You might equally well ask whether it is safe to supply a SP circuit from a TP DB; or what about for example, a lathe with a 3 phase motor which is not running, but a SP lamp and DRO which are.

I don't see a problem.  Graham's caveat concerning RCD protection is entirely reasonable, but even then, taking one phase off a 3 phase plug isn't any different from using a BS 1363 plug in a circuit which (due to age) has no RCD protection.

Perhaps the OPs fear is the commonly held mis-belief that taking a single phase load through a three phase RCD will lead to it tripping....

Chris Pearson said:

taking one phase off a 3 phase plug isn't any different from using a BS 1363 plug in a circuit which (due to age) has no RCD protection.

Well, yes ... and no.

The issue is two-fold simply because it's a place of work:

(a) risk assessments need to be reviewed when standard change ... and this has been changed for a very long time, so the duty-holder should be involved in the discussion.

(b) More importantly, anyone installing, manufacturing or supplying equipment for the workplace will have to comply with Section 6 of the H&S@W etc. Act, and quite possibly the CDM Regulations, all of which put the onus on the installer (or designer if part of the electrical installation) to ensure safety ... i.e. that likely makes consideration of the lack of RCD your (the installer, or person fitting the plug) problem. To quote Section 6(3) of the H&S@W etc. Act 1974 (as amended):

(3) It shall be the duty of any person who erects or installs any article for use at work in any premises where that article is to be used by persons at work or who erects or installs any article of fairground equipment to ensure, so far as is reasonably practicable, that nothing about the way in which the article is erected or installed makes it unsafe or a risk to health at any such time as is mentioned in paragraph (a) of subsection (1) or, as the case may be, in paragraph (a) of subsection (1) or (1A) above.

Agree those are the rule,  but it is hard to see how removing the ability to unplug something and making no other changes makes it any safer to maintain. There are many bits of kit where a 30mA RCD would be up and down like the proverbial. This is where the documented exception and RA come into play - and for more or less fixed kit. There are things one can do (double earthing is one) to remove the single fault to danger that  is the ' what if the CPC comes detached in the plug' that is the big worry that makes high leakage kit without an RCD dangerous. It rather depends how the plug and socket are used - if the lead is trailed across a  car park and pulled in and out every day, it is far more of a worry than a tidily mounted plug socket pair on the wall at waist height obviously isolating  an adjacent  milling machine or something that is unplugged only when the man is called out and the machine is being repaired.

Mike.

I can see that a risk assessment for inaccessible sockets or those in restricted areas might permit the absence of RCD protection, but not where an ordinary person might plug in a piece of office equipment, a portable power tool, or a vacuum cleaner, etc.

So how long do we allow for a building to be upgraded?

Now I am going to struggle not to take my tester in to work (in a government building). It might be interesting to see the response to a tripped RCD.

mapj1 said:

Agree those are the rule,  but it is hard to see how removing the ability to unplug something and making no other changes makes it any safer to maintain

I'm not saying they do ... the manufacturer may well have a recommendation for an RCD in any case.

mapj1 said:

here are many bits of kit where a 30mA RCD would be up and down like the proverbial. This is where the documented exception and RA come into play - and for more or less fixed kit.

Agreed, and there is guidance in GN5 (9th Ed 2022) to help with what other measures ought to be in place if the RCD is not fitted ... one issue, of course, is protective conductor currents ... and I know of many machines that DO have high protective conductor currents, that have NOT been installed in accordance with manufacturers instructions and BS 7671, despite it being spelled out in big bold letters.

mapj1 said:

It rather depends how the plug and socket are used - if the lead is trailed across a  car park and pulled in and out every day, it is far more of a worry than a tidily mounted plug socket pair on the wall at waist height obviously isolating  an adjacent  milling machine or something that is unplugged only when the man is called out and the machine is being repaired.

I disagree if the person is touching the machine when it's in use - we're not necessarily just talking about flexible lead failures that additional protection according to BS 7671 might afford. There may well be flexible connections to parts of the machine too.

The requirement in H&S@Wetc Act is "SFARP", and it's generally really difficult to argue that it wasn't reasonable to fit an RCD (or add in some of the other precautions in lieu, per GN5, when the risk assessment is undertaken).

mapj1 said:

Agree those are the rule,  but it is hard to see how removing the ability to unplug something and making no other changes makes it any safer to maintain

I'm not saying they do ... the manufacturer may well have a recommendation for an RCD in any case.

mapj1 said:

here are many bits of kit where a 30mA RCD would be up and down like the proverbial. This is where the documented exception and RA come into play - and for more or less fixed kit.

Agreed, and there is guidance in GN5 (9th Ed 2022) to help with what other measures ought to be in place if the RCD is not fitted ... one issue, of course, is protective conductor currents ... and I know of many machines that DO have high protective conductor currents, that have NOT been installed in accordance with manufacturers instructions and BS 7671, despite it being spelled out in big bold letters.

mapj1 said:

It rather depends how the plug and socket are used - if the lead is trailed across a  car park and pulled in and out every day, it is far more of a worry than a tidily mounted plug socket pair on the wall at waist height obviously isolating  an adjacent  milling machine or something that is unplugged only when the man is called out and the machine is being repaired.

I disagree if the person is touching the machine when it's in use - we're not necessarily just talking about flexible lead failures that additional protection according to BS 7671 might afford. There may well be flexible connections to parts of the machine too.

The requirement in H&S@Wetc Act is "SFARP", and it's generally really difficult to argue that it wasn't reasonable to fit an RCD (or add in some of the other precautions in lieu, per GN5, when the risk assessment is undertaken).