Up front Rcd question

[Dirk boddington]

After talking to someone who's Rcd failed them recently during a regular test it got me thinking.

I have a TT system and regularly test my rcd however if it failed to operate during a phase to earth fault then there is no backup.

I considered upgrading the consumer unit and adding rcbos but that's well beyond my budget right now. (they may fit in mine now but it will be very tight)

My second thought was adding a rcd in line with my existing consumer unit to act as a backup. All of my circuits are fed from this single rcd at present as its not a split load unit.

I've spoke to several electricians about it and I've had varying answers so thought I would ask a wider audience.

Can this be done with 2 30ma rcd or would this create nuisance tripping or would one have to be a 100ma?

 

[Rpa07]

@Midwest was referring back to #33 and beyond - he's still pondering. Good thread and nice post @hightower - full circle.

 

[Ian1981]

Hi Dirk,

I believe one of the electricians you spoke to was my uncle, he had a call this week and the conversation went very much like your post. Got to affirm his suggestion - installing a 100mA time delayed RCD before the CU. This would give you your belt and braces and would meet discrimination requirements. This is the industry suggested setup for TT systems.

Also remember, it's not just about disconnection in the case of a fault. If you have bonding correctly installed this will limit the touch voltages while a fault is present, and if the RCD fails to operate the risk of shock is still minimised. Might be worth checking your bonding etc rather than getting pent up about a second 30mA RCD.

Any questions, just ask.

The only way a time delay will offer discrimination is if it's in place of or before a main switch and it's protecting a db with either rcbo s or it's dual rcd.
It's a bit pointless as the 30mA rcd when tripped turns the whole board off anyway in this scenario so what does it discriminate against?

 

[hightower]

Fair point Ian

 

[Dirk boddington]

Hi Dirk,

I believe one of the electricians you spoke to was my uncle, he had a call this week and the conversation went very much like your post. Got to affirm his suggestion - installing a 100mA time delayed RCD before the CU. This would give you your belt and braces and would meet discrimination requirements. This is the industry suggested setup for TT systems.

Also remember, it's not just about disconnection in the case of a fault. If you have bonding correctly installed this will limit the touch voltages while a fault is present, and if the RCD fails to operate the risk of shock is still minimised. Might be worth checking your bonding etc rather than getting pent up about a second 30mA RCD.

Any questions, just ask.

Could well be, I ended up speaking to a few as the conversations with the first two I rang for quotes were very different opinions.

It looks like ill be going down the 100ma route, just waiting for the isolator to be installed from the supplier.

Out of curiosity, you say the bonding will minimise the shock, my bonding if anything may be over the top but if earth goes live and therefore everything that is bonded, am I right in thinking its only minimising the shock as all exposed conductors will be at the same potential? I think my worry comes from it making the bathroom live such as the shower

 

[Midwest]

And I'm still trying to figure out where you think I said it did?

The point I'm trying to make (not very well) is if the OP's 30mA RCD fails to function, 'you' are suggesting a 100mA RCD will provide fail safe personnel protection, it will not. Discuss?

 

[hightower]

The point I'm trying to make (not very well) is if the OP's 30mA RCD fails to function, 'you' are suggesting a 100mA RCD will provide fail safe personnel protection, it will not. Discuss?

And the point is trying to make was that bonding should be considered before worrying about a second RCD. Are you trying to suggest that the OPs request should be carried out? I'm not saying it can't be but should it be? Wonder if 30mA inline will make it's way in to the 18th

 

[D Skelton]

Let me clear this up once and for all. The use of a 100mA s-type is for back up protection only. It is not there to provide fault protection. The fault protection is provided by the 30mA RCD.

If the 30mA RCD fails however, the fact that your s-type won't meet the required disconnection time is the least of your worries, and anyway, there is no requirement for it to do so because A. ADS is only required to provide protection under single fault conditions, and B. Because the s-type is not a requirement in the first place.

To summarise; if the 30mA RCD fails, what's better, no back up protection and every exposed and extraneous conductive part becoming and staying live, or an s-type that trips in 0.3 seconds??

Who the hell cares if it's a tenth of a second slower to trip than required of the 30mA RCD?!

 

[Ian1981]

Let me clear this up once and for all. The use of a 100mA s-type is for back up protection only. It is not there to provide fault protection. The fault protection is provided by the 30mA RCD.

If the 30mA RCD fails however, the fact that your s-type won't meet the required disconnection time is the least of your worries, and anyway, there is no requirement for it to do so because A. ADS is only required to provide protection under single fault conditions, and B. Because the s-type is not a requirement in the first place.

To summarise; if the 30mA RCD fails, what's better, no back up protection and every exposed and extraneous conductive part becoming and staying live, or an s-type that trips in 0.3 seconds??

Who the hell cares if it's a tenth of a second slower to trip than required of the 30mA RCD?!

Time to move on from the thread I think
The Op can do what he likes it's his property.
Think he's got all the info he's gonna get about the matter

 

[D Skelton]

Indeed he can do what he wants. I'm clarifying however for the sake of accuracy that those telling him he can't fit an s-type because "it won't meet disconnection times" are wrong.

They are right in that it won't meet disconnection times, but again, there is no requirement for it to do so.

 

[Ian1981]

Indeed he can do what he wants. I'm clarifying however for the sake of accuracy that those telling him he can't fit an s-type because "it won't meet disconnection times" are wrong.

They are right in that it won't meet disconnection times, but again, there is no requirement for it to do so.

I get what your saying that 300ms as aposed to 200ms is trivial but for compliance to bs 7671, it is what it is.
Anyway it's been a decent discussion

 

[D Skelton]

Again, it doesn't have to comply with BS 7671 with regards to disconnection times.

That's the whole point.

 

[Ian1981]

Again, it doesn't have to comply with BS 7671 with regards to disconnection times.

That's the whole point.

Again, it doesn't have to comply with BS 7671 with regards to disconnection times.

That's the whole point.

so okay the live tail links inside a metal consumer unit short onto the metal casing,
You are within your right to protect the tails via a 100mA time delay rcd either before the db or in place of the main switch.
The tails links inside the consumer unit will carry a max 1 second disconnect time for a TT as if there was no rcd at all the case will undoubtedly become live as the live to earth fault current will be to low to operate the cut out fuse.
Then the time delay like any over device needs to meet a disconnect time of not more than 1 second and it's max time allowed anyway is 500ms during an rcd test.
I get the fact that in this case , that's not the issue.

 

[Dirk boddington]

No fighting guys

Don't worry, I am able to digest the information and I understand what's being said, I'm not a compete novice, it's just been a lot of years since I did any electrical work and it wasn't domestic or circuit design.

I originally figured 2 x 30ma rcd would work as my logic was "why fit a 100ma if I can have another 30ma".

My main concern was that if the 30ma failed I have no backup and my house becomes one big heating element and the occupants get a new frizzy hair do.

I'm satisfied that the 100ma time delay will do the job just nicely.

Thanks for the input though and it's interesting to read the responses (except the ones trying to wind me up)

 

[Ian1981]

No fighting guys

Don't worry, I am able to digest the information and I understand what's being said, I'm not a compete novice, it's just been a lot of years since I did any electrical work and it wasn't domestic or circuit design.

I originally figured 2 x 30ma rcd would work as my logic was "why fit a 100ma if I can have another 30ma".

My main concern was that if the 30ma failed I have no backup and my house becomes one big heating element and the occupants get a new frizzy hair do.

I'm satisfied that the 100ma time delay will do the job just nicely.

Thanks for the input though and it's interesting to read the responses (except the ones trying to wind me up)

You've got the best name on the forum Dirk I'll give you that!

 

[hightower]

Out of curiosity, you say the bonding will minimise the shock, my bonding if anything may be over the top but if earth goes live and therefore everything that is bonded, am I right in thinking its only minimising the shock as all exposed conductors will be at the same potential? I think my worry comes from it making the bathroom live such as the shower

Sorry, missed this post. Yes, anything connected to your main earth terminal will rise to a live voltage. Your shower pipes might not rise to the exact same voltage as your radiator pipes, but the goal is to get it close enough so that dangerous current cannot pass through you when holding them both. If an earth fault occurs and everything rises to the same potential but the RCD fails, you'll still have a fairly safe scenario. This added with regular testing of the RCD should keep you safe.

 

[Risteard]

Tho yes the equation Ra x Idelta N <

Not sure what you meant here?

 

[D Skelton]

so okay the live tail links inside a metal consumer unit short onto the metal casing,
You are within your right to protect the tails via a 100mA time delay rcd either before the db or in place of the main switch.
The tails links inside the consumer unit will carry a max 1 second disconnect time for a TT as if there was no rcd at all the case will undoubtedly become live as the live to earth fault current will be to low to operate the cut out fuse.
Then the time delay like any over device needs to meet a disconnect time of not more than 1 second and it's max time allowed anyway is 500ms during an rcd test.
I get the fact that in this case , that's not the issue.

As you've already said, different circumstances. In the OPs circumstances there is no requirement for it to meet disconnection times as he's not using it for fault protection.

 

[D Skelton]

Sorry, missed this post. Yes, anything connected to your main earth terminal will rise to a live voltage. Your shower pipes might not rise to the exact same voltage as your radiator pipes, but the goal is to get it close enough so that dangerous current cannot pass through you when holding them both. If an earth fault occurs and everything rises to the same potential but the RCD fails, you'll still have a fairly safe scenario. This added with regular testing of the RCD should keep you safe.

That's all well and good until you provide a bridge between the equipotential zone and terraferma.

Oh and by the way, that pair of cast iron lamps at the bottom of the garden path that the postman brushes past daily before delivering your letters, well they've just become a pair of lethal weapons.

 

[Midwest]

I take the points mentioned.

I think we are all agreed, that 100mA RCD will not protect a person from potentially receiving a fatal electric shock, whereas a 30mA RCD will provide that protection.

In a possible scenario (without human interaction) of a 30 mA RCD failing to function, at the time of a fault , the 100mA will eventually trip. Jobs a good'un.

However, in the scenario with human interaction, and the 30mA failing to function, the human will receive a possible fatal electric shock, before the 100mA functions.

As an S Type RCD, has not been designed for the above scenario, i.e. designed for personnel protection, so therefore why should it be conceived as fit for purpose, to provide 'back-up'. What if the 100mA fails at the same time, what's the 'back-up to the back-up'? To provide 'back-up', or fail safe suggests that the secondary device will provide protection if the primary device fails.

As said, the OP can do what he wishes, but it should not be suggested to design a system, which has no endorsements from manufacturers or recommendations from BS7671.

One point I do agree on , is the unnecessary bonding of metal pipes etc, that are not extraneous, reg 415.2.2.

 

[Wilko]

Hi - coming in late to the party (please be gentle) my understanding is - the upfront S 100mA in a TT could help clear a fault inside the metal CU (rare, but perhaps use all RCBO instead) and it could control bonded touch voltages if there was some fault to earth and the 30mA RCD had previously failed (which must also be rare?). But it couldn't replace the 30mA RCD in providing protection for a person who has touched live while in contact with earth potential as its too slow and allows too much current. Have I got it correctly

 

[Risteard]

An RCD cannot control touch voltages - all it can do is disconnect.

 

[Wilko]

An RCD cannot control touch voltages - all it can do is disconnect.

Correct ! Sorry my sloppy wording

 

[D Skelton]

I take the points mentioned.

I think we are all agreed, that 100mA RCD will not protect a person from potentially receiving a fatal electric shock, whereas a 30mA RCD will provide that protection.

Yes we are

In a possible scenario (without human interaction) of a 30 mA RCD failing to function, at the time of a fault , the 100mA will eventually trip. Jobs a good'un.

However, in the scenario with human interaction, and the 30mA failing to function, the human will receive a possible fatal electric shock, before the 100mA functions.

As an S Type RCD, has not been designed for the above scenario, i.e. designed for personnel protection, so therefore why should it be conceived as fit for purpose, to provide 'back-up'. What if the 100mA fails at the same time, what's the 'back-up to the back-up'? To provide 'back-up', or fail safe suggests that the secondary device will provide protection if the primary device fails.

It will provide protection, albeit not ideal protection, but protection nonetheless. In your scenario with human interaction, if no s-type is fitted and the 30mA RCD fails, then there will be no disconnection. At least an s-type will provide disconnection within 500ms, which might not be perfect, but it's far better than the alternative.

Further, the idea behind providing back up protection is that in the event of a fault occurring and the 30mA RCD failing, the s-type will clear the fault in most cases before there is human interaction.

As said, the OP can do what he wishes, but it should not be suggested to design a system, which has no endorsements from manufacturers or recommendations from BS7671.

So your advice is basically; "because BS 7671 doesn't make any recommendations about providing some sort of back up protection in TT systems, don't make the installation safer"?

Sound advice that is!?

The point I'm making is simply this; RCDs are extremely prone to failure, so in an installation where you are relying on a single RCD as your SOLE means of fault protection, it would be prudent to allow for some form of fail safe mechanism. A cost effective means of achieving this is an s-type RCD up front.

The s-type is merely there as a last resort, nothing more. But it's a last resort that could easily save someone's life should the 30mA RCD fail as they often do.

 

[Midwest]

Yes we are



It will provide protection, albeit not ideal protection, but protection nonetheless. In your scenario with human interaction, if no s-type is fitted and the 30mA RCD fails, then there will be no disconnection. At least an s-type will provide disconnection within 500ms, which might not be perfect, but it's far better than the alternative.

Further, the idea behind providing back up protection is that in the event of a fault occurring and the 30mA RCD failing, the s-type will clear the fault in most cases before there is human interaction.



So your advice is basically; "because BS 7671 doesn't make any recommendations about providing some sort of back up protection in TT systems, don't make the installation safer"?

Sound advice that is!?

The point I'm making is simply this; RCDs are extremely prone to failure, so in an installation where you are relying on a single RCD as your SOLE means of fault protection, it would be prudent to allow for some form of fail safe mechanism. A cost effective means of achieving this is an s-type RCD up front.

The s-type is merely there as a last resort, nothing more. But it's a last resort that could easily save someone's life should the 30mA RCD fail as they often do.

I acknowledge your idea, that this design would cause the isolation of the install in the event of a fault and a failure of the 30mA RCD. I also acknowledge that RCD's, like any similar device is prone to failure. This demonstrates the importance of regular testing of them, something advised by the manufacturers.

I respect your views and knowledge, which you seem not to do of others. You often do make well informed remarks as an electrician, but suggesting the use of an RCD, not intended for personnel protection, would mitigate the effects of an electrical shock to a human, is even outside your expertise. All the articles I've read on the risk of electrocution, talk of a very small continuous electric current, 40mA and above, at just over 40ms;
' to cause irreversible damage to the normal cardiac cycle (‘ventricular fibrillation’) or death (‘electrocution’).

I'm no physician, so I'll tend to stick with that advice, and not suggest anything else.

I'm not saying it's a bad design, but it might not prevent fatal consequences in the scenarios discussed. Something the OP should be clear of, and not muddled by an incorrect assessment of the risks of electrocution.

Regular testing of RCD's is important in all designs, and all types of supplies.

 

[hightower]

So what happens if an RCD fails in a TN system? Okay the breaker might disconnect but if the fault is caused by human touch you're still gonna get a couple hundred amps up your arm. I'm having a hard time grasping people's arguments. All you can do is regular testing of the RCD and ensure your bonding is up to scratch. I agree that a 100mA is better protection than none but also agree that it could still cause someone's death. So what are we to do? Start installing RCDs in series as back up? Start installing RCD monitoring devices? Or do we do as suggested to us and push for correct bonding and regular testing. The amount of installs I've seen with no bonding, I'd argue that would cause a death long before a failed RCD would.

 

[Ian1981]

So what happens if an RCD fails in a TN system? Okay the breaker might disconnect but if the fault is caused by human touch you're still gonna get a couple hundred amps up your arm. I'm having a hard time grasping people's arguments. All you can do is regular testing of the RCD and ensure your bonding is up to scratch. I agree that a 100mA is better protection than none but also agree that it could still cause someone's death. So what are we to do? Start installing RCDs in series as back up? Start installing RCD monitoring devices? Or do we do as suggested to us and push for correct bonding and regular testing. The amount of installs I've seen with no bonding, I'd argue that would cause a death long before a failed RCD would.

100mA time delay main switch and a dual rcd db is all you need for this tt.
(Protective bonding as Hightower says as well)
Op doesn't want rcbo s for individual circuits so it's a no brainier for me.
Too be fair there's no requirement to do anything one rcd not ideal but not a massive issue

 

[Ian1981]

An RCD cannot control touch voltages - all it can do is disconnect.

True but the 50v is there to satisfy the calculation of 50/Idelta N giving you the max resistance before the rcd will disconnect in the required time before dangerous voltages appear
Or in this case Ra xIdelta N =<50 volts

 

[DPG]

So what happens if an RCD fails in a TN system? Okay the breaker might disconnect but if the fault is caused by human touch you're still gonna get a couple hundred amps up your arm. I'm having a hard time grasping people's arguments. All you can do is regular testing of the RCD and ensure your bonding is up to scratch. I agree that a 100mA is better protection than none but also agree that it could still cause someone's death. So what are we to do? Start installing RCDs in series as back up? Start installing RCD monitoring devices? Or do we do as suggested to us and push for correct bonding and regular testing. The amount of installs I've seen with no bonding, I'd argue that would cause a death long before a failed RCD would.

No you're not - Ohm's law will prevent it

 

[D Skelton]

I acknowledge your idea, that this design would cause the isolation of the install in the event of a fault and a failure of the 30mA RCD. I also acknowledge that RCD's, like any similar device is prone to failure. This demonstrates the importance of regular testing of them, something advised by the manufacturers.

I respect your views and knowledge, which you seem not to do of others.

I respect the views and knowledge of anyone who's views aren't formed on the basis of a misinterpretation of mine, and who's knowledge is formed on the basis of objective truth.

You often do make well informed remarks as an electrician, but suggesting the use of an RCD, not intended for personnel protection, would mitigate the effects of an electrical shock to a human, is even outside your expertise.

No it's not. It's entirely correct to say that the use of a time delayed RCD will mitigate the effects of an electrical shock to a human far better than nothing at all.

All the articles I've read on the risk of electrocution, talk of a very small continuous electric current, 40mA and above, at just over 40ms;
' to cause irreversible damage to the normal cardiac cycle (‘ventricular fibrillation’) or death (‘electrocution’).

I'm no physician, so I'll tend to stick with that advice, and not suggest anything else.

This information is most likely correct, and is in line with all the information I have read on the effects of electric shock. That said, it's entirely irrelevant to the debate.

I'm not saying it's a bad design, but it might not prevent fatal consequences in the scenarios discussed. Something the OP should be clear of, and not muddled by an incorrect assessment of the risks of electrocution.

No incorrect assessment on the effects of electric shock has been made. The point which you continually seem to miss is that a time delayed RCD will do a far better job of preventing electrocution upon the failure of a 30mA RCD than nothing at all.

Installing an up front time delayed RCD to a TT installation reliant on a single 30mA RCD for fault protection is not a substitution for anything, and is certainly no substitution for regular testing of the RCDs, it's an addition. An addition which will increase the safety of the installation exponentially.

Regular testing of RCD's is important in all designs, and all types of supplies.

I agree

 

[DPG]

Got to say I agree with Mr. Skelton - I would sooner rely on a time-delayed RCD than nothing at all.

 

[hightower]

No you're not - Ohm's law will prevent it

Okay, but you're still going to get enough of a belt to kill you...

 

[DPG]

Okay, but you're still going to get enough of a belt to kill you...

Yes, indeed. Wasn't meaning to be picky, but it's amazing how many people think that you get a bigger belt from a cooker circuit rather than a lighting circuit, just based on the MCB rating.

 

[D Skelton]

So what happens if an RCD fails in a TN system? Okay the breaker might disconnect but if the fault is caused by human touch you're still gonna get a couple hundred amps up your arm. I'm having a hard time grasping people's arguments. All you can do is regular testing of the RCD and ensure your bonding is up to scratch. I agree that a 100mA is better protection than none but also agree that it could still cause someone's death. So what are we to do? Start installing RCDs in series as back up? Start installing RCD monitoring devices? Or do we do as suggested to us and push for correct bonding and regular testing. The amount of installs I've seen with no bonding, I'd argue that would cause a death long before a failed RCD would.

I think you're confusing additional protection with fault protection here buddy.

In a TN system, fault protection is achieved with the use of circuit breakers and/or fuses, and rarely with RCDs. In a TT system (99.9% of domestic TT systems at least), fault protection is more often than not provided by a sole 30mA RCD.

In the event of a L-E fault on a TN system the circuit breaker would disconnect, if that failed you potentially have an RCD, and after that you have the main fuse. The reason for this is because of the low earth fault impedance. A TT system rarely takes advantage of a low earth fault impedance, thus we mainly rely on RCDs, which are prone to faliure in ways circuit breakers and fuses are not.

I hope that clarifies things for you.

 

[Midwest]

I respect the views and knowledge of anyone who's views aren't formed on the basis of a misinterpretation of mine, and who's knowledge is formed on the basis of objective truth.



No it's not. It's entirely correct to say that the use of a time delayed RCD will mitigate the effects of an electrical shock to a human far better than nothing at all.



This information is most likely correct, and is in line with all the information I have read on the effects of electric shock. That said, it's entirely irrelevant to the debate.



No incorrect assessment on the effects of electric shock has been made. The point which you continually seem to miss is that a time delayed RCD will do a far better job of preventing electrocution upon the failure of a 30mA RCD than nothing at all.

Installing an up front time delayed RCD to a TT installation reliant on a single 30mA RCD for fault protection is not a substitution for anything, and is certainly no substitution for regular testing of the RCDs, it's an addition. An addition which will increase the safety of the installation exponentially.



I agree

Can't seem to do the multi quote thing. But lets keep it simple.

Everything I've ever read & heard on the subject of the use of RCD's to try and negate the risk of electrocution, state the use of 30mA RCD. How will the use of a 100mA RCD achieve this, if 40mA is enough to prove fatal?

 

[D Skelton]

Can't seem to do the multi quote thing. But lets keep it simple.

Everything I've ever read & heard on the subject of the use of RCD to try and negate the risk of electrocution state the use of 30mA RCD. How will the use of a 100mA RCD achieve this, if 40mA is enough to prove fatal?

You are also confusing fault protection and additional protection.

If it makes it easier, don't think of it as a 100mA RCD, think of it as a 0.1A circuit breaker that just doesn't provide overcurrent protection.

An RCD in a TT system (in the context of this debate) is there to provide fault protection, not additional protection. It just so happens that boards come preloaded with 30mA RCDs which in the case of a TT system kill two birds with one stone.

Assume a TT system has no requirements for additional protection. No sockets for general use, no buried cable, no bathroom etc... there's no requirement to use a 30mA RCD for fault protection. You could use a 100mA one, a 300mA one, a 1000mA one, assuming a suitable stable Ra value (500 ohms, 167 ohms, 50 ohms respectively)

 

[Midwest]

You are also confusing fault protection and additional protection.

If it makes it easier, don't think of it as a 100mA RCD, think of it as a 0.1A circuit breaker that just doesn't provide overcurrent protection.

My lack of knowledge on that subject to one side, if I touch a live wire with my hand and my bare foot touches earth, I'm led to believe the current flow would be insufficient to operate the fuse or ocpd, but would operate an RCD due to the imbalance. Or is that incorrect?

 

[DPG]

My lack of knowledge on that subject to one side, if I touch a live wire with my hand and my bare foot touches earth, I'm led to believe the current flow would be insufficient to operate the fuse or ocpd, but would operate an RCD due to the imbalance. Or is that incorrect?

I would say that is correct.

 

[D Skelton]

My lack of knowledge on that subject to one side, if I touch a live wire with my hand and my bare foot touches earth, I'm led to believe the current flow would be insufficient to operate the fuse or ocpd, but would operate an RCD due to the imbalance. Or is that incorrect?

I've edited and added some more to my previous response.

In answer to your question, you are correct, and in that scenario, the use of an RCD to prevent a shock as you describe is to use it for the purpose of additional protection.

Additional protection is not a factor in this debate.

 

[Midwest]

So therefore a 30mA RCD would limit the minimise the current flow to 30mA (or there about), whereas a 100mA would minimise it to 100mA (or there about).

40mA is sufficient to prove fatal.

 

[Ian1981]

30mA rcds need a flow and return of the same current in line and neutral if you touch a live part then some current will flow through you to earth and not return through the rcd.
Rcd will detect an imbalance between live and neutral and operate the coil,as long as it 30mA or near that and turn off the device.
Think that's the basic way of explaining it anyway.
Apologies if you know already know this