Seems to have now been posted to BSI, comprising the changes suggested by Lyle and Andy (I haven't checked if they are verbatim).

If this is all it is... I confess I need to check the books but is this not already a requirement?

Also, to have been rushed out in this fashion, presumably this is in response to a specific incident(s)? Either way would it not be reasonable for BSI/IET to publish a supporting note justifying the change; this should illustrate where there is a (percieved) deficiency / gap in the existing standard.

826.1.2.2

But it seems that not every installation that containing local generation would be a PEI and subject to part 8 - a simple (inverter only) PV system simply injects all available power into the system - regardless of the needs of the local installation or indeed the wider grid. It has no management system. Thus it seems to not meet the requirements for a PEI (e.g. 823 (iv) or 822.1).

Presumably the implication is that the new requirements apply not just to the PV circuit itself, but also everything that connects it back to the grid (presuming it's capable of export at times).

So this incomer would be a problem:

  - Andy.

In relation to that proteus main switch, it's not a protective device so I don't think it would be captured.

But I've done some survey of the standards and the language about directionality changes between them. Notably MCCBs have to use an arrow symbol and not the words "supply" & "load".

One concern I've got is that the underpinning standards only make fleeting references to directionality and have poor definitions of what directionality in fact is. For example any AC load at non-unity power factor technically has bidirectional flow of power and it would be a more natural use of language to refer to these devices as been uni/bi-directional in respect to flow of energy.

I think we can categorised the concerns into the following broad classes:

1) are the devices able to interrupt a fault current in the reverse direction without suffering damage? This appears very difficult to achieve because the electronics must withstand the Isb [>6kA] being interrupted adjacent and causing a very large transient right on its terminals likely far exceeding the surge arrestors capability.

2) are the devices able to interrupt a residual current in both orientations? This appears trivial - hence almost every RCCB on the market presently is already bidirectional.

3) are the devices likely to give rise to a hazard if fed from the "load" side? This is a potential concern but it's not actually addressed in the product standards - it's quite possible that there's bidirectional RCBOs out there that could for example develop an internal fault which depends on MCB function to interrupt and therefore could be hazardous in these appliancations - this proposed ammendment does nothing about this problem and if anything it could create false security. My gut is that we should ensure any source of power has sufficient protection at it's terminals to protect against this kind of fault - but that's not trivial either because inverters have a relatively low short circuit current so how can you protect against an L-L fault when the Ipf is in say 32A?

I'm wondering what evidence such as incident reports there are behind this proposal - BEAMAs report notably lacks case studies.

In relation to that proteus main switch, it's not a protective device so I don't think it would be captured.

The proposed 530.3.201 seems to include "equipment for protection, isolation, switching, control and monitoring".

I have a feeling that a lots of devices got labelled supply/load not because of any physical limitation of the device, but more as a convenience to the installer (so when in place in a typical CU the terminals would have helpful labels) or indeed the manufacturer (saves having to put the labels on the CU part of the assembly or have an extra manufacturing step to label the devices after insertion).

   - Andy.

Ouch, you're correct.

I was noticing the definitions wording which is limited to protective devices but proposed 530.3.201 as presently drafted has a far wider scope.

How can equipment for isolation possibly be directional? Thermonic emissions of electrons? Well if it's glowing there's bigger issues to worry about.

Keir Stitt said:

How can equipment for isolation possibly be directional? Thermonic emissions of electrons? Well if it's glowing there's bigger issues to worry about.

A very simple example is a switch-fuse where the fuses remain connected to the pole side, where, to be able to replace fuses without operating another isolator, you'd of course want to connect a source to the switch-break side only.

With DC the arc is certainly directional, and one side of the contact gap is more consistently eroded by the arc than the other.

A very simple example is a switch-fuse where the fuses remain connected to the pole side, where, to be able to replace fuses without operating another isolator, you'd of course want to connect a source to the switch-break side only.

Fair point - and known issue when using single pole devices to feed 3-phase loads, for example. Which suggests again this is really a matter of voltages being present, rather than the actual direction of power flow.

As an interesting "historical curiosity" BS 7671 2008 contained:

  -  Andy.

AJJewsbury said:

As an interesting "historical curiosity" BS 7671 2008 contained:

Agreed - that applied only to PV, and ONLY to devices for isolation and switching, NOT protective devices (even though they must also be suitable for isolation if used for ADS).

However, BS 7671 no longer has that requirement.

AJJewsbury said:

Thus it seems to not meet the requirements for a PEI (e.g. 823 (iv) or 822.1).

I give you Schrödinger's RCBO. Bidirectional in the MI's, unidirectional according to it's terminal markings.

In fairness I've brought this to attention of Proteus who are changing their terminal markings.

I give you Schrödinger's RCBO. Bidirectional in the MI's, unidirectional according to it's terminal markings

I've a feeling there may be a lot like that - I asked another manufacturer about their single module RCBOs who replied that there wasn't a safety problem with them feeding a PV system, but if I wished to upgrade they're recommend the 2-module version (which didn't have the SUPPLY/LOAD labels). A new version of badge engineering perhaps.

   - Andy.

I use Acti9 iC60H plug on neutral - 1P + Ns. These have the markings for Lin and Lout etc, does that mean the Schneider, most expensive RCBO in the world is not bidirectional? 

I use Acti9 iC60H plug on neutral - 1P + Ns. These have the markings for Lin and Lout etc, does that mean the Schneider, most expensive RCBO in the world is not bidirectional?