Michael Fox said:

It would make sense to me to have them on the TT earth to minimise any differences in earth potential. Further, a future planned extension to the garden wiring (automated gates) would put both wiring systems in close proximity.  Would this make sense?

Regulation 411.3.1.1 requires simultaneously-accessible exposed-conductive-parts to be connected to the SAME earthing system, so it's possible the future planned extension might not conform to BS 7671 unless everything is on the same system.

Michael Fox said:

Before I finalise the installation I am considering converting the lighting system to be fed from the new TT earthed building.  This would appear to offer a safer solution as at the moment you can stand in the garden and touch on the lamp posts.

1. See also 411.3.1.1.

2. There is no requirement for anything you've described in your e-mail to be on 'TT', separated from PME, including, most of the time, the outbuilding. However, existing installations were sometimes done like that in the past. You need to consider Regulation 411.3.1.1 as above for anything fed from the house, but also note that 'going TT' doesn't necessarily mean there isn't a potential difference (from diverted neutral currents causing ground potential to rise) between an exposed-conductive-part of the TT system, and your feet, unless you separate from buried metalwork underground by 6-10 m (20-33').

note that 'going TT' doesn't necessarily mean there isn't a potential difference

Agreed - and not just from malign PME influence. While PME has its problems, TT isn't perfect either. The relatively high impedance of the Earth on a TT system means that normal earth leakage currents (a.k.a. protective conductor currents) can raise the potential on a TT earthing system and outdoor systems are notorious for getting damp and leaking more than intended. BS 7671 allows voltages of up to 50V before anything has to trip - far higher than what's likely to persist in a PME system in normal conditions (short of a broken PEN event for example). E.g. five TT circuits leaking say 20mA each and a 200Ω electrode might yield 20V touch voltage - which can persist indefinitely. Then there's the issue of the reliability of RCDs - if the RCD is faulty and an earth fault does occur in the installation, 230V can remain on the TT earthing system permanently (until someone noices...). Some studies suggest that around 7% of RCDs in service wouldn't perform as required.

Don't get me wrong, TT has its place (and I'm no fan of PME either), but it's definitely a case of swings and roundabouts and you might want to consider some mitigating measures (e.g. two tiers of RCDs for a TT system). Hence why ordinary steel public street lighting columns are almost always on PME.

   - Andy.

AJJewsbury said:

BS 7671 allows voltages of up to 50V before anything has to trip

I don't think that's true, at least if there's no 'fault' ... also, there is no "voltage operated" circuit-breaker any more, except for the device described in 722.411.4.1 (iii), which is on PME?

Regardless, I think mixing earthing systems in an installation can be more problematic where ADS is used, than using a TN-C-S or TT system, one or two specific cases of "special locations" excepted.

I don't think that's true, at least if there's no 'fault'

I was thinking of 411.5.3 (ii) - as residual or protective conductor currents < I_Δn won't necessarily cause the RCD to trip (whereas faults of negligible impedance always should) and these currents will naturally cause a p.d. across R_A - (ii) puts a limit on that.

  - Andy.

AJJewsbury said:

I was thinking of 411.5.3 (ii) - as residual or protective conductor currents < I_Δn won't necessarily cause the RCD to trip (whereas faults of negligible impedance always should) and these currents will naturally cause a p.d. across R_A - (ii) puts a limit on that.

Yes, that would be the case for a residual current fault, but currents flow in protective conductors and earthing systems for other reasons (for example, they could carry diverted neutral currents from shared extraneous-conductive-parts).

Theoretically, it would be possible to use a 115-0-115 V centre-tapped transformer with earthed mid-point to supply an installation that is TT, and in that case, 'leakage currents' could be any value and not operate an RCD as they would then be "common mode" - only a line to PE (or exposed-conductive-part or earth) would be detected by the RCD. RCDs would have to disconnect all live conductors, and in such a system there would be two line conductors, but no distributed neutral.

Theoretically, it would be possible to use a 115-0-115 V centre-tapped transformer

Indeed - or rather more simply, just a 3-phase system. Perhaps I shouldn't have presumed a single phase TT system and specified unbalanced leakage (or protective conductor) currents, as it's really only the current flowing to Earth via a significant impedance (R_A or similar) that's the issue - currents that cancel out within the installation don't count.

   - Andy.

Theoretically, it would be possible to use a 115-0-115 V centre-tapped transformer

Indeed - or rather more simply, just a 3-phase system. Perhaps I shouldn't have presumed a single phase TT system and specified unbalanced leakage (or protective conductor) currents, as it's really only the current flowing to Earth via a significant impedance (R_A or similar) that's the issue - currents that cancel out within the installation don't count.

   - Andy.

AJJewsbury said:

Indeed - or rather more simply, just a 3-phase system.

Only a perfectly balanced three-phase system, with equal equipment on each phase at all times - if you have single-phase circuits in a three-phase system, the 'leakage' will not be balanced.

HOWEVER, this concept does have its uses. I have in the past specified a three-phase arrangement to supply 3 sets of equal single-phase AC to DC power supplies that had high values of protective conductor currents, so that 3-phase 4-pole 30 mA RCDs could be used for safety ... the protective conductor currents effectively 'cancelled' (or at least reduced to a net residual current smaller than that required to operate the RCD).

Only a perfectly balanced three-phase system

There can be a degree of cancelling on any polyphase system - it's not all or nothing. Even on a 115-0-115 system it unlikely that leakage currents from both lines will be perfectly equal - differences in manufacturing tolerances of the capacitors in filters, or if the leakage is caused by damp, unequal resistances to start with. Even so cancelling can be useful - 30mA from one line and 40mA from another and both the RCD and the voltage across Ra would only see 10mA.

The effect on addition protection might be more dubious though - adding 50mA body current to a system that was already leaking 25mA  but 'out of phase' might not be as detectable as we might want.

   - Andy.

AJJewsbury said:

might not be as detectable as we might want

But as you say, we accept that risk (if it can be quantified) with three-phase systems, where leakage currents per phase can be in the Amperes range (and not all 'covered off' by an RCD).