Multiplus2 And Grounding

Hello everyone

I was with a customer who has problems with the shore connection on a new vessel that is under warranty and asked me for my opinion because I have installed Victron equipment several times. I just looked at the matter and looked through the plans and tried different scenarios of the system’s operation via the Cerbo screen. So the problem becomes in the marina when you want to connect to the Shore. Because at some point on the pier the RCD switch disconnects.

When checking the connection of the multiplus system, I noticed the following things: The housing grounding is not done, the grounding is also not present on the AC IN…
The grounding from AC Out1 is connected to the wire marked as Grounding for ACIN, there are no other grounding wires. I can’t find where the grounding leads in the plans so I added ? in the picture

The Service system consists of:

2x Multiplus 24/5000/120
1x Lynx smart BMS 500
3X Lynx dis. 1000
5x 25.6V/200Ah Smart

Is my thinking correct?
When the inverter is operating, N and PE are closed, so does the RCD trip on shore?

  • Necessarily ground the housings - both (I noticed a negative busbar nearby in this room, I’m 99% sure it’s from DC consumers),
  • connect the AC In grounding as planned?
    -AC Out Grounding from both multipluses is connected to the negative busbar in this room - like the housings?

The vessel is of course also equipped with a Galvanic isolator.

I’m adding a couple of pictures for easier understanding.

Here I have marked with a green line how the grounding is done with a wire for ACIN and ACOUT

Thank you in advance.

Since NO isolating transformer was installed in this setup… it appears that galvanic isolation of the protective conductors/grounding was attempted here.
In general, however, this is not a problem as long as the grounding relay on the boat is constantly activated and a RCD breaker has been installed here. The hull of the boat MUST, of course, be included in the grounding.
Is it a steel- or fiberglass hull?

Boat is from fiberglass. Doesn’t it bother you that the AC In and the multiplus housing are not grounded?

Since I am not familiar with the regulations and installations on site and do not know how the system is configured, I can only say… it does not look particularly good, but it would work. Perhaps the RCD switch of the marina supply always tripped and then this was built as an emergency solution… :thinking:

However, since it is a fiberglass boat, I would definitely connect the PE/protective conductor.
It would be interesting to see the boat’s technical equipment… Perhaps there are metallic parts of the system/engine/bow thruster, etc. that come into contact with the water… This must then, of course, be considered separately so that galvanic corrosion does not occur when connecting to the marina supply…

So the boat is made in the EU and built according to EU standards. But the boat is 78 feet (23.7744 m) long, and there is a lot of electrical equipment. In the engine compartment there is another system 2x multipluss 10000VA with MG bms and batteries, there is also a generator in the vessel, everything works normally on the generator…

I have a diagram of the entire vessel, but what confused me the most was the grounding in the service, which I mentioned in the first post. (2x5000va multiplus)

A technical drawing of the entire system would be helpful here…
The isolating transformer does not have a protective conductor… so if the supply line for the multis came from the isolating transformer, using the protective conductor would be useless…

The complete documentation is only 383 pages :slight_smile:

383 pages :slight_smile: … Professional, of course :wink:
Therefore, I would be surprised if the PE had been “forgotten.”
The only interesting part would be the section on" isolation transformers to the multis" …

Here is the shore input to the boat, if you look at the plan, the consumers are only marked with PE. So everything goes to the busbar… I think

so there is a missing wire leading to this location?

When using an isolation transformer, the PE conductor must NOT be connected… see circuit diagram and Victron drawing…

The boat must be earthed completely separately from the shore connection… Otherwise, the isolation transformer for galvanic isolation is usesess!!!

That could be the reason why the RCD tripped from the marina’s power connection…

1 Like

Do you know the difference between an isolation transformer and galvanic protection?

If you look at my schematics again, only PE is interrupted with a Galvanic ISOLATOR, This vessel does not have an ISOLATION TRANSFORMER

It’s not unknown for a Multiplus to trip a RCB on its input as it switches over to AC in supply. This is due to the protective sequencing of connecting the input L&N and then disconnecting the NE bridge on AC out. This leaves input N&E connected for a brief time - well sufficient to trip an RCD.
Those Galvanic isolator blocks are just big non polarised capacitors. Whilst the will stop DC, they wont, and aren’t designed to stop AC. So an AC leakage current can still flow through the hull to shoreside.
Fitting an isolating transformer IS the ONLY solution to this type of problem. Also floating the input PE conductors is not a good idea without the transformer. The case of the Multiplus is internally connected to the PE input & output grounds - these are all common connected. So having the case stud ‘not connected’ is not an issue.

By Appendix B of the Victron manual for this Multi, both AC-In and AC-Out both have a PE connection; AC-In to the source and AC-Out to the panel board and its PE bus. They both are connected to the case. Note also that Appendix B shows a case ground and the case ground has not been discussed by the OP.

The governing document for electrical systems on boats in the EU is ISO 13297 the source for this excerpt:

11.5 A separate DC equipotential conductor shall be connected from the metallic case or chassis of the inverter or inverter/charger to the main grounding/earthing point or its bus, and be of an amperage rating equal to the DC positive conductor. This conductor shall not be connected to the DC negative at the inverter or inverter/charger.

Regarding the nuisance tripping of the shore RCD:

As was discussed; there is a life safety requirement to bond neutral (N) and ground (PE) at all electrical sources that provide AC to a vessel. These sources include:

  • Shore power: N/PE bond is at the shore side transformer providing power to the shore pedestal.
  • Onboard generator
  • Secondary of an isolation transformer
  • Output of an inverter
  • Output of an inverter/charger when inverting

From Mike D Dec 29: This is due to the protective sequencing of connecting the input L&N and then disconnecting the NE bridge on AC out. This leaves input N&E connected for a brief time - well sufficient to trip an RCD.

Victron inverter/chargers have two relays in the AC-In electrical path: an input relay and the N/PE relay. When shifting modes from pass through to inverting or from inverting to pass through, the two relays are controlled so that the required action of the N/PE relay has completed before the input relay is closed. This ensures that the shore side RCD will never sense an improper N/PE bond aboard.

From MikeD Dec 29: Those Galvanic isolator blocks are just big non polarised capacitors. Whilst the will stop DC, they wont, and aren’t designed to stop AC. So an AC leakage current can still flow through the hull to shoreside.

A galvanic isolator is designed to block galvanic current, hence the name, while still providing a path back to the shore power source for AC fault current. The galvanic current blocking is accomplished by two diodes in series pointing towards the vessel and two diodes in series pointing away from the vessel. Since diodes can be biased into conducting by low level AC, a capacitor is placed in parallel with the two pairs of diodes to provide a path for AC and to prevent the diodes from conducting.

Since a galvanic isolator is in the PE circuit, it is a life safety piece of equipment as the PE circuit must remain intact. Since diodes can fail in either shorted (loss of galvanic isolation) or open (loss of a path for fault current), galvanic isolators.

From ISO 13297: 6.9 When a galvanic isolator is fitted in the protective conductor, failure of the isolator shall not result in an open circuit.

In the USA, this requirement formerly met by either electronically monitoring the condition of the PE circuit or by use of “Fail Safe” galvanic isolator. With the advent of the requirement for RCD’s on shore power pedestals, the monitoring system has all but disappeared because this system momentarily connected PE > N and the RCD would trip so we must install Fail Safe GI’s.

If I was faced with this problem, I would:

  1. Install the PE connections as shown in the techdocs.
  2. Install a properly sized DC case ground for each Multi back to the main B- bus.
  3. Ensure that there are separate N busses; one for the non-inverter supplied loads and the other for the inverter supplied loads.
  4. Ensure by hand over hand or by ringing out that circuit discipline was maintained when the N from the two sets of loads were landed on a N bus. It will only take one conductor from an inverter supplied load landed on the N bus for the non-inverter supplied load to cause this issue.