Metal CU, TT earth and wiring for SPD

Chris Pearson: 
 

Zoomup: 
 

I come across many TT installations here in the sticks. I normally install a time delayed 100mA R.C.D. in an insulated enclosure before new all metal consumer units, and then R.C.B.O.s in the new consumer units. I have had success with this method in all cases but one.

Seems sensible to me, but you don't need an all RCBO board.

It may even be possible to combine it with SPD. https://www.tlc-direct.co.uk/Products/WYREC2SPD.html?source=adwords&ad_position=&ad_id=315107931576&placement=&kw=&network=u&matchtype=&ad_type=&product_id=WYREC2SPD&product_partition_id=987624279487&campaign=shopping&version=finalurl_v3&gclid=EAIaIQobChMIroCC5ojT8QIVyJ7tCh11ngDlEAQYASABEgKwDvD_BwE

I like all R.C.B.O. boards. Update. Whilst working today I remembered that at the nuisance tripping house it was originally TT earthed. We managed to get an official P.M.E. earth terminal installed so the 100mA S type R.C.D. was then unnecessary.

Z.

Define “nuisance tripping “.

Removing a RCD that is functioning correctly and doing what is supposed to be doing seems the preferred way of resolving “nuisance tripping“ for some people.

Nusisnce tripping is a catch all term which describes a tripping RCD utter regardless of what the cause is.

Thanks for the replies, certainly some food for thought.

I had missed 534.4.7 but I think at least some manufacturers don’t recommend SPD’s downstream of a RCD, will need to check again.

I hadn’t thought about build up of leakage current when there are multiple RCBO’s in a board, I wonder how likely this is to occur in a domestic environment, I could see it as a major concern in a office building full of IT equipment.

Looking at the Wylex RCD device from TLC that Chris pointed out, its in a polycarbonate, not metallic case. My understanding that under regulation 421.1.201 this device would be considered as a similar device and would need to be non-combustible in domestic environments? 

Looking at the data it’s not quite clear if the enclosure is wired with the SPD as a stub, if it’s a stub I think I would need an additional 80A fuse to protect it;  this is something I can ask the manufacturer. But it highlights and area I have been struggling with, we are required to make sure that SPD have OCPD following the manufacturers recommendation but it’s far from easy to find this information.

Regarding the need for an RCD as a main isolator for an all RCBO board the OSG provides an example consumer unit for TT use with this configuration in fig 3.6.3 (i). There is a similar recommendation in NAPIT OSG section 3.5, page 102. Of course there will be different opinions on the suitability of the suggested solutions.

But given both documents state that with a metal consumer unit on TT it’s not recommended to combine a standard isolator on the input with dual RCD because of the risk of abrasion to the connecting cables I assume the same conclusion has to be drawn about interconnections for SPD’s.

I am also wondering if I can put third party SDP’s in consumer units, there are certainly SPD’s available from various companies. But if I can’t mix and match standard RCBO’s and CB’s from different manufacturers in a CU why would it be acceptable to introduce a third party SPD? safe option for now is to use items from same manufacturer.

Overall it looks like my best way forward is a time delayed RCD on the input and then probably a dual RCD arrangement to keep costs down and do more research on what OCPD is required for various SPD’s on the market. Most boards pre populated with a SDD that I have looked at do not include a OCPD to protect the SPD but don't state if anything is required. Hager is an example where so far I have failed to find any data on what if any OCPD is required.

It would be fantastic if someone with suitable knowledge and authority would do some joined up thinking and create a document on consumer unit configurations with SPD’s, taking a special look at TT applications. In my opinion (as someone very new to the industry) it’s a complex subject that needs clarifying and making the wrong decisions could increase rather than decrease risk.

Sparkingchip: 
Define “nuisance tripping “.

Ah! That's where you get the design wrong and OCPDs and RCDs do the job which they were intended to do.

Lewden or Fusebox with built in SPD with OCPD would be my choice. Time delay RCD fitted in the same type of enclosure used by energy suppliers for there isolator.

Just swap out for a Wylex 100ma time delay RCD

some SPDs (Wylex and probably other manufacturers too) have a built-in fusible link  - which when blown, causes the window to turn red because a spring-loaded lever is no longer held down by the link. This removes the need for a separate OCPD, as long as as the backup protection is suitable - a 100A BS 1361 / 88-3 being suitable.

Do please note that SPDs for systems with a TT supply will normally include a gas discharge device (spark gap in a controlled atmosphere ,usually a ceramic pellet) These when they fail, fail  open circuit, unlike the solid state devices that can often fail short. On a TNx set-up with a low Zs, the short circuit failure from live to earth is OK, as the fuse opens. On TT, unless the electrodes are unusually good, this will not happen, and the dead SPD just makes everything live, and nothing trips.

The solid state arrestors are rather better in terms of less of a voltage excess  between no break-over and always break-over voltages,  and do not follow through like an arc can - once struck the voltage has to fall very low for an arc  to go out. 

so SPD units for TT and TNS look different inside, as the TT one will have a gas discharge tube in the earth path.

For TT (from here) The thing with the arrow heads represents the gdt = spark gap.

Looking at the Wylex RCD device from TLC that Chris pointed out, its in a polycarbonate, not metallic case. My understanding that under regulation 421.1.201 this device would be considered as a similar device and would need to be non-combustible in domestic environments? 

Quite a few would argue that it's NOT “similar switchgear" - likewise a simple RCCD in an insulating enclosure. The definitions of a consumer unit, distribution board (of which a CU is one case), and switchgear all involve distinct incoming and outgoing circuits (hence must contain overcurrent protective devices, by the definition of a circuit) or main and auxiliary switching equipment (e.g. main switch and separate controls for outgoing circuits) in the case of switchgear.  If we start over-interpreting the definitions we'll end up thinking we need metal clad cooker control units, shower switches, henley blocks and FCUs.

   - Andy.