Isolation Transformer - A discussion

[AdrianUK]

Following on from a recent discussion that moved slightly off topic and into the pros & cons of isolation transformers, I thought I would start a separate learned discussion

So, the scenario :

A temporary outside decoration is to consist of a 20m length of festoon cabling with 40 coloured LED BC lamps. This a 230V powered festoon intended to be powered from the mains. This festoon string has only two cores in its mains lead - it does not have an earth core & has no provision for any kind of earthing.

Lets consider the possible supply arrangements:

i). Plug directly into the nearest available 13A socket (which may or may not have RCD protection).

ii). Plug into the nearest 13A socket via a plug-in type 30mA RCD (which may in addition to any permanently installed RCD).

iii). Power from an 'isolation transformer'. Here is were the discussion really starts. This transformer may have a number of possible configurations:

a). Full floating - the secondary has no connection to earth.
b). Connected to the mains earth - one leg of the secondary is connected to the installation earth (which may be PME or TT).
c). Connected to an independent earth rod - one leg of the secondary is connected to a separately driven copper earth rod (lets assume, for the purpose of discussion, it is well clear of any other buried metalwork which maybe connected to a PME system).

Any of the above three transformer configurations may be used with a plug-in type RCD on the secondary side (or primary side.....)

Please note: this is a topic for learned discussion, I'm not asking for advice on how to do this, I'm starting a discussion

Over to you, Gentlemen!

 

[Harry Bloomfield]

iii + a is the safer configuration, assuming the isolation transformer is a good quality one, fully up to spec. and undamaged. The only risk is that of making contact with both poles of the transformer and there is no way of protecting against that. It is also the standard method of powering an electrical workbench supply, especially TV repair benches.

b and C offer no safety advantage over a, but adds a risk.

 

[bernardgreen]

For what it is worth :-



The festoons with coloured lamps along the houses are 1 Watt LED bayonet cap lamps and directly mains fed via RCDs, no isolation transformer. All out of reach for a person unless they are using a ladder.

The lights on the tree are a mixture of festoons that can be touched by a small person on the ground, the lamps are not easily removed from the cable.
Four strings of 150 lamps at 24 Volt AC incandescent fed from 2 toroidal transformers
Two strings of 150 lamps rated at 12 Volt DC supplied from 2 switch mode power supplies (with a lower voltage to match the brightness of the incandescent lamps)
The mains supply to tree is via an RCD in the base of a lamp post.

This arrangement is considered to be safe and there have been no adverse incidents.

 

[JohnW2]

iii + a is the safer configuration, assuming the isolation transformer is a good quality one, fully up to spec. and undamaged.

I agree.

The only risk is that of making contact with both poles of the transformer and there is no way of protecting against that

Exactly - the very point I've been making in the other thread.

b and C offer no safety advantage over a, but adds a risk.

Agreed, but (as in the other thread) I would go further than that - 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.

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

Kind Regards, John

 

[AdrianUK]

Gentlemen,

Thank-you for your responses so far - none of which I disagree with.

Lets explore the scenario a little further...

So, if we run the system 'earth free' (no part of the transformer secondary is connected to earth), lets consider what would happen in the event of a 'fault'.

The string consists of 40 lamps, each of which relies on a simple seal around the neck of the lamp to keep out the wet. It the wet gets in, lets consider what happens.

If the water connects both prongs of the BC lampholder a current will flow. This almost certainly won't blow the fuse supplying the system (1kVA isolation transformer with a 13A BS1361 on the primary & a 3A BS1361 on the secondary) so it would probably just dry itself out.

What if the water only makes contact with one prong? No current will flow - the system isn't deliberately earthed, but will now be (impedance) 'earthed' by this first fault. The system has no means of detecting this so will continue to function.

What if the water now enters a second lampholder? We could have a current between the two lampholders. Again this will not be detected unless it exceeds a sufficient magnitude to blow a 3A fuse - unlikely. So this situation could persist for many hours.

Discuss

 

[bernardgreen]

If the water connects both prongs of the BC lampholder a current will flow.

This will initially create a small leakage current if the water is clean. However this current will mean some electrolytic action will occur between the various metal bits in contact with the water. This could result in the water plus the products of electrolysis becoming a low impedance short circuit Live to Neutral

What if the water only makes contact with one prong? No current will flow

True

What if the water now enters a second lampholder? We could have a current between the two lampholders.

Unless the cable sheath is conductive there will be no third wire for the leaking current to flow from holder to holder.

( some festoon cables seem have very high impedance conductive sheaths. This may be intended to enable faults to be detected by monitoring the voltage on the sheath )

 

[SUNRAY]

( some festoon cables seem have very high impedance conductive sheaths. This may be intended to enable faults to be detected by monitoring the voltage on the sheath )

I've found this with black festoons running at 230V in wet weather and accordingly stopped handling them live, it doesn't seem to happen with the yellow variety though.
I had always assumed it was down to price/quality of the seals.

 

[AdrianUK]

I've found this with black festoons running at 230V in wet weather and accordingly stopped handling them live, it doesn't seem to happen with the yellow variety though.
I had always assumed it was down to price/quality of the seals.

Are you thinking of the original festoon that was made by tightening the self piercing 'beantree' lampholders onto a two core flat rubber cable?

 

[SUNRAY]

Are you thinking of the original festoon that was made by tightening the self piercing 'beantree' lampholders onto a two core flat rubber cable?

No I haven't come across any of that for yonks it's all moulded stuff now.

 

[JohnW2]

Gentlemen, Thank-you for your responses so far - none of which I disagree with. Lets explore the scenario a little further... So, if we run the system 'earth free' (no part of the transformer secondary is connected to earth), lets consider what would happen in the event of a 'fault'.

OK.

The string consists of 40 lamps, each of which relies on a simple seal around the neck of the lamp to keep out the wet. It the wet gets in, lets consider what happens. .... If the water connects both prongs of the BC lampholder a current will flow. This almost certainly won't blow the fuse supplying the system (1kVA isolation transformer with a 13A BS1361 on the primary & a 3A BS1361 on the secondary) so it would probably just dry itself out.

Quite so. The impedance through the water probably would be quite high, so it's likely that relatively little current will flow through it and hence very little at all will 'happen'. The bulb might well even stay alight, maybe a little dimmed.

What if the water only makes contact with one prong? No current will flow - the system isn't deliberately earthed, but will now be (impedance) 'earthed' by this first fault. The system has no means of detecting this so will continue to function.

Well, for a start, that's a very unlikely scenario, given the very confined space for the water to occupy. In any event, water within the lamp holder/base which was in contact with 'one prong', and nothing else, would not 'earth' anything - the water which was in contact with one prong would have to somehow be in contact with something earthed (e.g. 'the earth') for that to happen.

What if the water now enters a second lampholder? We could have a current between the two lampholders. Again this will not be detected unless it exceeds a sufficient magnitude to blow a 3A fuse - unlikely. So this situation could persist for many hours.

If water were in contact with the second lampholder in the same fashion as the first ('very improbably' in contact with just one prong) then ... if it were the same prong in both, then the water in both would be at the same potential so, even if there were 'water continuity' between that water and the ('touchable') outside world, there would be no risk. If the water in the two lampholders were in contact with (only) two 'different prongs), then those two bits of water would have the full supply voltage between them - so, if there were somehow 'water continuity to the touchable outside world' in both cases then there would be a theoretical shock risk if one touched both bits of water. However, that would seem to be an incredibly improbable scenario, in the 'once in a blue moon, if ever' category!

It may be easier to forget about water ingress and think just about 'frayed cables'. If the insulation of one of the conductors became damaged such that the conductor could be touched then, if it were a fully earth-free system, then if one touched that exposed conductor whilst simultaneously touching anything else, one would not get a significant shock. However, if one side of the secondary were earthed, then there would be a 50% chance (i.e. if one was touching the 'non-earthed side of the supply) of one getting a serious shock if one touched the exposed conductor at the same tiome as being in contact with something earthed (like 'the earth').

If the insulation were damaged in two places, then ... if it were the same conductor in both cases, and one touched both simultaneously then the situation would be as above - i.e. if it were a fully earth-free system, then one would not get a significant shock. However, if one side of the secondary were earthed, then there would be a 50% chance (i.e. if one was touching the 'non-earthed side of the supply) of one getting a serious shock if one touched both exposed conductors at the same time as being in contact with something earthed (like 'the earth').

If the two exposed conductors were on different sides of the supply, then one would get a potentially serious shock regardless of whether or not the secondary was earthed - and nothing can be done to protect people against that (other than mechanical things to reduce the chance of such contact).

Does any of that help?

Kind Regards, John

 

[AdrianUK]

As I said at the beginning, this isn't a request-for-advice thread. Its more of a Sunday afternoon discussion thats now run over into Monday. I understand the pros & cons of running an earth free system. Most of the suspects I expected to join-in have, with the exception of Eric who I'm sure will be along in due course.....

Well, for a start, that's a very unlikely scenario, given the very confined space for the water to occupy. In any event, water within the lamp holder/base which was in contact with 'one prong', and nothing else, would not 'earth' anything - the water which was in contact with one prong would have to somehow be in contact with something earthed (e.g. 'the earth') for that to happen.

I'm thinking this festoon wouldn't be in free space, it would must likely hung on something wet & earthy, like a tree - so there would be an (admittedly high impedance) path to true earth via the tree. The system now becomes a impedance earthed system (IT). What happens with the second fault now ?

If the water in the two lampholders were in contact with (only) two 'different prongs), then those two bits of water would have the full supply voltage between them - so, if there were somehow 'water continuity to the touchable outside world' in both cases then there would be a theoretical shock risk if one touched both bits of water. However, that would seem to be an incredibly improbable scenario, in the 'once in a blue moon, if ever' category!

With the tree scenario above, the tree would provide the (semi-conductive) path between the two lampholders. A current, probably of tiny magnitude, would flow, between the two lampholders via the tree.

I'm not much thinking of a shock risk here, I think its clear to us all that earth free system vastly reduces that. I get that with such a system one would have to make simultaneous contact with both poles to receive a shock. I am however, thinking more about other faults and their consequences, how they might be detected, or even if it matters.

An just for a bit more food for thought..... lets introduce a class I colour changing LED floodlight, on a spike, into the system

 

[JohnW2]

I'm thinking this festoon wouldn't be in free space, it would must likely hung on something wet & earthy, like a tree - so there would be an (admittedly high impedance) path to true earth via the tree. The system now becomes a impedance earthed system (IT). What happens with the second fault now ?

I really do think that the probability of just one side of the supply (one 'prong' of the lampholder) developing a significant (significantly low) impedance path to earth (via a tree or whatever) would be close to zero. The two 'prongs' of a lampholder are only a few mm apart, and it hard to see how water would be in contact with one, and not the other, with the water also being 'in continuity' with the tree/whatever. If the water were in similar contact with both 'prongs' (and the tree/whatever) then, again, there would be no significant risk.

With the tree scenario above, the tree would provide the (semi-conductive) path between the two lampholders. A current, probably of tiny magnitude, would flow, between the two lampholders via the tree.

You're now requiring that, in two separate lampholders, water would be in contact with only one prong of the lampholder (and also with the tree) and also that it was a different 'prong' in the two lampholders. Given that, as above, I think that the probability of that happening in one lampholder would be 'close to zero', I'm really not sure what language I should use in relation to the probability of that happening in two separate lampholders, involving two different sides of the supply, and with there being a significant path between them through the tree - "even closer to zero", I suppose

I'm not much thinking of a shock risk here, I think its clear to us all that earth free system vastly reduces that. I get that with such a system one would have to make simultaneous contact with both poles to receive a shock.

Quite so.

I am however, thinking more about other faults and their consequences, how they might be detected, or even if it matters.

I can't really comment on that without knowing what "other faults and their consequences" you had in mind. None come immediatelt to my mind!

An just for a bit more food for thought..... lets introduce a class I colour changing LED floodlight, on a spike, into the system

Other than that the spike might offer a more likely 'through water' connection between one prong of the lampholder and earth than would a tree, I don't think that alters the generality of anything I've said.

Kind Regards, John

 

[SUNRAY]

An just for a bit more food for thought..... lets introduce a class I colour changing LED floodlight, on a spike, into the system

There are 2 points I'd consider:
1. is the spike is sufficient for an adequate earth, or is an additional earth spike required
2. is a suitable mains earth available.

Additional to that I'd be forced to consider additional connexions to it, such as DMX etc

 

[JohnW2]

There are 2 points I'd consider: 1. is the spike is sufficient for an adequate earth ...

The spikes of the 'floodlights on spikes' I have appear to be plastic In any event ....

... or is an additional earth spike required

I thought we were discussing the 'earth-free' situation, downstream of an isolating transformer with no earth connection to the secondary, weren't we?

3. is a suitable mains earth available.

See above. No sort of earth is 'suitable' for an earth-free circuit!

BTW, where is your point (2)?

Kind Regards, John

 

[SUNRAY]

There are 3 points I'd consider:
1. is the spike sufficient for an adequate earth,

The spikes of the 'floodlights on spikes' I have appear to be plastic In any event ....

Quite clearly then the answer to 1. is: NO

2. or is an additional earth spike required

I thought we were discussing the 'earth-free' situation, downstream of an isolating transformer with no earth connection to the secondary, weren't we?

We are

3. is a suitable mains earth available.

See above. No sort of earth is 'suitable' for an earth-free circuit!

I replied directly to Adrians post about a festoon on twin cable and the very common requirement to add an additional load partway along the system. 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, Are you saying the earth connexion on a classI product should be ignored? I will not accept 'excluding a classI product from a floating supply'.

BTW, where is your point (2)?

Kind Regards, John

Changed back to the original 3 as you quoted just as I wrote my 1st draft.