Isolation Transformer - A discussion

[SUNRAY]

I need to clarify that I have a great deal of respect for all recent cotributers and your regular comments, I don't necessarily always agree but that's the nature of the electrical beast. In this instance I'm trying to get to the bottom of the differing opinions rather than just accept them.

If I were to use a generator to run a domestic heater in a building with a mains supply, I'd do an in the head, risk assesment to decide what action needed to be taken.

Perfect examples occurred last winter:
A shed containing tropical fish and the warm air heating from the main building failed, however the mains supply to the shed was only 16A and the various heaters, pumps and lights accounted for about 10A and water temp was dropping. The owner plugged in a fan heater and fairly quickly tripped the supply MCB.

My first and immediate fix was a 5.5KVA generator and 2 fan heaters to warm the air quickly.
No I didn't bond the supplies, both heaters were class I albeit plastic bodied and there was plenty of space between the different supplies, in the order of 4m or so.

However also last year the same generator was used to power fridges and freezers during a brown out, the 5 stood side by side in a row and within easy touching distance are ovens and other appliances, all class I and stainless steel in a commercial setting. Straight away I added the earth/bond no question. Even if it had only been one of those appliances, in the situation my decision and that of my boss would have been the same.

If I were putting in the panel heater (on a genset) into a domestic situation, chances are I'd likey not bond, if the genset was not earthed and only for a short period of time.

I don't believe there is any distinction between generators, for size or number of phases, most generators have an earth point with a warning it must be earthed however the bigger versions frequently have links to be able to dictate the earthing arrangements and running them IT is often quite feasible but I know I'd ever be able to install one of those big heaters into any situation where 'sign off' is required without the CPC of the heater being earthed regardless of the supply status. And most certainly not in any licenced public entertainment situation where multiple inspection layers occur.

In the other thread I mentioned a situation with a battery powered public address system where the on site inspection insisted I use a RCD on a contained (ie with no route earth) inverter supplying several class II devices.
I regularly install self contained PA systems for sound stages and they also get inspected and signed off including checking the bond to the stage structure and that is a 24V battery powered system.

There are British Standards for the methods of earthing/bonding generators/battery supplies, I've been going back through some of my documentation but so far not found any references.

 

[Harry Bloomfield]

I need to clarify that I have a great deal of respect for all recent cotributers and your regular comments, I don't necessarily always agree but that's the nature of the electrical beast. In this instance I'm trying to get to the bottom of the differing opinions rather than just accept them.

If I were to use a generator to run a domestic heater in a building with a mains supply, I'd do an in the head, risk assesment to decide what action needed to be taken.

Which all really that you can do. There is no absolutely correct and absolutely safe method, both methods have their own risks if the wrong circumstances happen. All you can do is weigh up which is the more likely failure mode and make that a part of your decision.

 

[JohnW2]

I need to clarify that I have a great deal of respect for all recent cotributers and your regular comments, I don't necessarily always agree but that's the nature of the electrical beast. In this instance I'm trying to get to the bottom of the differing opinions rather than just accept them. .... If I were to use a generator to run a domestic heater in a building with a mains supply, I'd do an in the head, risk assessment to decide what action needed to be taken.

Fair enough. However, you still haven't answered my question about the big genny and big heater(s) whose photos you keep posting - if that big genny was feeding (just) the big heater(s), what would you do, and why?

However also last year the same generator was used to power fridges and freezers during a brown out, the 5 stood side by side in a row and within easy touching distance are ovens and other appliances, all class I and stainless steel in a commercial setting. Straight away I added the earth/bond no question. Even if it had only been one of those appliances, in the situation my decision and that of my boss would have been the same.

What "earthing/bonding" are you talking about? Assuming that the output of the generator was 'floating', are you saying that you would connect one side of the genny output to the exposed-c-ps of the fridges/freezers and/or that you would also connect them (including one side of the genny supply) to some sort of actual 'earth', or what? - and whatever you mean, what do you think you would be achieving (other than, quite probably, increasing the risk of electric shock in the more plausible of possible 'fault scenarios')?

Kind Regards, John

 

[SimonH2]

and say that, in a situation like this, if one earths either side (or a centre tap) of the secondary of an isolating transformer, then one might as well not have bothered with the isolating transformer in the first place - since, in this situation, it is the 'isolation from earth' that is the whole point in having a transformer.

"It depends". Part of the problem is that we don't generally have a "proper" earth, but one that's tied to the neutral in TNC systems.

This all seems so 'obvious' to me that I get a bit concerned when I see 'discussion' about it

We disagree about the "obviousness" of it.

As per the other thread, a lot of the problem is that, as you suggested, a lot of what are called "isolation" transformers are not - since indirectly the output is closely tied to the input, CT of the secondary being tied to the supply earth which is itself (usually) hard tied to the supply neutral. Plug that same transformer into an IT system and it is now an isolating transformer.

Festoons are commonly used to light a pedestrian route and additional loads added at strategic points, like a floodlight or 'EXIT' sign at an access point or obstruction. In this case it happens to be a classI product ...

In which case it is not permissible to connect it to the 2 core festoon cable due to the lack fo CPC. If you add a CPC (or make the festoon cable 3 core) then you can - but I don't think it makes a huge amount of difference to the discussion.

1... Using a Class I item when the supply is floating is, in itself, not safe in the first place, even if its exposed-c-ps are earthed. The whole point of having earthed exposed-c-ps in a Class I item is so that a protective device will operate if a 'live' (L not N) conductor comes into contact with the exposed-c-p.

No. The SOLE point of earthing exposed conductive parts is to prevent those conductive parts having a dangerously high touch voltage relative to the local "earth" that a person is stood on. Tripping of upstream OCP is a by-product of having a supply referenced to earth - and mostly we have this arrangement so that a single fault will trip the OCP and cause disconnection of the supply - and thus trigger investigation as to why "it's not working".

But it's equally valid to simply bond everything remotely conductive into an equipotential zone - c.f. the people who hang out of helicopters and work on live lines while the helicopter is connected to one of the phases. The touch voltage has been eliminated by bonding, but nothing is tripped.

As a system gets larger, the probability of their being a fault somewhere in the system, and/or an accumulation of leakage currents to hazardous levels starts to approach 1. So the safety advantages of floating systems are lost and it is better to have a system with a well-defined relationship to earth.

Another alternative is to have a supply that is only weakly tied to earth, then implement fault monitoring by measuring the voltages between the supply conductors and earth. This is arguably the safest system for a small scale supply and I belive is used in some medical situations, but again it doesn't scale up well.

This is common in marine environments - where it's definitely not "small scale". But then you have a system where a few grand spent on earth fault monitoring (EFM) equipment is "back of the sofa" loose change, and the system is operated by trained operatives. I'm involved in just such a system in my day job - equipment is most definitely class 1 and earthed to the hull for safety. It's interesting looking at the data from the test lab where it's possible to see the "virtual neutral"* move away from earth as the EFM kit is charging up the capacitance in the system. In the event of a fault, everything carries on running, and the EFM will not only alert that there is a fault, but also localise it down to major circuit level (basically wherever the monitoring kit has been installed down to) - below that there's handheld kit that can be used to localise it further.

* In the lab, the "virtual neutral" is mathematically derived by the measuring equipment from the 3 phase voltages. It's what you'd get at the star point if you put 3 resistors onto the 3 phases.

That's not materially different from what I wrote, is it? In the context we are discussing (a fully-floating supply), having earthed exposed-c-ps will not facilitate the operation of any protective device (OPD or RCD). As for "limiting touch voltage" (which I would have thought was technically bonding, rather than earthing), the only touch voltage it will limit will be that between the exposed-c-p and something else earthed and, again, in the context of a fully-floating supply a hazardous touch voltage cannot really arise (since any pd between the exposed-c-p and earth will be via stray capacitance etc., hence too high impedance a path to be hazardous).

That's a wild assumption. In the case of a string of lights, perhaps valid. By the time you get to systems with inverter drives sized in 100s of kW, then the input filters on those (not to mention the accumulated capacitance across a number of them) are definitely not "too high impedance a path to be hazardous". On the system I'm working on at the moment, there's a consideration of whether the EFM can locate a fault quickly enough before filter capacitors overheat.

The point I was making to Sunray was that if an item is designated as Class I, the implication is surely that the "basic insulation" (plus "additional insulation" if present) is not considered adequate to provide the required degree of protection against electric shock (i.e. does not qualify as "Class II") . If that is the case, and if (as above) earthing the exposed-c-p does not really afford any additional protection against electric shock, then it surely follows that Class I items cannot be 'safely' used when the supply is floating, doesn't it?

You are drawing a correct conclusion if the starting assumption was valid. But the assumption that "if (as above) earthing the exposed-c-p does not really afford any additional protection against electric shock" is not valid in the general case.

Since, with a floating supply, earthing the exposed-c-p does not provide any additional protection, if a Class I item with an earth-referenced supply is not regarded as 'safe' unless the exposed-c-p is earthed, then nor can it be 'safe' with a floating supply, even if the exposed-c-ps are earthed.

Again, incorrect conclusion due to incorrect starting assumption.

Yes, but surely, for the isolated supply there is no earth.
That is, no path back to the supply either through the wiring or the planet.

No, earth is still there. It's just that the supply is not referenced to it. As above, there are plenty of situations where an "unearthed" conductive part could become dangerously live relative to the "earth" that a person is standing on. By earthing such parts, which really means bonding them all together and to the "earth" as experienced by the person, we prevent dangerous voltages appearing - that is the purpose of the CPC.
As above, it's really hard to get a truly isolated and floating supply. If you consider such a hypothetical thing then you'd be correct, but in real life it's hard to do.

Isn't that the point and purpose of isolated supplies?

No
The point of a truly floating/isolated supply is that there IS still an earth (as in the large conductive ball we inhabit), but you can't get a circuit between it and the supply. If you mean it doesn't have a CPC connected to earth - that depends : e.g. a shaver socket doesn't have one on the output, but a site transformer does.


But back to the original question ...
The safest option would be a fully isolated transformer, secondary not earthed, and an EFM in place - and operated by trained personnel on a 24/7 basis.

As for the thread that kicked this off, the same would apply to site supplies. But the chances of even the majority of site supplies being properly setup, by properly trained people, with a fully functional EFM, is ... rather small. So the safest option is to assume that there's going to be problems, that users won't care about bare conductors draped on the scaffolding, etc, etc, etc - and restrict the possible touch voltage to something reasonably safe given the conflicting requirements for a higher voltage to reduce operating currents and a lower voltage for safety.


Incidentally, I have two such strings of lights in the garden. The lights are about the right brightness when run from around 55-60V. One string runs from an auto-transformer (N&E shared with supply, L selected from transformer taps) I knocked up, the other runs from a "yellow box" my late father put together and I've never investigated how it's wired internally.

 

[JohnW2]

As per the other thread, a lot of the problem is that, as you suggested, a lot of what are called "isolation" transformers are not ....

True - but I have been talking throughout about (and only about) isolating transformers which really are 'isolation transformers' and are being used for 'isolation' - which, in practice, essentially means isolating the secondary from earth.

The discussion got rather confused when people started talking about 'yellow bricks' - which, for a start, involve a voltage change as well as anything else and, secondly, usually have the centre tap of their secondary connected to earth. That does not correspond in any way to what I regard as 'isolation', even if (as has been illustrated) some of the products have "isolation transformer" written on their cases! What exactly do they think is being 'isolated' from what?

No. The SOLE point of earthing exposed conductive parts is to prevent those conductive parts having a dangerously high touch voltage relative to the local "earth" that a person is stood on.

I don't really agree. In the event of a fault between (earth-referenced) L and an 'earthed' (to installation's MET) exposed-c-p fault, the 'touch voltage' ("to the local "earth" that a person is stood on") would be close to the full supply potential in a TT installation, and could be up to about half of the supply voltage in the case of a TN installation - and, in the absence of any upstream protective devices, that would persist - at least, until some cable melted (in the case of TN installation).

But it's equally valid to simply bond everything remotely conductive into an equipotential zone ...

Indeed - not only 'valid' but inherently safer. However, that "everything remotely conductive" would have to include your "the local "earth" that a person is stood on", which clearly can present problems!

That's a wild assumption. In the case of a string of lights, perhaps valid. By the time you get to systems with inverter drives sized in 100s of kW, then the input filters on those (not to mention the accumulated capacitance across a number of them) are definitely not "too high impedance a path to be hazardous".

You have moved the goalposts a million miles from anything which is (was!) being discussed or which is remotely relevant to a DIY forum or any sort vof domestic/consumer products.

You are drawing a correct conclusion if the starting assumption was valid. But the assumption that "if (as above) earthing the exposed-c-p does not really afford any additional protection against electric shock" is not valid in the general case.

Again, I was talking specifically about single Class I items, essentially those likely to be encountered in domestic environments. If the item is only deemed safe if the exposed-c-ps are 'earthed', how could it be considered safe if the supply were not earth-referenced, since the fact that it was 'earthed' would be of no interest to such a supply?

No ... The point of a truly floating/isolated supply is that there IS still an earth (as in the large conductive ball we inhabit), but you can't get a circuit between it and the supply. If you mean it doesn't have a CPC connected to earth - that depends : e.g. a shaver socket doesn't have one on the output, but a site transformer does.

That's why, to my mind, a shaver socket does provide 'isolation' whereas (no matter what is written on the case!) a site transformer does not. As I asked above, if you (and the people who write those labels) believe that the yellow brick is an 'isolation transformer', exactly what do you believe is being 'isolated' from what?

Kind Regards, John

 

[SimonH2]

The discussion got rather confused when people started talking about 'yellow bricks' - which, for a start, involve a voltage change as well as anything else and, secondly, usually have the centre tap of their secondary connected to earth. That does not correspond in any way to what I regard as 'isolation', even if (as has been illustrated) some of the products have "isolation transformer" written on their cases! What exactly do they think is being 'isolated' from what?

Being pedantic, I think you should be able to do an IR test and find that the output is isolated from the input. It's arguable that it's the supply being earthed that defeats that isolation Plug it into a floating supply (e.g. a generator without a N-E link) and I think you'll find that the input and output really are isolated.

I don't really agree. In the event of a fault between (earth-referenced) L and an 'earthed' (to installation's MET) exposed-c-p fault, the 'touch voltage' ("to the local "earth" that a person is stood on") would be close to the full supply potential in a TT installation, and could be up to about half of the supply voltage in the case of a TN installation - and, in the absence of any upstream protective devices, that would persist - at least, until some cable melted (in the case of TN installation).

Err, are we discussing the same thing ?
If you've a floating supply, which is what I thought we were talking about, then a L-E fault within that Class I item would simply pull the "L" to earth - and then nothing would happen. The N is now "live" and the L is no similar to what we are used to with the neutral. How does the touch voltage be anything significant ? How does any current flow ?

Hmm, checks back and yes :

Using a Class I item when the supply is floating is, in itself, not safe in the first place, even if its exposed-c-ps are earthed

So apply a short to earth of any single point in the system and nothing will happen. There will be no dangerous touch voltages, no dangerous current will flow, no cables will melt.
Only in the event of a SECOND fault will problems arise.
What would be a problem is if the two faults are in equipment with vastly differing ratings. Then the common fault current may overload the smaller circuit IF that circuit isn't double pole protected and the N-E fault is in the smaller circuit - so the L-E fault current of the larger circuit flows through the N & E of the smaller circuit and it's possible that the OCP of the larger circuit doesn't trip (in a timely fashion, or at all). Isn't that why some types of installations are required to have double pole protective devices ?

As I mentioned, in the marine environment it's normal to have floating systems and earth fault monitoring (to detect and locate that first fault) - and a LOT of Class I equipment. It's my understanding that the ability to "have a single fault and everything keeps working" is a significant factor in that.

 

[JohnW2]

Being pedantic, I think you should be able to do an IR test and find that the output is isolated from the input. It's arguable that it's the supply being earthed that defeats that isolation Plug it into a floating supply (e.g. a generator without a N-E link) and I think you'll find that the input and output really are isolated.

I think you are agreeing with me In practice, the feed to an isolating transformer will almost always be from an earth-referenced supply (since it would surely be rare that one would consider it necessary/appropriate to have an isolating transformer between a floating supply and the load?) - so, as I've been saying, in practice it is the earthing of any part of the secondary (either end, centre-tap or whatever) which 'defeats' the isolation (rendering the 'isolating transformer' useless in most senses).

Err, are we discussing the same thing ?

It seems not ...

If you've a floating supply, which is what I thought we were talking about ....

I took your comment about the "SOLE" reason for earthing an exposed-c-was a general one, since I presumed that you would have understood traht my comments about it being to facilitate operation of a protective device related to earth-reference supplies (since, in the single-fault scenario (and many 2-fault ones} , it would not achieve that with a floating supply).

... then a L-E fault within that Class I item would simply pull the "L" to earth - and then nothing would happen. The N is now "live" and the L is no similar to what we are used to with the neutral. ...

All agreed - but, as above, I was not talking about a floating supply.

... Hmm, checks back and yes : ... So apply a short to earth of any single point in the system and nothing will happen. There will be no dangerous touch voltages, no dangerous current will flow, no cables will melt. Only in the event of a SECOND fault will problems arise.

I suspect that your 'checking back' did not involve reading of all I have written in this thread, since you seem not to have understood the point I was making ...

... my point was that if an item is classified as "Class I" ("exposed-c-ps MUST be earthed") the implication is that the insulation is neither 'double' nor adequately 'reinforced' to be deemed adequate to protect from electric shock by contact with live parts (i.e. "Class II") - that presumably relating to mechanical damage to the insulation 'exposing' live parts within ... and if the 'live parts within' become touchable, it's at least possible that both sides of a floating supply would become touchable.

If one accepts that (i.e. that the insulation alone is not deemed adequate {'robust enough'} to keep fingers away from live parts, such that an earthed exposed-c-p is required for additional 'protection'), then it can't really be safe with a floating supply, since earthing the exposed-c-ps would offer no protection against that particular risk.

Kind Regards, John

 

[SimonH2]

I took your comment about the "SOLE" reason for earthing an exposed-c-was a general one, since I presumed that you would have understood traht my comments about it being to facilitate operation of a protective device related to earth-reference supplies (since, in the single-fault scenario (and many 2-fault ones} , it would not achieve that with a floating supply).

I stand by my comment. The SOLE purpose of earthing (more correctly, bonding as part of an equipotential zone) exposed cps is to prevent them becoming "live" and presenting a danger - regardless of supply type. It is NOT to trip any OCP - it's the other way round. The OCP tripping is a means of dealing with the fault current that results - and in some cases, that may not happen, the system could be such that it's impedance limited (c.f. the impedance limited, stall-able, motors used in central heating valve actuators).

... my point was that if an item is classified as "Class I" ("exposed-c-ps MUST be earthed") the implication is that the insulation is neither 'double' nor adequately 'reinforced' to be deemed adequate to protect from electric shock by contact with live parts (i.e. "Class II") - that presumably relating to mechanical damage to the insulation 'exposing' live parts within ... and if the 'live parts within' become touchable, it's at least possible that both sides of a floating supply would become touchable.

If one accepts that (i.e. that the insulation alone is not deemed adequate {'robust enough'} to keep fingers away from live parts, such that an earthed exposed-c-p is required for additional 'protection'), then it can't really be safe with a floating supply, since earthing the exposed-c-ps would offer no protection against that particular risk.

I disagree. Because I don't "accept that" ...
If two class I items develop faults such that their exposed cps are connected to different lines/phases, then the bonding will prevent dangerous touch voltages - that is all that's required. If you think differently, then perhaps you could do a diagram demonstrating the dangerous situation because I don't see it.
Naturally, there would then be a circuit between the two lines/phases via the CPCs - it is for the circuit design to deal with the fault currents that would flow, most likely by way of ADS (i.e. tripping of one or both OCPs).

But if we're talking about touching parts WITHIN a device - i.e. removing covers - then all discussions are mute because that's not normal operation and whoever takes the covers off is then responsible for not touching the live parts inside.

 

[Harry Bloomfield]

I stand by my comment. The SOLE purpose of earthing (more correctly, bonding as part of an equipotential zone) exposed cps is to prevent them becoming "live" and presenting a danger - regardless of supply type. It is NOT to trip any OCP - it's the other way round. The OCP tripping is a means of dealing with the fault current that results - and in some cases, that may not happen, the system could be such that it's impedance limited (c.f. the impedance limited, stall-able, motors used in central heating valve actuators).

.

Sorry John, but I agree with Simon, the protection of life is the absolute first priority and as an aside, I have made good use of impedance and current limiting to good effect several times when designing.

I disagree. Because I don't "accept that" ...
If two class I items develop faults such that their exposed cps are connected to different lines/phases, then the bonding will prevent dangerous touch voltages - that is all that's required. If you think differently, then perhaps you could do a diagram demonstrating the dangerous situation because I don't see it.
Naturally, there would then be a circuit between the two lines/phases via the CPCs - it is for the circuit design to deal with the fault currents that would flow, most likely by way of ADS (i.e. tripping of one or both OCPs).

But if we're talking about touching parts WITHIN a device - i.e. removing covers - then all discussions are mute because that's not normal operation and whoever takes the covers off is then responsible for not touching the live parts inside.

Exactly!

 

[EFLImpudence]

But we only 'bond' exposed-c-ps that are connected, by their respective CPCs, to the same 'earth' and even then only in bathrooms and the like.

If a single exposed-c-p of an isolated supply is not 'earthed' to the house supply MET then there is nothing to 'bond' it to.

 

[JohnW2]

I stand by my comment. The SOLE purpose of earthing (more correctly, bonding as part of an equipotential zone) exposed cps is to prevent them becoming "live" and presenting a danger - regardless of supply type. It is NOT to trip any OCP ...

I think you have yourself illustrated the way in which the two concepts are getting muddled. The purpose of bonding in to prevent dangerous PDs existing between two simultaneously touchable parts (whether parts of the electrical installation {exposed-c-ps} or otherwise), and/but the purpose of earthing is to facilitate the operation of a protective device if an exposed-c-p becomes 'live' (even if any other simultaneoudlsy-touchable parts are, incidentally or by virtue of bonding, equipotential with that live part).

The (required) connection of exposed-c-ps to a CPC is, and is generally thought of, and described as, earthing, not bonding. What you seem to be describing is Supplementary Bonding which,these days, is rarely required if exposed-c-ps are 'adequately earthed'.

I disagree. Because I don't "accept that" ...

I find it hard to see why you don't, so maybe I was not clear enough - so let me try again ...

... in the name of "protection against electric shock" a Class II item is required to have "double or reinforced insulation. If, as very often is the case, the item has no exposed-c-ps (i.e. only plastic is touchable), then the requirement is clearly nothing to do with 'touch voltages'. The only rational (and reasonable) reason for the requirement for "double or reinforced insulation) is as an adequately robust (or 'with redundancy') means of preventing someone coming into contact with live parts within the (plastic) enclosure, presumably in the event that 'inadequately robust' insulation may become damaged/break thereby allowing one to touch live parts. Clearly nothing to do with anything being 'equipotential' with the plastic-enclosed item.

But if we're talking about touching parts WITHIN a device - i.e. removing covers - then all discussions are mute because that's not normal operation and whoever takes the covers off is then responsible for not touching the live parts inside.

As above, that is, indeed, precisely what I was talking about in the part of the discussion you have picked up on - but not in terms of people "removing covers" but, rather, in the event that inadequately 'reinforced' single insulation became damaged. As above, what else could be the reason for the requirement for "double or reinforced insulation" in a plastic-enclosed item?

Kind Regards, John

 

[JohnW2]

Sorry John, but I agree with Simon, the protection of life is the absolute first priority ....

Exactly, which takes me back to where I got sucked into this part of the discussion (in the original thread) ...

... I didn't understand, and still don't understand, why (other than perhaps in very unusual circumstances) anyone would want to increase the risk to life and limb by unnecessarily connecting to earth any part of the secondary of a so-called 'isolating transformer', thereby converting the 'isolated' ('floating', not earth-referenced) supply into an earth-referenced supply.

By all means bond together all simultaneously-touchable conductive parts since, as Simon has said, it is that bonding that will considerably reduce the risk of significant/serious electric shock. However, the moment one connects any part of the otherwise-floating supply to earth, one greatly increases the risk to life and limb in some scenarios, and I am far from convinced that there are any 'balancing advantages' of earth-referencing an otherwise-floating supply, other than in the most unusual/contrived of scenarios.

I therefore wondered (and still do) why others appear (by virtue of the practices they advocate) to be less concerned about the "protection of life" than I am.

Kind Regards, John

 

[SUNRAY]

I really started doubting my RAMS and deliberately stayed away from this thread to see which direction is drifted into. My strong belief is the external metal being earthed/bonded to other touchable metal (within reason of course) is purely and simply to prevent an electric shock hazard.
Now if that 'prevention of shock hazard arrangement' happens to offer a secondary benefit such as automatic disconnexion (incidently something which was not available for a very long time after we started earthing the metalwork) then all I can say to that is bring it on.

I fully understand all of the points being offered here, (yes I really do even if I argue against them) but trying as hard as I can there is no way I can come down on the side of JohnW or EFL in what I think I interpret as short sighted.
Sadly I live and work in the real world with real faults and real botches. I have received more shocks than I've ever wanted from faulty/incorrectly installed equipment.

Just as sadly I have worked in situations where floating installations without earths exist and it's a PITA constantly checking everthing, and I truly do mean EVERTHING with the neon and LED screwdrivers, yes we do carry both!! and yes I'm referring to those very pieces of test eqipment that we regularly see mentioned on here as useless and unreliable!! but invaluable in preventing shocking experiences when the simple task of making metalwork safe is deliberatly neglected.

Thank you Simon for bringing my head back down to earth - pun very much not intended but as it happens very appropriate.

 

[JohnW2]

I really started doubting my RAMS and deliberately stayed away from this thread to see which direction is drifted into. My strong belief is the external metal being earthed/bonded to other touchable metal (within reason of course) is purely and simply to prevent an electric shock hazard.

As I recently wrote, they are two different things, but I think you are probably perpetuating the confusion by writing "earthed/bonded", since they are two different things.

As I recently wrote, bonding (connecting together simultaneously-touchable conductive parts) is definitely a good thing that will consdierably reduce the risk of electric shock. That remains the case whether or not the bonding is also connected to earth (which it often will be, since at least one of the bonded items will often have a path to earth).

As I also wrote, earthing of an exposed-c-p exists to facilitate operation of a protective device in the event that the exposed-c-p - but that will only work with an earth-referenced supply (i.e. will NOT work if the supply is floating). In other words, earthing (as opposed to bonding) doesn't really achieve anything useful (but can increase some other hazards) if the supply is floating.

... Now if that 'prevention of shock hazard arrangement' happens to offer a secondary benefit such as automatic disconnexion ... then all I can say to that is bring it on.

As above, one can get the benefits of bonding (in terms of reducing risk of shock) by bonding simultaneously parts together, even if those parts are not also connected to earth.

Connecting one side of the 'isolated' supply to earth is a totally different matter. It converts the isolated/floating supply into an earth-referenced one, which means that anyone touching the non-earthed side of the supply and simultaneously touching anything 'earthed' would/could get a serious shock - something that would not happen if the supply had been left 'floating'.

I have received more shocks than I've ever wanted from faulty/incorrectly installed equipment.

With respect, if you utilised sufficiently safe practices, you would not receive shocks no matter how faulty or incorrectly installed the equipment was. Your practices should always 'assume (and anticipate) the worst, and the 'extremely unlikely''.

Kind Regards, John

 

[SUNRAY]

As I recently wrote, they are two different things, but I think you are probably perpetuating the confusion by writing "earthed/bonded", since they are two different things.
Kind Regards, John

And I know that. The reason I write earth/bond is either may be required depending on the situation.

As I also wrote, earthing of an exposed-c-p exists to facilitate operation of a protective device in the event that the exposed-c-p -
Kind Regards, John

With respect you have written

I think you need to read more carefully what I have written, several times. What I have said is:
The whole point of having earthed exposed-c-ps in a Class I item is so that a protective device will operate if a 'live' (L not N) conductor comes into contact with the exposed-c-p.
Kind Regards, John

And as you say you have written that several times and now you appear to be changing your mind.

With respect, if you utilised sufficiently safe practices, you would not receive shocks no matter how faulty or incorrectly installed the equipment was. Your practices should always 'assume (and anticipate) the worst, and the 'extremely unlikely''.

Kind Regards, John

Yes of course that is the case, but After you have turned the main switch and every MCB in a CU off, would it occur to you to measure the voltage between 2 adjacent metal boxes linked with plastic conduit... ? No neither did I.
Or After you have turned the main switch and every MCB in a CU off, would you expect to find 200V between the back box and the socket when removed... ? No neither would I. And after you have turned the main switch and every MCB in a CU off, would you expect to be able to light a 40W bulb with that 200V? No neither did I.

It may not be obvious to you but I do try to work in a safe manner and deliberately work in a way that I expect to not get electric shocks.