neutral busbar. So that means the entire  neutral route, wether its grid side, or load side, is all  hard connected to one neutral, the eskom neutral. Which on face value is ok, but what about the eskom neutral not being there anymore, when we islanding.? 

Edited July 3, 20232 yr by BritishRacingGreen

21 hours ago, TaliaB said:

That is exactly the problem if you have a floating neutral on the inverter output due to the lack of bonding there is no refrence point to earth,should there be a faulty appliance( live or neutral shorted to metal casing) the RCD will not detect the fault current as it cannot create an imbalance between live and neutral to activate the RCD. You touching the metal case provides a path for the fault current to flow to earth only then might the RCD trip hence the reason 30ma is used in domestic applications should it be 100ma @50hz it could be fatal. 

True, but as I understand it (I'm no sparky) a few conditions need to take place together for a person to get shocked/electrocuted when this fault happens:

1. The metallic case is NOT grounded. This shouldn't be the case with modern appliances. It could happen if socket ground is disconnected, or if a 3 prong to 2 prong plug adapter is used for some reason. But still, this shouldn't be the default with most appliances or household sockets.

2. The person touching the metallic enclosure is actually touching ground somehow. I remember an old rule my mom used to say (not sure she fully understood why): Never connect electric appliances when you are barefoot. Even so, I would think that in the household, we are fairly isolated from true ground. If so, there isn't really a path for current to flow. 

I do know of one case where an amateur "sparky" (not certified) removed an RCD from a house because the tenants were complaining of nuisance tripping. 

Sadly enough, a young lady died due to electrocution, while taking a shower. This does make sense as there was probably a fault to begin with, and there was a return path to ground via water pipes. 

My opinion is that the risk is minimal, as both conditions above have to occur (in addition to a fault) in the exact few seconds that the grid comes back and the relay opens (untill inverter reconnects to grid). This is of course with a proper 30mA RCD in place. 

Again, I'm no expert, but my simple logic is that it's preferable this way (with relay). 

Revised diagram after electrician corrected my mistakes.
This will be suitable/usable for Sunsynk/deye type inverters up to 5kW 8kW in the South Africa context only.

 

Edited July 4, 20232 yr by WannabeSolarSparky
not for 8kW deye/sunsynk

58 minutes ago, meetyg said:

True, but as I understand it (I'm no sparky) a few conditions need to take place together for a person to get shocked/electrocuted when this fault happens:

1. The metallic case is NOT grounded. This shouldn't be the case with modern appliances. It could happen if socket ground is disconnected, or if a 3 prong to 2 prong plug adapter is used for some reason. But still, this shouldn't be the default with most appliances or household sockets.

2. The person touching the metallic enclosure is actually touching ground somehow. I remember an old rule my mom used to say (not sure she fully understood why): Never connect electric appliances when you are barefoot. Even so, I would think that in the household, we are fairly isolated from true ground. If so, there isn't really a path for current to flow. 

I do know of one case where an amateur "sparky" (not certified) removed an RCD from a house because the tenants were complaining of nuisance tripping. 

Sadly enough, a young lady died due to electrocution, while taking a shower. This does make sense as there was probably a fault to begin with, and there was a return path to ground via water pipes. 

My opinion is that the risk is minimal, as both conditions above have to occur (in addition to a fault) in the exact few seconds that the grid comes back and the relay opens (untill inverter reconnects to grid). This is of course with a proper 30mA RCD in place. 

Again, I'm no expert, but my simple logic is that it's preferable this way (with relay). 

Your insulated body nevertheless still has capacitance to adjacent ground, which provides an AC path. That’s why you can still get a shock from a live wire by touching just one terminal, even if you're not touching ground or another live wire the AC will alternately charge and discharge your body capacitance, and if the voltage is high enough, the resulting current can be fatal. 

In this very non perfect world we have here in SA. I would say do whatever you can do to make the Earth leakage trip in the event of a fault. Using a relay bond does this. Using a permanent bond also does this. 

But whatever you do, don't do nothing as in the event of loadshedding you have zero protection & a floating neutral. 

We can split hairs all night long. Make the Earth leakage functional during the loss of grid is the ultimate goal here. How you want to achieve that is your business. 

Edited July 3, 20232 yr by Steve87

Agreed.

But also, which one will be acceptable to cover an insurance claim. Both?

12 hours ago, PowerUser said:

@Coulomb without dragging you into this whole discussion, can you please just confirm for us, if the Voltronic inverters have a mechanism in place should the internal relay bond fails while islanding?

No, there is no test that the bond is working when it should, if that's what you mean.

11 hours ago, WannabeSolarSparky said:

Revised diagram after electrician corrected my mistakes.
This will be suitable/usable for Sunsynk/deye type inverters up to 8kW in the South Africa context only.

1. For the sake of DIY'ers, just say for the 5kw unit, as the 8kw connection layout differs. (Grid-Gen-Load, + no pe)

2. Get your Sparky back to fix his drawing err. 😉

10 minutes ago, VirWat said:

1. For the sake of DIY'ers, just say for the 5kw unit, as the 8kw connection layout differs. (Grid-Gen-Load, + no pe)

2. Get your Sparky back to fix his drawing err. 😉

my error, i did not read the 8kW manual

10 minutes ago, WannabeSolarSparky said:

my error, i did not read the 8kW manual

No, not point 1.😄

point 2.🤨

6 minutes ago, VirWat said:

No, not point 1.😄

point 2.🤨

Please enlighten me... maybe i drew wrong from the work he did...
I did the drawing, if you are referring to the PE not shown at the inverter, it is connected on my inverter as per sunsynk manual, just not shown in the drawing.

This Neutral.

9 minutes ago, VirWat said:

This Neutral.

That was done from the sunsynk manual.
All neutrals before rcd's are common to grid neutral.
He is coming today to do the COC Tests will confirm with him.
 

Edited July 4, 20232 yr by WannabeSolarSparky

Just asking.

What's the point of having split Neutral bars and then combine them with a bridge?

58 minutes ago, VirWat said:

Just asking.

What's the point of having split Neutral bars and then combine them with a bridge?

I should have just opened up the db board

He removed that yesterday but forgot to mention it to me 🤣

Surely at the end of the day if your earth leakages are tripping with and without the grid present then that is an indication of a functional bond?

In a typical installation with TN-C-S system, your neutral and earth are bonded at your main distribution board. Your earth is shared between the Grid side and Load side of the inverter because you are supposed to earn everything to the utility earth. When the Load side earth and neutral are bonded, the earth cable provides a bond between the Grid neutral and the Load neutral. If you are doing a permanent N-E bond on the Load side, you are effectively bonding the Grid neutral with the Load neutral as well.

That goes completely against the requirements to have split neutral, isn’t that so?

 

For what it's worth -

 

Alternative Energy.pdf

Edited July 4, 20232 yr by VirWat

8 minutes ago, VirWat said:

For what it's worth -

 

Alternative Energy.pdf 695.96 kB · 2 downloads

I like the way that allows non-essentials to be fed from the inverter too.

Nice electrical drawings showing in detail.

49 minutes ago, Superfly said:

I don't understand the the non-/essential bus bars  - they feed nothing. Shared neutrals without RCB downstream is OK as the MCB will only trip on overload.

 I do not have any non-essential loads, that is for future use if needed.

On 2023/06/27 at 7:58 AM, Scorp007 said:

Just out of interest I find this method in the Sunsynk manual with respect to the MCB on the earth wire at a SPD. 

I found an explanation for the reasons behind the MCB on the Earth wire and I remembered the discussion in this thread. Anyway, apparently this is done in some cases in the States but I don't think such protection is commonly used in SA.

Hi there,

I would like some advice on the Earth Neutral bonding for a Sunsynk 5kW inverter. There is a lot of debate surrounding using relay vs permanent bond. If using a relay for the bonding however, what rating should the relay be? I have seen a few photos where a 10A relay is used. Others suggested based on the inverter power (in my case then, 25A), although if current would flow between this bond, this should trip your Earth leakage, right? I was considering a 20 Amp DIN rail relay.

Additionally, what should the wire rating be for a) the wire creating the N-E bond (probably related to the relay rating).

So my questions are:

a) N-E bond relay A rating

b) N-E bond wire rating (probably depending on to the relay rating)

c) The rating for the wire that is used for the relay coil (connected to ATS 240)