From: Andy Wade )

The need to get away from the supplier's TN earth is relevant where
the
equipotential zone concept is difficult or impossible to apply -
hence the
reference to greenhouses and Class I equipment used outdoors.
Remember that
TN earths can rise to potentially dangerous voltages due to (for
example)
cable faults on TN-S systems, or the classic broken service neutral
on a PME
supply. OK, these things are pretty rare, but they do happen, which
is one
reason why equipotential bonding is required. The big advantage of
TT
earthing, properly implemented and maintained, is that your earth
really is
at the local ground potential, and will stay there. Even if the
earth
electrode has the maximum recommended resistance value of 200 ohms
(see
OSG), then to trip a 100mA RCD shouldn't lift the earth by more than
20V.


I think this is something of a misunderstanding, and a very common
one. An RCD doesnt limit fault voltage at all, just as it does not
limit fault current either. What it does is ensure that fault
voltage/current is short lived. With a TT system a leakage fault is
liable to give you more or less the full 240v on your hand drill, RCD
or no RCD. The RCD simply cuts the shock time down.

Given that appliance leakage faults are far more common than PME
earths going live, the advantage is clearly not with the TT option.


The downside of TT is the reliance on RCDs, which don't always fail
safe,
and the need to ensure that the earth system is maintained.

There is that too: RCDs do fail, and earths are normally not
maintained. I've seen whole house earths disconnected at the rod
before now.

In short TT should not be used unless there is no safer method
available. TT allows fault voltages to rise to 240v, and relies
entirely on an RCD to disconnect such faults. Neither of these
potential risks occur with supplier-earthed systems.

If this is going to go into an FAQ this should really be cleared up
first - unless I'm mistook somewhere along the line.


Regards, NT

"N. Thornton" wrote in message
om...

From: Andy Wade

The need to get away from the supplier's TN earth is relevant where
the equipotential zone concept is difficult or impossible to apply -
hence the reference to greenhouses and Class I equipment used
outdoors. Remember that TN earths can rise to potentially dangerous
voltages due to (for example) cable faults on TN-S systems, or the
classic broken service neutral on a PME supply. OK, these things
are pretty rare, but they do happen, which is one reason why
equipotential bonding is required. The big advantage of TT
earthing, properly implemented and maintained, is that your earth
really is at the local ground potential, and will stay there. Even
if the earth electrode has the maximum recommended resistance value
of 200 ohms (see OSG), then to trip a 100mA RCD shouldn't lift the
earth by more than 20V.

I think this is something of a misunderstanding, [...]

Well, I think you missed my point a bit - although I agree that the last
sentence should have ended something like "then with a 100 mA RCD any
sustained lift in the voltage of the installation earths will be limited to
20 V."

An RCD doesnt limit fault voltage at all, just as it does not
limit fault current either. What it does is ensure that fault
voltage/current is short lived. With a TT system a leakage fault is
liable to give you more or less the full 240v on your hand drill, RCD
or no RCD. The RCD simply cuts the shock time down.

Yes, perfectly true, but the requirements for RCD tripping times are based
on coordination with the IEC touch voltage curve, such that the shock should
not cause death or serious injury.

There's a lot of information on shock protection in Paul Cook's Commentary
on BS 7671 [ISBN 0852962371]. See for example Table 4C: in dry conditions a
touch voltage of 220 V is not considered dangerous (in the sense mentioned
above) if its duration is less than 180 ms. Under such conditions the
residual current will be at least 4.4 times the RCD tripping current (220
V/50 V) and the tripping time will be very short.

Given that appliance leakage faults are far more common than PME
earths going live, the advantage is clearly not with the TT option.

Advantage from what point of view? Both systems will give the indirect
contact protection required by BS 7671 - for what you rather ambiguously
call "leakage faults" ('fault' usually means a dead-short). The TT system
has the additional advantage in limiting the sustained touch voltage which
can occur between the installation's protective conductors and the local
'true' ground to a safe value, something which the TN systems don't do at
all. The TN system circumvents this problem by using equipotential bonding,
but that concept fails when Class I equpment is used outdoors, for example.

Moreover, the touch voltage in a TN system in the presence of an appliance
earth fault might be higher than you've realised. This comes about because
of the use of reduced-size circuit protective conductors (CPCs) in (for
example) twin & earth cables. The worst case is with the 4mm^2 cable, which
has a 1.5mm^2 CPC. Thus the resistance of the CPC is nearly 2.7 times that
of the phase conductor, which makes the touch voltage about 73% of the
supply voltage - i.e. about 170V - and it will increase during the fault as
the CPC heats up more than the phase conductor. Also the fault can take up
to 400 ms to clear, or 5 s in the case of fixed equipment. This combination
is outside the safe area of the IEC touch voltage curve (which 'allows' 100V
for 400 ms) but is permitted - by tradition, I guess - except in those wet
areas where local supplementary bonding is reguired. Supplementary bonding
is simply a measure to reduce the touch voltage that can occur between the
accessible exposed-conductive-parts and extraneous-conductive-parts in the
locality.

There is that too: RCDs do fail, and earths are normally not
maintained. I've seen whole house earths disconnected at the rod
before now.

Quite true - although (IME) disconnected earths are more likely to be found
in old house installations with an ELCB and an earthing conductor of puny
size, 1.5 or 2.5 mm^2. With modern practice with a 16mm^2 (or conduit
protected) earthing conductor and a proper joint to the earth electrode in
an earth pit, there should be much less to worry about and 'maintenance' is
really limited to checking the resistance during periodic inspections. The
greater weakness is probably that most people will not bother to do the
3-monthly RCD test.

In short TT should not be used unless there is no safer method
available. TT allows fault voltages to rise to 240v,

Yes, but only outside the equipotential zone, and then only very briefly,
whereas a TN system can produce a touch voltage as high as ~170 V for 5 s
within the equipotential zone (except where supplementary bonding is used).

and relies entirely on an RCD to disconnect such faults. Neither of
these potential risks occur with supplier-earthed systems.

But different risks obtain in the TN case, as I hope I've manged to show.

If this is going to go into an FAQ this should really be cleared up
first - unless I'm mistook somewhere along the line.

I wasn't aware of any FAQ proposal - although I have been meaning for ages
to write a piece on "Electricity Supplies to Outbuildings" for ages. Now
where's that round tuit gone?

Anyway, thanks for an interesting post.

--
Andy

"Andy Wade" wrote in message ...
"N. Thornton" wrote in message
om...

From: Andy Wade

The need to get away from the supplier's TN earth is relevant where
the equipotential zone concept is difficult or impossible to apply -
hence the reference to greenhouses and Class I equipment used
outdoors. Remember that TN earths can rise to potentially dangerous
voltages due to (for example) cable faults on TN-S systems, or the
classic broken service neutral on a PME supply. OK, these things
are pretty rare, but they do happen, which is one reason why
equipotential bonding is required. The big advantage of TT
earthing, properly implemented and maintained, is that your earth
really is at the local ground potential, and will stay there. Even
if the earth electrode has the maximum recommended resistance value
of 200 ohms (see OSG), then to trip a 100mA RCD shouldn't lift the
earth by more than 20V.

I think this is something of a misunderstanding, [...]

Well, I think you missed my point a bit - although I agree that the last
sentence should have ended something like "then with a 100 mA RCD any
sustained lift in the voltage of the installation earths will be limited to
20 V."

An RCD doesnt limit fault voltage at all, just as it does not
limit fault current either. What it does is ensure that fault
voltage/current is short lived. With a TT system a leakage fault is
liable to give you more or less the full 240v on your hand drill, RCD
or no RCD. The RCD simply cuts the shock time down.

Yes, perfectly true, but the requirements for RCD tripping times are based
on coordination with the IEC touch voltage curve, such that the shock should
not cause death or serious injury.

There's a lot of information on shock protection in Paul Cook's Commentary
on BS 7671 [ISBN 0852962371]. See for example Table 4C: in dry conditions a
touch voltage of 220 V is not considered dangerous (in the sense mentioned
above) if its duration is less than 180 ms. Under such conditions the
residual current will be at least 4.4 times the RCD tripping current (220
V/50 V) and the tripping time will be very short.

Given that appliance leakage faults are far more common than PME
earths going live, the advantage is clearly not with the TT option.

Advantage from what point of view? Both systems will give the indirect
contact protection required by BS 7671 - for what you rather ambiguously
call "leakage faults" ('fault' usually means a dead-short). The TT system
has the additional advantage in limiting the sustained touch voltage which
can occur between the installation's protective conductors and the local
'true' ground to a safe value, something which the TN systems don't do at
all. The TN system circumvents this problem by using equipotential bonding,
but that concept fails when Class I equpment is used outdoors, for example.

Moreover, the touch voltage in a TN system in the presence of an appliance
earth fault might be higher than you've realised. This comes about because
of the use of reduced-size circuit protective conductors (CPCs) in (for
example) twin & earth cables. The worst case is with the 4mm^2 cable, which
has a 1.5mm^2 CPC. Thus the resistance of the CPC is nearly 2.7 times that
of the phase conductor, which makes the touch voltage about 73% of the
supply voltage - i.e. about 170V - and it will increase during the fault as
the CPC heats up more than the phase conductor. Also the fault can take up
to 400 ms to clear, or 5 s in the case of fixed equipment. This combination
is outside the safe area of the IEC touch voltage curve (which 'allows' 100V
for 400 ms) but is permitted - by tradition, I guess - except in those wet
areas where local supplementary bonding is reguired. Supplementary bonding
is simply a measure to reduce the touch voltage that can occur between the
accessible exposed-conductive-parts and extraneous-conductive-parts in the
locality.

There is that too: RCDs do fail, and earths are normally not
maintained. I've seen whole house earths disconnected at the rod
before now.

Quite true - although (IME) disconnected earths are more likely to be found
in old house installations with an ELCB and an earthing conductor of puny
size, 1.5 or 2.5 mm^2. With modern practice with a 16mm^2 (or conduit
protected) earthing conductor and a proper joint to the earth electrode in
an earth pit, there should be much less to worry about and 'maintenance' is
really limited to checking the resistance during periodic inspections. The
greater weakness is probably that most people will not bother to do the
3-monthly RCD test.

In short TT should not be used unless there is no safer method
available. TT allows fault voltages to rise to 240v,

Yes, but only outside the equipotential zone, and then only very briefly,
whereas a TN system can produce a touch voltage as high as ~170 V for 5 s
within the equipotential zone (except where supplementary bonding is used).

and relies entirely on an RCD to disconnect such faults. Neither of
these potential risks occur with supplier-earthed systems.

But different risks obtain in the TN case, as I hope I've manged to show.

If this is going to go into an FAQ this should really be cleared up
first - unless I'm mistook somewhere along the line.

I wasn't aware of any FAQ proposal - although I have been meaning for ages
to write a piece on "Electricity Supplies to Outbuildings" for ages. Now
where's that round tuit gone?

Anyway, thanks for an interesting post.


Good points, but 3 points tilt the table a bit:

1. Appliance leakage faults are many many times more frequent than a
supplier earth going live. This rather tilts what is more favourable.
The TT system performs worse on such faults.

2. The fact that with most domestic TT installs, the earthing and RCD
functions are never tested after the installer leaves, not even once.

It is these 2 primarily that lead me to conclude that the odds of a
nasty will be higher with a TT system.

Note also the 170v scenario you present with TN only occurs when the
live to earth fault resistance introduced is zero, and that is
something rarely achieved by real world faults. So the situation with
TN is perhaps not as grim as it may have appeared.

Ultimately both systems have their imperfections, but with the death
rate from fixed wiring being at probably less than 1 per year,
considering all the other hazards in life it begins to become academic
in the end.


Regards, NT

Note also the 170v scenario you present with TN only
occurs when the live to earth fault resistance introduced
is zero, and that is something rarely achieved by real
world faults.

If it isn't zero, then the calculations for earth fault tripping via an MCB
are false, and the MCB may well take more than 10 seconds to trip, which may
be a fire hazard, depending on the nature of the fault. Remember that fire
protection is actually vastly more important than shock protection. Hardly
anyone gets killed by electrocution in the home, whilst many are killed by
electrical fires from appliance earth faults.

Christian.

"N. Thornton" wrote in message
m...

Good points, but 3 points tilt the table a bit:

1. Appliance leakage faults are many many times more frequent than a
supplier earth going live. This rather tilts what is more favourable.
The TT system performs worse on such faults.

Worse on touch voltage, but probably better on fault clearance time. But
does it matter, both systems comply with BS 7671 if properly designed and
built?

2. The fact that with most domestic TT installs, the earthing and RCD
functions are never tested after the installer leaves, not even once.

From posts here it seems to be coming quite common for a periodic inspection
to be done when a property changes hands.

It is these 2 primarily that lead me to conclude that the odds of a
nasty will be higher with a TT system.

Bear in mind the thread's subject. I'm certainly not advocating use of TT
earthing for house installations in preference to use of an available
distributor's earth terminal. That would be counter to the trend in the
supply industry, which is definitely towards universal PME (which principle
is ensconced in the ESQC regulations.) What I continue to maintain though
is the value of the TT system in outbuilding installations where the
equipotential zone principle is difficult or impossible to apply - e.g.
garages, sheds and workshops with damp floors, greenhouses or outdoor
equipment (Class I) with exposed-conductive-parts.

Note also the 170v scenario you present with TN only occurs when the
live to earth fault resistance introduced is zero, and that is
something rarely achieved by real world faults. So the situation with
TN is perhaps not as grim as it may have appeared.

See Christian's post. For design purposes you only need to consider s/c
faults. Any 'leakage' situation which produces enough voltage drop across
the CPC to give a dangerous touch voltage will generate an awful lot of
localised heat, and in practice will quickly turn into a
negligible-impedance fault as insulators track and carbonise, or whatever.
Certainly an RCD will give earlier disconnection in the event of this type
of failure (and might even prevent a fire) and in the TT installation you
get such protection 'for free' (the RCD has to be there anyway). But you
can fit RCDs in TN installations to give similar protection if you choose -
and you then get into the arguments about nuisance tripping (always a sign
of a definite problem, IME) and fire safety, premature loss of lighting from
exit route [see Peter Parry, /passim/].

Ultimately both systems have their imperfections, but with the death
rate from fixed wiring being at probably less than 1 per year,
considering all the other hazards in life it begins to become
academic in the end.

Tell that to the advocates of Part P. (Though I think it's a lot more than
one a year if you include deaths in fires started by wiring problems.)

--
Andy