This is not quite as simple to answer as it once might have been thought.

• Can you be sure the structured cabling will always be used for Ethernet? PSTN for example is not necessarily SELV or PELV.
• Unfortunately the BS EN 62368 series of standards no longer uses SELV and PELV definitions, and the new definitions ES1 and ES2 don't map directly to SELV sources - especially in single-fault conditions. This is particularly problematic for Power Over Ethernet applications ... but potentially plain Ethernet too (although at the moment isolation is still included in the Ethernet standard).

Ethernet has signals of a couple of volts peak to peak at the transmitter end, falling to about half that at the far end for the fastest sequence of 10101 that s supported, lower frequencies, such as 11001100 experience less loss. The idea that regs requiring an RCD apply to this would be  a disaster waiting to happen.

Power over ethenet rides a current limited Dc between the two earth references of the line pairs on which the data is imposed, and in some designs on unused pairs as well. The normal power ADS considerations are not really suitable. I suggest that B7671 is the wrong standard, and those relating to network cabling are more suited.

PSTN should be installed as per the BT SNs unless there is a good reason to deviate

Mke

110.1.2 (vii) specifically includes this in scope.

Also, BS EN 50174-2 has HD 60364 as a normative reference.

mapj1 said:

Ethernet has signals of a couple of volts peak to peak at the transmitter end, falling to about half that at the far end for the fastest sequence of 10101 that s supported, lower frequencies, such as 11001100 experience less loss. The idea that regs requiring an RCD apply to this would be  a disaster waiting to happen.

Power over ethenet rides a current limited Dc between the two earth references of the line pairs on which the data is imposed, and in some designs on unused pairs as well. The normal power ADS considerations are not really suitable. I suggest that B7671 is the wrong standard, and those relating to network cabling are more suited.

Agreed ... but whilst the source of those signals or voltages under BS EN 60950-1 used to be SELV/PELV, now that has been replaced by BS EN 62368 series, which doesn't have the equivalent of SELV and PELV,, it becomes more tricky.

That doesn't mean every product is not safe nor not compliant, but how is that demonstrated now? More to the point, are there products out there that are not equivalent to SELV/PELV?

And of course, an RCD is not feasible for that type of signal - neither, at present, could it be done for pure DC regardless of whether it's LV or ELV.

Well either BS7671 needs to develop a section on network installation, and include things like ensuring cable impedance, high bandwidth termination methods, special kinds of splice etc, and also sections on telephones, broadband,  TV antenna cables and door bells, audio video and every other thing you can think of, or more sensibly should say - We have other standards that apply to those. Use them instead.  I think it is a hiding to nothing to assume that the entire body of standardization can all be encompassed by one BSI document, when other organizations have been doing a perfectly good job elsewhere.
Mike

mapj1 said:

Use them instead. 

BS 7671 refers to the telecomms installation standards .in the respects you state. Those standards refer back to BS 7671 for matters of electrical safety (and provide context on that also). The installation standards work together in this respect.

Some of the product standards, however, are not aligned with the basic safety publication IEC 61140.

So it sounds like your options are either to run it in earthed steel conduit/trunking/ducting or just ignore BS 7671's specific requirements in favour of a 'common sense' approach (a precedent already set by the caravanning fraternity who almost universally ignore BS 7671's demand to use rubber flex for hook-ups, as it's only available in matt black, instead preferring brightly coloured PVC that can be seen by those mowing the grass).

   - Andy.

AJJewsbury said:

BS 7671's demand to use rubber flex

Wording in 721.55.2.6 for caravan hookup is different (includes 'or equivalent') to that in 717.52 for mobile & transportable units, which really does 'demand' H07RN-F.

I guess the 'or equivalent' could be debated until the cows come home - for my money, H05RN-F and H07RN-F are 'general purpose cables' in BS EN 50525-2-21, similarly H05VV-F to BS EN 50525-2-11 are 'ordinary duty cables"  - and in both standards the cables are intended for connection of appliances to a supply by a flexible connection - the difference being that of insulation and sheathing material. (In this case, not water resistance, as that would require H07RN8-F.)

I guess the 'or equivalent' could be debated until the cows come home

Indeed - I guess my thinking has been influenced by the fact it used to say "H07RN-F or H05VV-F" (back in the 16rh Ed days) - and someone somewhere decided that H05VV-F should be deleted. If the committee really did think they were equivalent when why go to the bother of deleting it? Besides I don't think I'd feel comfortable standing in the dock vs someone like our JP trying to claim that a 500V rated cables was equivalent to a 700V rated one. One that didn't quite have a harmonised  approval but was physically identical to a H07RN-F I'd be happier with.

   - Andy.

We seem to have got to a very silly conversation here. The problem is the foolish "feature creep" of supposedly power electrical specifications, and the way in which every standard in existence tries to reference every other standard. The result is a total mess, leading to questions like this. What has Ethernet got to do with power electrical installations? Answer: Nothing whatever. POE does not change this one jot, it provides battery level power, which of course is completely isolated from any mains supply. Why the reference above to RCDs? Ethernet (base T) was designed to be intrinsically safe, connecting two points with cable isolation at both ends. You can cut and hold onto the cables until the cows come home with zero risk. POE slightly changes this as multiple cables may have core centre taps that are common, but still they present no shock risk. None of this is thus the slightest bit relevant to BS7671, although it seems to have become included. Very odd.