Cable size between equipotential earth bonding bar and distribution board in a Group 1 medical location

AJJewsbury said:

More the other way around, I'd suggest. Supplementary bonding provides no guarantees to reduce Zs, but rather it limits the available touch voltages within the location.

Yes ... and no.

There's no guarantee the touch-voltage will be limited to 50 V AC or 120 V DC (in general, Section 415, or in medical locations, 25 V AC or 60 V DC, Section 710). However, if the touch-voltage is not limited to those voltages, a protective device will operate.

The voltages are therefore only limited to the stated voltages for fault currents that won't operate the protective device in time ... so where, according to Section 419, disconnection times can't be met, and supplementary local equipotential bonding is applied, the touch-voltages are limited to 50 V AC or 120 V DC (in general) or 25 V AC or 60 V DC (in medical locations according to Section 710), because the prospective fault current has not reached I_a.

Anas SAHILI said:

This ensures that protective devices operate quickly enough to prevent hazardous touch voltages.

Well, again, not quite.

As described above, the real situation is a combination of the responses from Anas SAHILI and AJJewsbury (48c484574606166ec78373fa22d9df26): .

In short:

Supplementary local equipotential bonding limits the touch-voltage between simultaneously-accessible exposed-conductive-parts in faults where protective devices cannot operate in the specified time. In cases where protective devices operate in the specified time, the touch-voltage can be reduced, but unless R≤U_{touch-voltage-limit}/I_pf, the touch-voltage will not necessarily be limited to the relevant touch voltage limit U_{touch-voltage-limit} (25 V AC, 50 V AC, 60 V DC or 120 V DC depending on whether general rules or Section 710 applies) for the given prospective fault current I_pf.

gkenyon said:

In cases where protective devices operate in the specified time, the touch-voltage can be reduced, but unless R≤U_{touch-voltage-limit}/I_pf, the touch-voltage will not necessarily be limited to the relevant touch voltage limit U_{touch-voltage-limit} (25 V AC, 50 V AC, 60 V DC or 120 V DC depending on whether general rules or Section 710 applies) for the given prospective fault current I_pf.

In practice, calculating what the actual touch-voltage might be between any two simultaneously-accessible exposed-conductive-parts is very difficult because of parallel paths, which is why we need to stick to 'theoretical maxima' based on some current or other, either I_a or I_pf. Unless the CPCs connecting the two simultaneously-accessible exposed-conductive-parts are relied upon for the bonding, The actual touch-voltage is typically less than the theoretical maximum for the given current, because not all of that current flows in one conductor.

so where, according to Section 419, disconnection times can't be met, and supplementary local equipotential bonding is applied, the touch-voltages are limited

Although it us unusual for section 419 to apply - in common circumstances ADS for all the circuits in the location will work well within the times specified in section 411 (let alone 5s), but still supplementary bonding (in medical locations) is required. So I suspect it's as much a case of reducing touch voltages during ADS disconnection times as far as reasonably practical, rather than just covering cases where ADS doesn't occur (in time).

Ditto in bathrooms - supplementary bonding is still demanded where not all circuits have 30mA RCD protection - even if all circuits have disconnections times <0.4s by MCB or fuse.

   - Andy.

AJJewsbury said:

So I suspect it's as much a case of reducing touch voltages during ADS disconnection times as far as reasonably practical, rather than just covering cases where ADS doesn't occur (in time).

In general, yes, for socket-outlets and most equipment fed from the TN or TT supply where disconnection times cannot be met ... the medical IT system, slightly different perhaps as it's not designed to disconnect on first fault, but for those circuits, I_a would be selected for the protective device(s) that are intended to operate on a second fault.

AJJewsbury said:

Ditto in bathrooms - supplementary bonding is still demanded where not all circuits have 30mA RCD protection - even if all circuits have disconnections times <0.4s by MCB or fuse

Agreed, although that still permits up to 50 V AC touch-voltage, which, granted, has not been seen to be problematic in causing death, but certainly I think it's the current [excuse the pun] thinking that 50 V AC wouldn't necessarily protect someone who is submerged in a bath.

Sorry ... edited 3 Feb 2026 ... NOT AGREED after further reflection ... see response.

AJJewsbury said:

Ditto in bathrooms - supplementary bonding is still demanded where not all circuits have 30mA RCD protection - even if all circuits have disconnections times <0.4s by MCB or fuse.

 AJJewsbury 

Not agreed in full ... the disconnection time might be reduced, BUT the touch voltage is not necessarily limited to 50 V AC or 120 V DC by the bonding. The provision of the RCD can't limit the fault current, only affect the time that the person is subjected to the touch-voltage.

For the circuits where say a B32 provides ADS, the bonding resistance requirement is:

R ≤ (50 V)/( 160 A) Ω, or R ≤  0.31 Ω ... which could exceed Zs  for the circuit in question ... and is likely to be greater than R2. Certainly, it's not guaranteed to reduce the voltage to 50 V if operating "at the limit" of the formula in Reg 415.2.2.

Yes, in practice, the bonding resistance in a small bathroom might well be far, far lower than the maximum 0.31 Ω permitted ... and the actual touch-voltage would generally be reduced ...

BUT, clearly, "limiting the touch voltage to 50 V AC" is NOT the intention of additional protection by supplementary protective equipotential bonding according to Regulation Group 415.2 (otherwise the formulas provided would actually achieve that ... which they clearly are not intended to, as per the discussion above).

"limiting the touch voltage to 50 V AC" is NOT the intention of additional protection by supplementary protective equipotential bonding

Absolutely - I hope I didn't give the impression that I believed it did. I do however believe it's there to reduce touch voltages in general - even if it doesn't provide any unconditional absolute limit (not unlike main bonding in that respect). 50V (or 25V) for 5s just one representative point on the curve of what it must achieve.

   - Andy. 

 AJJewsbury Thank you. I just wanted to push back on this a little bit, for a very good reason.

Throughout my career, I've heard it said many times that BS 7671 has a 'maximum touch voltage' of 50 V AC or 120 V DC ... it certainly has been a belief among some on the industry that this is the case. I've heard it more, to be honest, from designers and engineers working on larger programmes, than electricians.

I would agree that in practice the effective touch-voltage can drop significantly, and may get down to 50 V AC or less. In fact, in industrial plant rooms and large infrastructure, where a common bonding network is in place, and each room has its own earth bar or bonding ring conductor, the expected touch-voltage from AC faults can be much lower than 50 V.

However, I can conceive of examples, particularly where there's a TN-C-S supply, where if you take the resistances to their maximum (say 0.31 Ω where circuits in the location are supplied with breakers with maximum rating of B32) then the touch-voltage might be reduced from, say, 110 V AC to 90 V AC, which isn't really that much of a reduction.

 AJJewsbury Thank you. I just wanted to push back on this a little bit, for a very good reason.

Throughout my career, I've heard it said many times that BS 7671 has a 'maximum touch voltage' of 50 V AC or 120 V DC ... it certainly has been a belief among some on the industry that this is the case. I've heard it more, to be honest, from designers and engineers working on larger programmes, than electricians.

I would agree that in practice the effective touch-voltage can drop significantly, and may get down to 50 V AC or less. In fact, in industrial plant rooms and large infrastructure, where a common bonding network is in place, and each room has its own earth bar or bonding ring conductor, the expected touch-voltage from AC faults can be much lower than 50 V.

However, I can conceive of examples, particularly where there's a TN-C-S supply, where if you take the resistances to their maximum (say 0.31 Ω where circuits in the location are supplied with breakers with maximum rating of B32) then the touch-voltage might be reduced from, say, 110 V AC to 90 V AC, which isn't really that much of a reduction.

Its an interesting point, as in reality as far as hurting people is concerned, what is needed is an over-voltage versus time curve.
It may be that the 'normal' disconnection times cannot be met - which assume exposure to either half supply voltage or almost all of it depending on TNx vs TT earths and time accordingly, but in practice there is a sliding scale of increasing danger from some slightly increased disconnection time, to a situation where the fault voltage persists forever and is never disconnected.

Personally I feel that 50V is rather high for the never disconnect and victim  good contact case, but it pops up in tests for RCDs and 1667 ohm Zs examples, but equally,  in many situations an ADS that should operate in 400ms operating in say 600ms may not be the end of the world if the voltage rise could be nobbled just a bit more or perhaps the contact impedance to the victim increased (thinking of that gravel around sub-stations).

The problem is it is totally impractical to calculate every combination, and even if you could, what would you do with the results. " Danger only for people with unusually sweaty palms and damp footwear, safe for everyone else" is not the pithy and clear sort of text for a triangular warning notice ! 

But in many cases the 50V has to be a bit aspirational as the exact fault current path may not be as expected, so which resistances should we measure, or there may be other factors that make the situation either better or worse, and usually of course, better late than never, something will operate and cut the power.

Mike.

mapj1 said:

Its an interesting point, as in reality as far as hurting people is concerned, what is needed is an over-voltage versus time curve.

You mean like this one?

https://commons.wikimedia.org/wiki/File:IEC_TS_60479-1_electric_shock_graph.svg

Standards committees use the IEC 60479 series for this type of data, and consideration of relevant approaches to protection against electric shock.

You mean like this one?

Although that's based on body current rather than voltage - which hides one of the unknowns - i.e. body resistance. It might well be convenient to assume 1kΩ and so read mA for volts but I suspect that's not always a safe assumption.

  - Andy.

AJJewsbury said:

Although that's based on body current rather than voltage - which hides one of the unknowns - i.e. body resistance. It might well be convenient to assume 1kΩ and so read mA for volts but I suspect that's not always a safe assumption.

There is other data in the series of standards that helps with the different variations - this particular one is body current vs duration as you say.

Voltage vs harm is very difficult, especially bathroom (if we are talking bathroom bonding), and certainly 50 V AC definitely not good for someone immersed in a bath indefinitely ... neither is the 90 V in my example, even disconnecting in 0.1 seconds !

mapj1 said:

Personally I feel that 50V is rather high for the never disconnect and victim  good contact case,

In dry conditions, it might be OK for most people wearing footwear. Bathroom, someone immersed, not so sure at all.

In any case, BS 7671 is not intended to protect everyone all the time ... just most people most of the time. The 5 s disconnection time for distribution circuits in TN systems, and 1 s in TT systems, is a case in point.

gkenyon said:

someone immersed in a bath indefinitely

gkenyon said:

In any case, BS 7671 is not intended to protect everyone all the time ... just most people most of the time. The 5 s disconnection time for distribution circuits in TN systems, and 1 s in TT systems, is a case in point.

Once you get past 5 s, the zones on the graph do not change.

Chris Pearson said:

Once you get past 5 s, the zones on the graph do not change.

Agreed, but the touch-voltage-for-time and touch-current-for-time (with the available voltages and currents) might well show an exposure time of less than 5 s is needed.