Spur line (non ring) on TT supply question.

[itstony]

Hi Guys,
Long time member and not a regular user here. I do pop in here when the odd unknown pop's up and unable to locate an electrical dilemma and cannot find a person that has the expected knowledge and can explain and help, as yet unable to find information even from respected persons.
Let me add whatever basic work I do, an accredited local electrician does inspect and able to check and sign off in my property as required.

Everything is spur, no ring. There is one long spur from the main consumer box to a pool house and another small consumer box. Another spur from pool house to a later added garage.
Decided to upgrade the original long spur 50m+/- with new 4mm2 cable.

I had upgraded the TT supply (non supplied earth) from main consumer box with 6mm2 cable & copper rod. Same on the pool house with another rod. Not on the garage 3m away to another consumer box. Everything has been linked on the earth side everywhere possible. Always OTT on grounding.

The main consumer box supply has a a 20A MCB (as isolation only) 40A 30mA RCD. At the pool house consumer box 20A MCB 30 mA. Only issue has ever been pool lights tripping and replacing.

Had a malfunction on a pump frequency converter at the pool house, MCB tripped not the RCD, only the main consumer box RCD.
During trouble shooting, the same main RCD was the only item tripping.

I would like to obtain some information that I can read and learn from on how a a spur line should be setup and perhaps installation of a double RCD's or better design.
Or comments on what you see from this please. Never too old to learn.

Apologies for any over explaining, just my engineering training to supply all info you can.

 

[pc1966]

Everything is spur, no ring.

The ring final circuit is basically a UK thing because we have fused plugs. That means we can have a 32A breaker feeding the sockets and not worry about small appliance cable size because the fuse in the plug protects that. Having 32A means you can feed a lot of sockets off one circuit, but that usually means a few rooms, and probably a loop. So by converting the circuit in to a ring you get more area covered for a given voltage drop / cable size.

But for most other countries without fused plugs the breaker in the CU has to protect both the fixed wiring and the appliance cable, etc. Hence the limit of 16A/20A in most cases. Less sockets, no need for a loop, etc.

There is one long spur from the main consumer box to a pool house and another small consumer box. Another spur from pool house to a later added garage.
Decided to upgrade the original long spur 50m+/- with new 4mm2 cable.

I had upgraded the TT supply (non supplied earth) from main consumer box with 6mm2 cable & copper rod. Same on the pool house with another rod. Not on the garage 3m away to another consumer box. Everything has been linked on the earth side everywhere possible. Always OTT on grounding.

Good grounding and more than one rod is always a good policy!

The main consumer box supply has a a 20A MCB (as isolation only) 40A 30mA RCD. At the pool house consumer box 20A MCB 30 mA. Only issue has ever been pool lights tripping and replacing.

Had a malfunction on a pump frequency converter at the pool house, MCB tripped not the RCD, only the main consumer box RCD.
During trouble shooting, the same main RCD was the only item tripping.

I would like to obtain some information that I can read and learn from on how a a spur line should be setup and perhaps installation of a double RCD's or better design.
Or comments on what you see from this please. Never too old to learn.

Apologies for any over explaining, just my engineering training to supply all info you can.

What you are seeing is a lack of selectivity (formally known as discrimination). That is the ability of cascaded protective devices to properly isolate the faulted segment downstream, without the upstream protection tripping.

For MCB feeding MCB you will get acceptable overload selectivity with a ratio of 1.6:1 or more (e.g. 32A feeding 20A) as then if you draw, say, 50A the 20A breaker goes every time and 32A is slower to act, so overload is disconnected before it heats to tripping point.

But MCB to MCB have very poor fault selectivity. So if there is a short and 400A flows, then both MCB will have hit the "instant" magnetic trip point (for B-curve that is 3-5*In, so max for 32A is 160A, etc) and so both MCB are already beyond the point of no return before either can disconnect the fault.

The same applied to RCD (or RCBO) in cascade. If you want selectivity there you need two things:

• Around a 3:1 difference in current threshold (say 100mA feeding 30mA, as typical RCD spec is you can trip anywhere between 0.5*In and In, so here it might go at 50mA, etc.) so a small fault current always trips the downstream device.
• A short delay, typically 200ms, between RCD so on large earth fault currents, when all RCD have their threshold exceeded, the upstream RCD will wait long enough for the downstream one to act, assuming it can.

However, if you go for something like a 100mA selective (type-S) RCD with a delay it won't protect you from electric shock to an acceptable degree, so it should be used only to feed cables (such as SWA) that are protected against a nail or something penetrating and leaving something live (as the earthed armour forces a fault to causes disconnection even if penetrated) , and used to feed RCD/RCBO for your final circuits such as sockets, etc.

 

[itstony]

Thank you for the reply.

Have not worked on a ring system for over 25yrs and only imho did see weaknesses in the ring system, but tbf there are in the TT also and this is all I work with now. They are just different systems that were chosen and established in different countries. Plus a need for more MCB's and RCD's which personally I am OK with.
Safety and functionality always my focus.

Totally agree on the grounding!

This below I had implemented based on only logic only. Thanks for removing luck.
** For MCB feeding MCB you will get acceptable overload selectivity with a ratio of 1.6:1 or more (e.g. 32A feeding 20A) as then if you draw, say, 50A the 20A breaker goes every time and 32A is slower to act, so overload is disconnected before it heats to tripping point**
Only the ratio 1.6.1 is not fully clear.


I will need to digest more from selectivity, this was my initial logic by stepping down on the RCD's.
At this moment it appears I was nearly there, but wrong choices on ratings. These comment below was key and very useful for me. Getting lucky twice is not really smart to expect. Thanks on this too.
Around a 3:1 difference in current threshold (say 100mA feeding 30mA, as typical RCD spec is you can trip anywhere between 0.5*In and In, so here it might go at 50mA, etc.) so a small fault current always trips the downstream device.

A short delay, typically 200ms, between RCD so on large earth fault currents, when all RCD have their threshold exceeded, the upstream RCD will wait long enough for the downstream one to act, assuming it can

However the last comments delayed type-S RCD gave me a shiver based on my grounding fears and concerns. I'd like not go that route as having a pool must be on point.

Before I hit reply and only based on initial thoughts, the area I need to focus on to improve and implement the 3:1 guide.
Starting by replacing the upstream consumer box RCD 40A-30mA with a 40A-100mA, or 50A-100mA.

And just reviewing the main consumer box and without confusing matters. The outside of properties have now been upgraded to smart meter's and installed with a protection after the main fuse and has to be reset after lightning of failure but not often.
Now I must fess up, although I have no separate isolation switch exterior of the con/box.
What I have are high rated MCB's and the RCD's after each one on all rows of breakers.
On an exterior utility room and lower apartment I have not put that extra RCD's as I had on the long spur. Those both rely on the RCD's i their own consumer box's and do trip alone well when if needed.

Now wondering if I really need that extra RCD in the main consumer box in that long Spur line? I can still isolate with the MCB as I do when needed in the house.
Hope that all makes sense with those las additional thoughts.

 

[pc1966]

This below I had implemented based on only logic only. Thanks for removing luck.
** For MCB feeding MCB you will get acceptable overload selectivity with a ratio of 1.6:1 or more (e.g. 32A feeding 20A) as then if you draw, say, 50A the 20A breaker goes every time and 32A is slower to act, so overload is disconnected before it heats to tripping point**
Only the ratio 1.6.1 is not fully clear.

It is the combination of upper and lower tolerances not overlapping on the thermal curves.

If you want actual limits on selectivity you need to check the manufacturer;s data. For example, if using Hager devices then get the commercial catalogue from here:

Hager Brochures, Catalogues & Documentations | Hager UK

Discover all the product documentation you need in our brochures & catalogues! Download what you need in one click straight from Hager Official website!

hager.com

And look at the tables around page 144 where it lists combinations of MCB and the selectivity limits. So to use the same example, a NBN20 (20A B-curve) from from NBN32 appears oddly not to be selective, but I suspect it is a mistake in the table and the 0.11 kA fault level listed for the 25A upstream case is applicable. I.e. if you fault is over 110A then both trip.

A short delay, typically 200ms, between RCD so on large earth fault currents, when all RCD have their threshold exceeded, the upstream RCD will wait long enough for the downstream one to act, assuming it can

However the last comments delayed type-S RCD gave me a shiver based on my grounding fears and concerns. I'd like not go that route as having a pool must be on point.

The feed to the pool or any similar thing must be on a 30mA 'instant' RCD, but any device feeding that RCD on safe cable could be a 100mA delay - as you are unlikely to be exposed to any fault there.

Before I hit reply and only based on initial thoughts, the area I need to focus on to improve and implement the 3:1 guide.
Starting by replacing the upstream consumer box RCD 40A-30mA with a 40A-100mA, or 50A-100mA.

Note that the RCD current limit is simply what it can carry without damage. RCD do not limit current! You need a RCBO for that protective aspect.

And just reviewing the main consumer box and without confusing matters. The outside of properties have now been upgraded to smart meter's and installed with a protection after the main fuse and has to be reset after lightning of failure but not often.
Now I must fess up, although I have no separate isolation switch exterior of the con/box.
What I have are high rated MCB's and the RCD's after each one on all rows of breakers.
On an exterior utility room and lower apartment I have not put that extra RCD's as I had on the long spur. Those both rely on the RCD's i their own consumer box's and do trip alone well when if needed.

Now wondering if I really need that extra RCD in the main consumer box in that long Spur line? I can still isolate with the MCB as I do when needed in the house.
Hope that all makes sense with those las additional thoughts.

I am afraid your description of the layout is not clear. A diagram would be a much better thing to post.

 

[itstony]

Been looking at the brochure on Hager over the weekend and trying to work out what I may need.

I don't have anything decent to make a wiring diagram just now and agree a drawing makes life much easier.
Have a lot on today and at least i can try and supply a little more info that hopefully make it clearer for now?

Main consumer box has 3 rows of nine with 29 options.

Main entry MCB is 60A from 2x10mm2 no supplied earth. Yes I really should have a separate and ext isolation switch, but don't as yet.

All house protection on the left to right. Top row 6xMCB'S 1x #6 RCD. Last two top row, MCB Isolation1xRCD** Pool house Spur. 4mm2. ** I believe this RCD 40A 30mA is not required and needs removing as per below and same as below and retain only isolation.

Rows two & three 6x MCB's #7RCD. 2xMCB isolation Utility room & lower floor with secondary consumer box's.
These two trip with no effect to the house.

All cable to MCB's and RCD's are 6mm2.

TBH only from your comments did it make me revisit and see this and hope that is correct.
Please don't beat me up on using logic again, but that's where I am without confirmation.

Back to the long Spur if correct I can now rely on 25A 30mA RCD's in the pool house small consumer and garage small consumer box which split end of Spur.
It now makes it (appear) clear why the extra why the upstream 30mA tripped and not the end of spur 30mA didn't.