Growatt 5000ES advice

Hi all
I am installing the above inverter at home ( yes , yes , I wanted to buy a Victron but the budget is not there ) . I need some pointers please :
As this inverter has pass-through does that imply that the inverter syncs the inverter freq to the AC grid freq ? Otherwise how does it add power from the grid to the power generated by the inverter ?
The next question concerns a common neutral. I am installing a second DB for the inverter loads ( at this time only one plug cct from the main DB ). Can I safely permanently link the main DB neutral bar to the neutral bar in the 2nd DB and only bring the ā€˜liveā€™ of the plug cct across to its own MCB in the 2nd DB or must I bring the neutral across as well and not bond the neutrals ? Surely on pass-through the incoming mains supply is placed in parallel with the inverter supply bonding the neutral together anyway.

Regarding neutrals. When in islanded operation (ie when the grid is down), the output neutral cannot be connected to the input neutral. Both has to be disconnected. Hence you cannot permanently tie the neutrals together, and that means the neutral has to come acrossā€¦ and go back.

The plug circuit you want to isolate: You need to isolate both its live and neutral and bring both across to the new DB.

Thank you Izak , much appreciated. So, the new DB will have an isolator / CB ( white / green toggles) , an earth leakage unit as well as an 20amp MCB for this new plug cct I presume.That is the way I planned it anyhow. My initial thinking was the same as yours but then this pass-through facility threw me a bit.

My recipe in the past is like this.

In my main distribution board I have a double-pole breaker (rated for the cabling) that goes to the sub-db. Yes, you can get away with just a single pole here, but I like double. This is tapped BEFORE the RCD, so that the inverter does not have an RCD on its input (it doesnā€™t need one).

In the sub-DB, from left to right, I install:

  • A double-pole main breaker
  • A double-pole changeover switch (typically Hager SFT240)
  • An RCD
  • One or more breakers for the backup circuits (typically lights and plugs).
  • A warning light indicating that the power is on

Wiring in the sub-DB is as follows:

  • The feed from the main DB (live and neutral) goes to one end of the changeover, and from there to the input of the inverter as well. On the SFT240 this is on the bottom of the device.
  • The output of the inverter (live and neutral) goes to the top of the main breaker
  • The output of the main breaker goes to the other side of the changeover switch. Also on the bottom.
  • The common side of the changeover (at the top) goes across to the top of the RCD.
  • The live conductor from the bottom of the RCD loops over to a busbar linking the other breakers.
  • The neutral from the RCD goes to a separate neutral bar.
  • All circuits on backup is them moved over (live and neutral) to the sub-db.

The logic here is that there is only one RCD for the backup circuits, the one in the sub-DB, and it is in used regardless of whether you have the inverter bypassed or not. Makes a little more sense imho.

The transfer switch simply bypasses the inverter and connects the RCD directly to the incoming feed.

You have overcurrent protection in the main-DB for everything downstream, but you also have a main breaker protecting power going through the inverter (if the inverter adds power, then you could have more current coming out of it than going in).

Finally, also pull across a nice proper separate earth conductor, Iā€™d say at least 4mm^2 or 6mm^2 if you can. The inverterā€™s inputs, outputs, and all circuits, are earthed together to the earth bar in the sub-DB, which is in turn earthed to the main DB using a nice thick conductor.

Thank you Izak , very well presented and understandable. Will get to work on my setup.

Nice big metal case makes me nervousā€¦ Without an RCD youā€™d need a full-current fault to trip the breaker, right?

I have decided to feed my inverter from the main DB , after the earth leakage , from a separate 40 amp CB and introduce another earth leakage in the sub DB on the inverter output.Can you foresee any problems ?

But before the changeover switch? So youā€™d only ever have the one in-circuit at a time?

If you are concerned about that, my suggestion would be to fit a 300mA earth leakage in front of the inverter. And mount the inverter high enough that kids canā€™t touch them. And ground it very well.

And then you are not allowed to hang any sockets off this circuit (before the inverter) either, but that goes without saying.

Of course fitting yet another double pole breaker can be an issue. Space is always at a premium in distribution boards. So then youā€™re looking at a combination overcurrent+rcd, a 300mA RCBO. You get them, and I have one installed in my system, but they cost a bit of money (somewhere between R500 and R1000).

There is a reason I prefer not to have two 30mA RCDs in series. Because one of them will be more sensitive than the other, and then you have no selectivity. Either one could trip (while on grid). You want the nearest one to trip (especially if the upstream one is in another building). If you must have two in series, the upstream one should either have a delay (type S) or a higher rating.

Yes, at our office and callcentre we also split RCD circuits very finely. Electricians thought we were crazy, but with so many switching power supplies on a circuit, you can get to 30mA easily. Especially when staff plug in some dodgy phone charger thatā€™s gone through the wash twice. In any case we now would only lose about 10 agents for a single trip. The dropped calls cost a lot more than the equipment.

For a home installation I would worry a bit about non RCD on the input, but then you automatically have two in series which can easily cause other problems. But I must say, whenever Iā€™ve spent money on a ā€œniceā€ RCD, I havenā€™t regretted it.

Putting the inverter on an RCD might cause a fault when neutral bonding/unbonding depending on how your neutral floats.

My Victronā€™s ground relay with a Schneider 30mA were good friends. ACDC, not so much.

Amen to that.

If the inverter properly disconnects the neutral before closing its internal TN bond, it should not trip the upstream RCD.

My issue with series RCDs is simply that you get into weird situations, such as a nuisance tripping issue that trips only the upstream RCD (but you donā€™t notice immediately because the inverter picks up the outage), of it there is a real issue, it trips both, making it hard to figure out where the problem is. If you are okay with a bit of weirdness, but all means put two 30mA RCDs in series. It works just fineā€¦ Iā€™ve done it myself.

I have an older MultiGrid, that used a slightly less RCD-friendly way to test TN bonding. In my previous house, this tripped both Hager (type AC) and ABB (type A) RCDs every time it did the relay test. I ended up with what ABB calls type A-APR (which is thankfully cheaper than a type F, which is basically unaffordable) and that solved the problem.

Moved to a new house. This place has good old CBI Samite type breakersā€¦ and they have no nuisance tripping issues. Totally worth the R700+ you pay for these things.

Buuutā€¦ the poster is asking about a Growatt thingamabob. No idea if that even does TN bonding on the output (many of the Axpert clones donā€™t, no idea where this one is on the spectrum, looks a bit like a rebadged Infini). No idea if it even does a relay test. All of these RCD-type issues might be moot in this case. A cheaper 30mA downstream and a slightly less-cheap 300mA RCBO upstream should be fine.

Thanks to you guys I can see ā€˜more clearly nowā€™. Getting all the bits and pieces together for the build. I have decided to connect the inverter through its own 40amp breaker ( before the EL ) in the main DB like the geyser , aircon , stove etc. In the secondary DB I will have a 30mA EL and all the other overload protection devices as required. At this time I will only transfer 1 x plug cct to the inverter output side. By next week I will install it all and test the pass-through function.
Also adding a 3rd DB with a 4 pole (2NC + 2NO) contacts to switch the DC from the PV array
to either the Elon 100 geyser controller or the inverter PV input. Currently all the PV power is going directly to the Elon 100 and by around 12h00 the geyser is on temp. This system has been working here for the last 3 months and I am very impressed. So the power generated by the PV goes to waste in the afternoon.
With pass-through the mains will add to the power generated by the inverter if there is a deficit.
If you hear an explosion in eastern Durbanville next week: ā€œIt was meā€
Will post more as the project progresses
Very interestingly , this Growatt 5000ES is very popular in the USA but at this time they only get the 230v version , so for their weird 120v requirements they need to add a auto transformer to create a centre point neutral for a 120-N-120v setup.

This goes without saying, but please make sure thatā€™s DC rated :slight_smile:

Battling a bit to find a contactor with DC rated contacts. I will , most of the time , switch on zero current flow.

The bigger brands have DC ratings in their spec sheets, even if the thing isnā€™t marketed as DC. You really want to get this part right, even though itā€™s tedious.

1 Like

The 100mA units are a lot cheaper then the 300mA units and the difference is big enough that an earth fault on the output only trips the output RCD. Personally I prefer the 100mA unit.

Using two 30ma Units in series normally results in both tripping with an earth fault on the output, leaving you with no AC supply to the inverter. (I have learned this the hard way on a clients site way in the beginning)