I have a Woef job for a campervan I looked at last week, I’ve failed it at this stage as I don’t think it complies. It was wired a few years ago and passed its first Woef.
The issue I have is with the ‘Dual supply systems’ and the changeover between the two.
The installation consists of two GPOs and one air-con unit. There is a standard single phase, 16amp appliance inlet for grid connection, and a 12VDC/230VAC inverter that only supplies the air-con unit when there is no grid supply.
I don’t think the relay/contactor that’s being used to switch between grid and inverter is adequate, and doesn’t satisfy the ‘break before make’ requirement, as per C7.8.1 of 3001?
C7.8.1 mentions, changeover-switch (or similar arrangement). Could such a relay be considered a ‘similar arrangement’? I think it needs to be replaced with a three position manually operated changeover switch? Or perhaps they could fit a dual contactor ATS arrangement? A manually operated switch will likely keep things simpler.
Unfortunately I didn’t get enough details off the inverter on my inspection, I only noticed it was a pure sine-wave type, I’ll have to re-check it. I’m guessing it’s an ‘isolated type’, as I’m pretty sure it didn’t have a built in RCD. So just to clarify, assuming it is an isolated type, the earth system needs to be connected to the neutral output of the inverter via a switch, and this is only connected when the inverter is the supply, as per C7.8.2.1? The air-con unit will have RCD protection when grid connected, but not when connected to the inverter supply.
I’ve attached a wiring diagram sketch, and a photo of the relay/contactor being used. Any advice much appreciated, Thanks!
Dual Supply Change-over
Re: Dual Supply Change-over
Correct that this arrangement does not comply with C 7.8.
It also doesn't comply with 3.4.4.3, so could not have been validly certified when new (if built since 1/4/2010)
That clause lists several other requirements; including voltage rating, and that it must provide "isolation' between load circuits and non-selected source. No relay can provide either, and nor can most contactors.
The relay's current rating may also be an issue, depending on the load of the a/c unit; but I suspect that like most it's only 10 A resistive.
However we don't get to apply the construction clauses when assessing for WoEF;
we only get to apply App C (plus Part 1 of the unit was built overseas)
C 7.8.1 (a) isolation between sources: fail. Relays can't provide isolation and only specially rated types of contactor can.
Installation requirement [3.4.4.3] is "when one source is connected to the installation, the second source is isolated with regard to all live conductors from the installation".
We can actually only test for non-continuity; there's no test we can do in the field that confirms "isolated".
But for any "open-circuit" or "no voltage" test result to be called isolation; it must be achieved using a device-for-isolation as per 2.3.2.2 of "3000".
And if that device is a switching device (which includes relays); it must be capable of being locked "off" .
C 7.8.1 (b) break before make: fail. The action of the relay will be pretty much simultaneous.
The requirement when installing [3.4.4.3] is "simultaneous connection of load to both supply sources shall not occur".
Not shall not normally occur, but shall (never) occur.
Break-before-make devices are more complex in order to achieve this, and therefore more expensive. Which is one reason for standard mains / genset switches to have an intermediate "off" position
C 7.8.2 supply source connections: fail
While "3001" [ 3.4.4.4] permits either isolated or RCD-protected inverters, as defined in AS/NZS 4763.
In either case there's a connection between one side of the output and earth; difference being whether it's made
- within the inverter and always in place as per 7.8.2.2; or
- within the CI and only made when loads are on inverter supply, as per 7.8.2.1.
However many inverters are not either of these classes, many being what's classed as "equipotentially bonded' instead.
These have the output earth connection bonded to case / frame, and also to d.c. negative; but no connection between thisbonding system and the output.
And some don't fit any of the classifications defined by the Standard.
Note to 3.4.4.4 tells us that a compliant "isolated inverter" is marked with the symbol for double insulation.
That's because to get the classification, there needs to be not just no connection between the output and 'earth";
but actually double / reinforced insulation "between output circuits and accessible conductive parts" - and also between input and output circuits.
This is not something we can test for, so if it isn't clearly marked we can't assume it.
Most inverters readily available and sold for use in RVs are simply not permitted by "3001".
If the inverter includes a factory-fitted RCD, you might assume it's an RCD-protected inverter; but to be sure should test polarity of output.
Continuity test N-E won't work because the RCD has to switch the N; so voltage A-N, A-E, and N-E.
You want readings as per a normal MEN-supplied circuit.
Also need to test operation of the RCD (as per c 7.2).
If the inverter has an isolated output, the c/o device needs an extra pole to switch the N-E connection;
and we have to test this switching function [7.8.2.1].
That relay is only 2 poles, not 3, so cannot be making the required N-E connection.
Despite the heading, this is NOT an "isolated supply" and can't be used as such.
It's an isolated output of the source of supply, which is then converted to simulate MEN.
There has to be a N-E connection and there has to be an RCD in the circuit.
If any final subcircuit doesn't have RCD protection (regardless of supply source); then C 6.9(d) requires a N-E link.
Another issue is voltage ratings (though not part of WoEF checks)
3.4.4.3 requires a minimum voltage rating to be i.a.w. "3010". Which means the maximum out-of-phase voltage between adjacent contacts.
So, if switching directly from one source to another, we need a 230 V x 2 = 460 V rating.
Which is the other reason most source selection switches have an "off" position between each "source selected" position, as it brings us back to a 230 V rating - being the max voltage between any source and the "off".
Bottom line:
Any "changeover device" for source selection that uses relays or contactors , or that doesn't have intermediate "off" position(s), is likely to be non-compliant in at least one respect.
It also doesn't comply with 3.4.4.3, so could not have been validly certified when new (if built since 1/4/2010)
That clause lists several other requirements; including voltage rating, and that it must provide "isolation' between load circuits and non-selected source. No relay can provide either, and nor can most contactors.
The relay's current rating may also be an issue, depending on the load of the a/c unit; but I suspect that like most it's only 10 A resistive.
However we don't get to apply the construction clauses when assessing for WoEF;
we only get to apply App C (plus Part 1 of the unit was built overseas)
C 7.8.1 (a) isolation between sources: fail. Relays can't provide isolation and only specially rated types of contactor can.
Installation requirement [3.4.4.3] is "when one source is connected to the installation, the second source is isolated with regard to all live conductors from the installation".
We can actually only test for non-continuity; there's no test we can do in the field that confirms "isolated".
But for any "open-circuit" or "no voltage" test result to be called isolation; it must be achieved using a device-for-isolation as per 2.3.2.2 of "3000".
And if that device is a switching device (which includes relays); it must be capable of being locked "off" .
C 7.8.1 (b) break before make: fail. The action of the relay will be pretty much simultaneous.
The requirement when installing [3.4.4.3] is "simultaneous connection of load to both supply sources shall not occur".
Not shall not normally occur, but shall (never) occur.
Break-before-make devices are more complex in order to achieve this, and therefore more expensive. Which is one reason for standard mains / genset switches to have an intermediate "off" position
C 7.8.2 supply source connections: fail
While "3001" [ 3.4.4.4] permits either isolated or RCD-protected inverters, as defined in AS/NZS 4763.
In either case there's a connection between one side of the output and earth; difference being whether it's made
- within the inverter and always in place as per 7.8.2.2; or
- within the CI and only made when loads are on inverter supply, as per 7.8.2.1.
However many inverters are not either of these classes, many being what's classed as "equipotentially bonded' instead.
These have the output earth connection bonded to case / frame, and also to d.c. negative; but no connection between thisbonding system and the output.
And some don't fit any of the classifications defined by the Standard.
Note to 3.4.4.4 tells us that a compliant "isolated inverter" is marked with the symbol for double insulation.
That's because to get the classification, there needs to be not just no connection between the output and 'earth";
but actually double / reinforced insulation "between output circuits and accessible conductive parts" - and also between input and output circuits.
This is not something we can test for, so if it isn't clearly marked we can't assume it.
Most inverters readily available and sold for use in RVs are simply not permitted by "3001".
If the inverter includes a factory-fitted RCD, you might assume it's an RCD-protected inverter; but to be sure should test polarity of output.
Continuity test N-E won't work because the RCD has to switch the N; so voltage A-N, A-E, and N-E.
You want readings as per a normal MEN-supplied circuit.
Also need to test operation of the RCD (as per c 7.2).
If the inverter has an isolated output, the c/o device needs an extra pole to switch the N-E connection;
and we have to test this switching function [7.8.2.1].
That relay is only 2 poles, not 3, so cannot be making the required N-E connection.
Despite the heading, this is NOT an "isolated supply" and can't be used as such.
It's an isolated output of the source of supply, which is then converted to simulate MEN.
There has to be a N-E connection and there has to be an RCD in the circuit.
If any final subcircuit doesn't have RCD protection (regardless of supply source); then C 6.9(d) requires a N-E link.
Another issue is voltage ratings (though not part of WoEF checks)
3.4.4.3 requires a minimum voltage rating to be i.a.w. "3010". Which means the maximum out-of-phase voltage between adjacent contacts.
So, if switching directly from one source to another, we need a 230 V x 2 = 460 V rating.
Which is the other reason most source selection switches have an "off" position between each "source selected" position, as it brings us back to a 230 V rating - being the max voltage between any source and the "off".
Bottom line:
Any "changeover device" for source selection that uses relays or contactors , or that doesn't have intermediate "off" position(s), is likely to be non-compliant in at least one respect.
- Rating: 33.33%
Re: Dual Supply Change-over
Thanks Alec, appreciate the detailed reply.
Regarding the ‘installation requirement’ for RCD protection for the air-con circuit, I believe it would be required regardless of the source of supply, as per clause 3.3.2(a) it mentions ‘every final subcircuit’, so this would apply for either grid or inverter supply?
I’m a bit confused when it comes to the Woef requirement. When the CI is grid supplied it complies with C6.9(b). Assuming the correct change-over switch was fitted, would it comply with C6.9(d) when it was inverter supplied? It states ‘neutral/earth link or connection’. If the inverter was an ‘isolated supply’, there would be a ‘neutral/earth connection’ at the changeover switch, as detailed in figure 3.1? Therefore making the air-con circuit comply with C6.9(d)?
So RCD protection should’ve been fitted at installation stage, but could maybe not be picked up on during a Woef check?
Regarding the ‘installation requirement’ for RCD protection for the air-con circuit, I believe it would be required regardless of the source of supply, as per clause 3.3.2(a) it mentions ‘every final subcircuit’, so this would apply for either grid or inverter supply?
I’m a bit confused when it comes to the Woef requirement. When the CI is grid supplied it complies with C6.9(b). Assuming the correct change-over switch was fitted, would it comply with C6.9(d) when it was inverter supplied? It states ‘neutral/earth link or connection’. If the inverter was an ‘isolated supply’, there would be a ‘neutral/earth connection’ at the changeover switch, as detailed in figure 3.1? Therefore making the air-con circuit comply with C6.9(d)?
So RCD protection should’ve been fitted at installation stage, but could maybe not be picked up on during a Woef check?
Re: Dual Supply Change-over
I included some references to installation requirements because while we can't enforce them via the WoEF process, they can assist our understanding of the WoEF requirements.
Yes, the installation requirement for all final subcircuits to have RCD protection applies at all times.
The FS can't have RCD protection sometimes (eg when on external supply) and not at other times (eg when supplied by inverter).
This requirement isn't carried through directly to WoEF checklist, because of the need to allow for older units that don't have RCD.
The N-E connection is similar.
Older caravans (only caravans) were required to have one.
These days Section 3 doesn't call for one - ie there's no installation requirement, either to install or not to install.
And if the installer put one in, the result would be that the RCD (if there is one) for the socket plugged into will trip.
The WoEF requirements are covered in C 6.9; where the unit must be in 1 (and only one) of 4 configurations;
a) if it's not a caravan; there must be no (permanent) N-E connection.
- but if there's an on-board supply, there will of course be a N-E connection for that; either permanent (for RCD-protected source) or switched (for isolated-output source).
b) if it's a caravan; and all FSs are RCD-protected (at all times); then no (*permanent) N-E link permitted
- but again, there will be N-E connections for in-board supplies
c) only relates to where RCD protection (for external supply) is in the supply lead as against within the unit
(a legacy provision from the period when this was the required configuration)
d) requires a (permanent) N-E connection if any subcircuit doesn't (always) have RCD protection.
C 6.4 is similar; where the current limitation is only required if there's no (permanent) N-E link.
To be honest, until now I've never even considered the possibility that a unit could be in a different category WRT C 6.9 depending on which supply source is selected. However I'm certain the clause wasn't written with that in mind; it was written solely to cover the fact that over time there have been different required configurations for the external supply.
The current (2008) edition was the first one to include any provisions for on-board supplies; so any unit that has an on-board supply (other than imported units) should have been built to comply with this edition. This one simply doesn't.
Agree the lack of RCD protection for the circuits supplied by the inverter in this case must have been missed in earlier WoEF checks.
Unfortunately a lot of inspectors don't bother to use a checklist, and some are woefully out of date.
There's a general principle that whenever we set up an alternative supply, the earthing configuration of each supply available must be the same.
So if the normal supply is MEN (TNCS); then any alternative supply has to be converted to an equivalent configuration.
Hence the N-E connection made within the inverter / genset for RCD protected sources, and by source selection switch for isolated output sources.
The reason is so that a circuit designed for one configuration doesn't get connected to a source of a different configuration.
Which brings up another problem with the unit you've been looking at.
The final subcircuits are required to have fault protection (requirement in "3000" clause 2.4, not modified by anything in "3001").
For external supply, this is provided, as usual, by automatic disconnection of supply. If an earth fault occurs, either an overcurrent device or an RCD trips within specified time.
But when supplied by this inverter, there's no fault protection. The inverter simply can't provide enough current to operate an overcurrent device within the time limit. An RCD would work within time on the low fault current available - but there isn't an RCD.
Can't get fault protection by separated supply, because the requirements for separated supply (clause 7.4 of "3000) have not been met.
In particular, 7.4.5.
Whoever designed this unit thought they were being clever,
But in fact they were just being ignorant - not only failing to follow the detailed requirements for on-board supplies, but also forgetting the underlying safety requirements. And as a result, the unit is not only non-compliant, but bordering on electrically unsafe.
Yes, the installation requirement for all final subcircuits to have RCD protection applies at all times.
The FS can't have RCD protection sometimes (eg when on external supply) and not at other times (eg when supplied by inverter).
This requirement isn't carried through directly to WoEF checklist, because of the need to allow for older units that don't have RCD.
The N-E connection is similar.
Older caravans (only caravans) were required to have one.
These days Section 3 doesn't call for one - ie there's no installation requirement, either to install or not to install.
And if the installer put one in, the result would be that the RCD (if there is one) for the socket plugged into will trip.
The WoEF requirements are covered in C 6.9; where the unit must be in 1 (and only one) of 4 configurations;
a) if it's not a caravan; there must be no (permanent) N-E connection.
- but if there's an on-board supply, there will of course be a N-E connection for that; either permanent (for RCD-protected source) or switched (for isolated-output source).
b) if it's a caravan; and all FSs are RCD-protected (at all times); then no (*permanent) N-E link permitted
- but again, there will be N-E connections for in-board supplies
c) only relates to where RCD protection (for external supply) is in the supply lead as against within the unit
(a legacy provision from the period when this was the required configuration)
d) requires a (permanent) N-E connection if any subcircuit doesn't (always) have RCD protection.
C 6.4 is similar; where the current limitation is only required if there's no (permanent) N-E link.
To be honest, until now I've never even considered the possibility that a unit could be in a different category WRT C 6.9 depending on which supply source is selected. However I'm certain the clause wasn't written with that in mind; it was written solely to cover the fact that over time there have been different required configurations for the external supply.
The current (2008) edition was the first one to include any provisions for on-board supplies; so any unit that has an on-board supply (other than imported units) should have been built to comply with this edition. This one simply doesn't.
Agree the lack of RCD protection for the circuits supplied by the inverter in this case must have been missed in earlier WoEF checks.
Unfortunately a lot of inspectors don't bother to use a checklist, and some are woefully out of date.
There's a general principle that whenever we set up an alternative supply, the earthing configuration of each supply available must be the same.
So if the normal supply is MEN (TNCS); then any alternative supply has to be converted to an equivalent configuration.
Hence the N-E connection made within the inverter / genset for RCD protected sources, and by source selection switch for isolated output sources.
The reason is so that a circuit designed for one configuration doesn't get connected to a source of a different configuration.
Which brings up another problem with the unit you've been looking at.
The final subcircuits are required to have fault protection (requirement in "3000" clause 2.4, not modified by anything in "3001").
For external supply, this is provided, as usual, by automatic disconnection of supply. If an earth fault occurs, either an overcurrent device or an RCD trips within specified time.
But when supplied by this inverter, there's no fault protection. The inverter simply can't provide enough current to operate an overcurrent device within the time limit. An RCD would work within time on the low fault current available - but there isn't an RCD.
Can't get fault protection by separated supply, because the requirements for separated supply (clause 7.4 of "3000) have not been met.
In particular, 7.4.5.
Whoever designed this unit thought they were being clever,
But in fact they were just being ignorant - not only failing to follow the detailed requirements for on-board supplies, but also forgetting the underlying safety requirements. And as a result, the unit is not only non-compliant, but bordering on electrically unsafe.
- Rating: 33.33%
Re: Dual Supply Change-over
Many thanks Alec. Yes, I believe it was wired this way to create easy change over from one supply to the other, with no manual switching required. Unfortunately not enough thought has gone into compliance.
Re: Dual Supply Change-over
It’s odd they haven’t clarified the ‘break before make’ requirement in clause 3.4.4.3 like they have in clause C7.8.1. They do mention ‘change over switch’ in 3.4.4.3, guess it’s standard for a change over switch to be break before make.
Re: Dual Supply Change-over
Agree it could have been better; and not just on this point.
2008 was the first edition to even consider in-board supplies; now they are common and sometimes more than one on-board source.
New edition will be published shortly and provides more guidance.
Including for where not everything is to be switched between supplies.
Eg if you have a battery charger to top up batteries when external supply is available; would be foolish to switch that subcircuit over to be fed from an inverter that's fed from those batteries.
However the requirement for isolation of non-selected sources remains, and that's what leads to many otherwise acceptable arrangements being non-compliant.
2008 was the first edition to even consider in-board supplies; now they are common and sometimes more than one on-board source.
New edition will be published shortly and provides more guidance.
Including for where not everything is to be switched between supplies.
Eg if you have a battery charger to top up batteries when external supply is available; would be foolish to switch that subcircuit over to be fed from an inverter that's fed from those batteries.
However the requirement for isolation of non-selected sources remains, and that's what leads to many otherwise acceptable arrangements being non-compliant.