Unbalanced loads on a 3 phase RCD

[binky]

I've been studying some diagrams of 3 phase RCDs, as far as I can see, because all 3 phases pass through the same detection coil, then any imbalances on the phases doesn't affect the RCDs' operation?

 

[Andy™]

yep. only an imbalance that misses the RCD, like to earth

if it didnt, id like to see you balance all 3 phases to within 30mA of each other

 

[Evans Electric]

I'd say thats been answered.

 

[binky]

Just wanted to check. I've been looking at a caravan site that has a pretty ropey set of TT electrics. They want solar panels, but there isn't enough space for a 3 phase system. A small single phase system is unlikely to back-feed much / anything to the grid, but the last thing I need is a system that starts tripping the incoming 100mA Type S 3 phase RCD.

 

[Geoff1946]

I would run that idea past the supply company before committing any expenditure. They may not be too keen on a micro-gen system that unbalances the supply.

 

[Tony S]

A single phase input to the local three phase grid shouldn’t be a problem. It’s no different to a house with a single phase supply which has a PV array feeding in to the grid.

If you start pumping 10’s of kW in to the local system then it could get hiccups but from what you say that shouldn’t happen.

 

[binky]

I would run that idea past the supply company before committing any expenditure. They may not be too keen on a micro-gen system that unbalances the supply.

not an issue, can't get a system big enough on this site to cause problems like that.

 

[Sidewinder]

It would be disconnection times I'd be worried about, feeding in solar, 5s disconnect time, downstream of an RCD which needs to disconnect in a lot less than that, unless you can guarantee that the solar is truly and totally unreferenced to true earth.

 

[binky]

the RCD in question is only protecting sub-mains and could be replaced with 300 or 500mA unit. 30mA protection is provided where it is needed to protect people.

 

[Sidewinder]

That still doesn't matter though.

If the solar has an earth reference to the inverter output, i.e. if in the event of the upstream device disconnecting, current can flow from the live of the inverter output to earth, then the disconnection time will be that of the solar inverter.

Which iirc according to product standards is 5s.

Whereas the max disconnection time on a TT is 1s.

 

[binky]

contravening info, 1 second for ccts / distribution ccts over 32A, but reading GN1, section on RCDs 3.7.7 in my copy, it talks about the actual requirement of TT systems is to restricting touch voltages to 50V. My older version of GN8 (2004) also describes this as being the actual requirement for TT systems and a disconnect time of 5 seconds as inferred from the appropriate equation.

 

[Sidewinder]

Well BS7671 says what is in the attached pdf.

View attachment TT Disconnection Times.pdf

 

[binky]

I don't have my big yellow book to hand, but that table refers to circuits protected by an OCCPD, not RCDs

 

[Andy™]

an inverter must not share an RCd with anything else. if someone comes into contact with a live wire and the RCD trips within 0.2 seconds, the inverter could still be supplying power for a further 5 seconds

also, if any device is providing ADS for any circuits then the inverter must be connected to the supply side (712.411.3.2.11)

 

[Sidewinder]

I don't have my big yellow book to hand, but that table refers to circuits protected by an OCCPD, not RCDs

No it relates to disconnection times.

Nothing to do with how disconnection is achieved, just that it MUST be achieved in those times.

 

[binky]

an inverter must not share an RCd with anything else. if someone comes into contact with a live wire and the RCD trips within 0.2 seconds, the inverter could still be supplying power for a further 5 seconds

also, if any device is providing ADS for any circuits then the inverter must be connected to the supply side (712.411.3.2.11)

I'm well aware of that Andy, but any circuit anyone is likely to get belted from has 30mA protection at sub-boards. The Sub- ain in question os buried from the incoming point to a distribution cupboard in the centre of the caravan site

 

[binky]

No it relates to disconnection times.

Nothing to do with how disconnection is achieved, just that it MUST be achieved in those times.

 I've been reading the GN books, now mine are a little out of date, ie 2004 and 2008. The earlier version states quite categorially that disconnection time is not the main criteria for TT when disconnection times cannot be met and RCD protection is introduced to compensate and limit touch voltages to <50V , by reference to a formula, which implies a disconnection time of 5 seconds is acceptable. This advice seems to be getting watered down over time.

 

[Rob_the_rich]

Would another 100mA s-type RCD installed at the output of the inverter provide the required disconnection time?

 

[steptoe]

 I've been reading the GN books, now mine are a little out of date, ie 2004 and 2008. The earlier version states quite categorially that disconnection time is not the main criteria for TT when disconnection times cannot be met and RCD protection is introduced to compensate and limit touch voltages to <50V , by reference to a formula, which implies a disconnection time of 5 seconds is acceptable. This advice seems to be getting watered down over time.

From my understanding the <50v touch has been superseded by time if disconnection, not purely limited by the voltage rise.

Which, IMHO, can only be a good thing, after all, 49v at 100amps is still one helluva kick for the next 4.9seconds.!

 

[Sidewinder]

I can't put quotes from the regs straight into the forum because the formatting is lost and it looks nonsense.

Attached, the relevant regs.

TT, disconnection times, must be met, 411.5.3.1.

View attachment Chapter 41 Protection Against Electric Shock Excerpt 1.pdf