Domestic SPD installation with a TNC-S earthing system.

If it is TNC-s and you are within a few feet of the N-E bond in the service head, then there is not really much scope to excite a large N-E voltage, and voltages L-N and L-CPC are more or less the same.

However, if there are other items earthed, so the MET is more than just a second neutral connection, and also as the separation from the NE bond increases (allow about a micro-henry per metre ) then the voltages on N and E will diverge. How much divergence is serious enough to warrant a break-over device to clamp the voltage is an 'interesting' question. I'd suggest a few metres.

I'm sure there will be some maker's advice based on very generic assumptions that are intended to sells slightly more SPDs than we really need that contradicts this.

YouTube video


I watched the video last night and have been considering the lack of an earth conductor. 


I thought an earth connection is required to discharge the surge and it is part of the less than 300 mm live earth loop that is required, that can be stretched to 500 mm, but not omitted altogether. Either way not connecting the SPD earth terminal to the MET seems a bit lack.


Andy Betteridge

Particularly when the terminal is only a couple of inches from a earth bar.


But that's not going to be less than a 300 mm live earth loop or is it?


Andy Betteridge

That setup has a cable length from the N-E link at the cut out to the CU of about 1m, and the length of the SPD leads to the CU is about 0.5m. So...


That model of SPD reduces transients across its L-N terminals to about 1kV. That means it reduces transients across L-N at the CU to 1kV plus inductance from 0.5m x 2 of cable; and reduces the transient across L-E at the CU to 1kV plus inductance from 1m of SPD cable, plus 1m N CU-to-link cable and 1m E CU-to-link cable. The BEAMA guide to SPDs implies that you can easily add another 1kV per m of cable, so devices downstream of the CU may see L-E transients only reduced to 4kV.


Having said that, the E SPD connection isn't required for the L-N part of its functionality to work.

Without the earth connection the surge on the live has to discharge to the neutral through two surge devices in series, does that work?

Without the earth connection the surge on the live has to discharge to the neutral through two surge devices in series,

Only if they're wired CT1 - wire them CT2 and there's only one between L and N. It's probably easier though to go for a single two-terminal SPD and just connect it L-N anyway.

does that work?

If you did have two devices in a CT1 arrangement then it would sort of work, but the limiting voltage would be the sum of the two devices - so not ideal.


  - Andy.

Hi all reading the calculations of the effect of wiring inductance how would an SPD work in a house with no N-E bond except one possibly out in the street? Seems like they wouldn't be terribly effective

Hi all reading the calculations of the effect of wiring inductance how would an SPD work in a house with no N-E bond except one possibly out in the street? Seems like they wouldn't be terribly effective

Indeed - which is why you'd normally need (at least) two. One connected L-PE and the other N-PE (CT1 - usually on TN systems) or one connected L-N and the other N-PE (CT2 - usually on TT but can also be used on TN).  Sometimes two SPDs are packaged up together into single 3-terminal device - just to confuse matters.


  - Andy.

AJJewsbury:

or one connected L-N and the other N-PE (CT2 - usually on TT but can also be used on TN).  Sometimes two SPDs are packaged up together into single 3-terminal device - just to confuse matters.

Which as it happens is an exact description of the SPD featured in the video that has the E terminal left unconnected.