RCD requirements for TT?

[HappyHippyDad]

Evening all...

I'm having a complete mind blank on what the mA requirement is for an RCD on domestic household with TT earthing, Ra = 57ohms. I'm not talking about final circuits.

I'm planning on splitting the meter tails into a henly, then into a switched fuse, then 50m SWA to consumer unit in garage which will contain a 30mA RCD or RCBO's.

I simply cannot remember or find in the regs if I need RCD protection for the buried SWA. If it was a TN then It wouldn't need it as it has mechanical protection, but it's TT.

I'm sure a 100mA S type would be OK in the switch fuse as replacement for main switch, but I just want to know where in regs it says this will comply, or even necessary?

 

[telectrix]

i f a 100A S will provide disconnection times the should be right.

 

[Flanders]

Think its table 41.5 with an earth loop of 100 ohms or less you can use a 500ma rcd

 

[Strima]

I'd look at using a 300mA to get your earth fault disconnection time.

I also test L-N to ensure the OCPD will operate in case of a dead short, not normally a problem on TT systems but I've had a couple that have been end of line and the PFC has been pretty low.

 

[suffolkspark]

Yes you need upfront rcd as fuses/mcbs wont provide fault protection, must be a time delay so it doesnt trip before downstream rcds, and if using rcbos downstream make sure they also break the neutral.

 

[Ian1981]

100mA S type for fault protection of the distribution circuit.
I don’t know why people bring up the L-N loop test when they wouldn’t dream of testing it if it’s a TN system, unless everyone claims to do the t=K2S2/I2 calculation for every circuit on a TN or TT system.
Just because the earth fault loop impedance is high on a TT , it’s unlikely that the L-N is high as you’d measure this at the origin on a PFC measurement anyway.

 

[i=p/u]

50/0.5=100ohm (') (')

 

[Deleted account]

Forget about the mA of the RCD. Just concentrate on can you achieve disconnection times For each circuit whether distribution or final.

30 mA, 100mA, time delay or not. These are chosen for discrimination and circuit requirements.

 

[Risteard]

it’s unlikely that the L-N is high

If it was you'd have problems with excessive volt drop too!

 

[pc1966]

If it was you'd have problems with excessive volt drop too!

Yes, checking PSCC is important not just for max MCB break capacity, but as it more or less confirms system voltage drop as well.

 

[pc1966]

As folk have said above, you need an up-front RCD to disconnect on any cable faults as the TT earth wont do so one the OCPD.

For selectivity that has to be x3 above the final circuit RCDs and to have some delay. Most common for domestic would be 100mA S-type (just to add confusion on AC/A/B/F/etc nomenclature!) and they ought to be fairly readily available.

With your rod impedance you could use a 300mA (Ra < 167 ohm) or 500mA (Ra < 100 ohm) delay RCDs. But it is unlikely to be needed as:

• A 100mA S-type copes will cope with up to 50mA normal leakage, which is unlikely in any sort of system (short of a massive amount of computers on single phase)
• You are probably not to cascading further RCDs for sub-DBs where you would need to have again a x3 sensitivity factor and longer delays for hard-fault selectivity.

A final concern for selectivity is the downstream RCD or RCBO must also disconnect the faulted circuit's neutral, otherwise a N-E fault (or hard L-E fault on a low impedance load) will still trip the main RCD as they won't stop current imbalance due to the spurious neutral path.

All standard RCD (e.g. dual RCD bard) do this, but many RCBO do not. In the UK it seems you choice for RCBO would be best from the new compact Fusebox, Wylex, or Crabtree RCBO that are neutral-switching DP style.

TL;DR Use 100mA delay up-front, then DP RCBO board (or dual RCD board if you must).

 

[pc1966]

I'm planning on splitting the meter tails into a henly, then into a switched fuse, then 50m SWA to consumer unit in garage which will contain a 30mA RCD or RCBO's.

Just read this again, if it is a TT system at the Henley blocks (supply end, rather than making the garage a separate TT system off a TN supply) then there has to be an RCD after the switched-fuse but before the SWA sub-main run.

Otherwise a fault in that sub-main will simply cause the earth rod to sit at ~230V and nothing will stop it!

 

[cliffed]

100mA S type for fault protection of the distribution circuit.
I don’t know why people bring up the L-N loop test when they wouldn’t dream of testing it if it’s a TN system, unless everyone claims to do the t=K2S2/I2 calculation for every circuit on a TN or TT system.
Just because the earth fault loop impedance is high on a TT , it’s unlikely that the L-N is high as you’d measure this at the origin on a PFC measurement anyway.

That’s a good point

 

[cliffed]

Going back to the L-N loop, it becomes a problem when the Zs on that circuit is high & if on a Tncs some just stick a Rcbo on it.
This protects the earth fault scenario but will the mcb protect the short circuit scenario.

 

[pc1966]

Perhaps over-thinking this, but what is the garage CU going to have?

If it is only lights and a few sockets, you could make the supply cable RCD a 30mA instant one and just have a switch and two MCB at the garage. However, if it really is going to be a separate office or workshop then you should be designing for selectivity, and in that case:

• Fused-switch as high as SWA cable rating will allow to get good selectivity with the final circuit MCB action.
• Ideally sub-main fuse 1.6 time smaller than the DNO fuse for selectivity in the L-N big-bang fault situation. So if 100A DNO fuse, then 63A is going to achieve total selectivity with it (or 80A/50A, 60A/40A)
• Delay 100mA RCD at the feed end, and DP RCBOs to feed each of the final circuits.
• If you are limited to smaller feed fuses, like 40A, then try to use MCB/RCBO no larger than 20A
• If it is a workshop then also consider some emergency lights in case power fails for any reason. Being plunged in to darkness with a lathe still spinning and sharp tools all around is not a good situation!