Ring Main - limits on length

[ericmark]

Line - earth needs to be 1.37Ω which at 230 volt = 167 amps using ohms law. To ensure a type B 32A MCB will trip.
Line - neutral needs to be 0.94Ω (assuming incomer at 0.35Ω) which at 230 volt = 244 amps using ohms law. To keep volt drop to 5%.

I have said before I can't think of a case where excessive volt drop would be dangerous. With say a fan heater the fan could stop and as a result it could over heat but for the fan to stop the volt drop would have to be huge and the thermal fuse would blow anyway. Yes it could damage refrigeration equipment where the old overload is used, but modern inverter units would not be effected unless a massive drop.

As to line - earth using even an EZ150 Martindale tester should ensure it is safe. OK it shows fault at 1.5Ω rather than 1.37Ω but only at centre of the ring would the trip not trip with a fault so if the Martindale tester shows OK on all sockets then in real terms only the centre sockets could have a problem and the RCD would likely trip anyway.

Why the Martindale EZ150 has the first light at 1.5Ω I don't know. But the price jumps from £50 to £250 if you want to use a proper loop impedance tester however for a full re-wire I think one should use a proper loop impedance tester. The same goes for the RCD tester, in my house where I know the RCD does work due to it tripping from time to time I still don't know if it trips within the 40 milliseconds without a proper tester.

But where one pays the LABC to be responsible for site safety then I would say you can reasonably expect them to test using a loop impedance tester and with a RCD tester to verify your readings. How many tests are done I suppose would depend on how happy the inspector is with your work. I suspect the LABC inspector is lazy and only tests one or two sockets, but at the end of the day should the DIY guy get it very wrong and as a result kill some one, it would be the LABC inspector who would be in court as the DIY guy has paid him a lot of money to do that inspection.

So in real terms getting it wrong could mean the circuit will fail inspection. However from what has been said that's very unlikely. Yes the LABC inspector who was responsible for my dad's house never set foot in the house after I had completed the work, likely seeing we did have the meters he was sure I would have done a good job, and since I do have the exams under my belt I could have not really have told a court I thought it was OK if it was not, the exams show I have the skill.

But assuming the DIY guy has not passed exams to show he knows what he is doing then the LABC inspector would have to visit or risk a court appearance should something go wrong.

 

[JohnW2]

Line - earth needs to be 1.37Ω which at 230 volt = 167 amps using ohms law. To ensure a type B 32A MCB will trip.

Indeed so. As you say, 167A is just about enough (160A needed) to ensure that a B32 trips magnetically.

Line - neutral needs to be 0.94Ω (assuming incomer at 0.35Ω) which at 230 volt = 244 amps using ohms law. To keep volt drop to 5%.

I am having difficulty in making much sense of that statement. When you say "assuming incomer at 0.35", I presume you are talking about the Ze - i.e. the L-E loop impedance (the same as L-N loop impedance for TN supplies) at the origin of the installation. That means that to achieve the required Zs of the circuit for adequate fault protection (1.37Ω), the within-installation L+CPC impedance (i.e. "R1+R2") needs to be a maximum of 1.02Ω (1.37Ω - 0.35Ω).

How that relates to VD (which depends upon R1+Rn) will depend upon what cable is being used, since the relationship between (R1+R2) and (R1+Rn) varies with cable size. The "incomer" (Ze) is obviously irrelevant to VD, which is measured only within the installation. Assuming I've done my sums correctly, for a ring final (if that's what you are talking about), the maximum effective (R1+Rn) (across both legs of ring) to produce an 11.5V VD (5% of 230V) with a 26A load applied at mid-point of ring would be 0.44Ω (or 0.36Ω for 32A applied at the mid-point of the ring). If the ring final were wired in 2.5mm² T+E those maximum R1+Rn figures would equate to maximum R1+R2 of roughly 0.57 (for 26A) and 0.47Ω (for 32A), implying maximum Zs figures (if one wanted VD ≤5%) of 0.92Ω and 0.82Ω respectively if one assumed a Ze of 0.35Ω.

However, I cannot really understand how your statement fits with this - so perhaps you could explain.

Kind Regards, John

 

[ericmark]

I am assuming a TN-C-S which is why I quoted the 0.35Ω which is considered as maximum incomer reading and since neutral and earth are combined near the head the reading will be the same for both line - earth and line - neutral.

Ring design current is 26 amp and with installation method 100 the tabulated current carrying capacity of the cable is 42 amp so the correction factor is 0.917 so the mV/A/M changes from 18 to 16.5. So impedance of line to neutral conductors is 0.79Ω to get a volt drop of 11.45 volt which is just under 5% I am working on an Ambient temperature of 30°C and a Max permitted operating temp tp of 70°C with Rating factor for grouping Cg at 1 and Rating factor for ambient temperature Ca at 1. I took the impedance at centre point as 0.94Ω and worked from that which actually calculated out at 106.773644469205 meters but that is close enough to 106 meters to show my calculations agree with what was told to me at the IET meeting so I am reasonably sure my java script is giving the correct figures.

Like you I did the rough calculations and could not get to the 106 meters quoted. So like you I enquired to find out why. In the main it is the mV/A/M correction and use of 26A as the design current which results in the 106 meters and I had with rough calculations used non corrected volt drop and 32A.

The formula in the BS7671 book is rather complex and I would not want to be doing that every time hence first I used Excel and when my new PDA stopped working with Excel I converted it to Java script and I did find some errors in the original Excel program.

Personally taking the ring final to the maximum size I think is silly, it was I am told 80 meters when it was 4% volt drop but never worked it out. At 106 meters however it is simple you use one reel of cable with 100 meters on the reel that has been the rule of thumb for years even when it was officially 80 meters.

If you measure the two impedances and subtract one from the other you have two readings both which can be slightly out, so in real terms one can only highlight it as a fault when it is well over the limit and although in theroy we can calculate the length in practice it's rather hit and miss. I altered the readings by just one number so 0.34 and 0.95 and volt drop increases to 11.85 and other direction 0.36 and 0.93 it becomes 11.07 this is with just 0.01Ω error with the meters. If the figures showed 20 volts volt drop OK clearly there is an error but to try and fault an installation with only a small error would be near impossible due to errors in measurement. The Java Script program may work to 11 decimal points but the meters do not and also we need the central socket and in real terms all we can do is find socket with worst readings.

So yes I can quote the figures but really it does not matter as our readings are just not good enough.

As to the earth loop impedance with a TN-S supply it can be 0.8Ω at the incomer which means the 1.368Ω is reached first. I can't see how one can accept using a RCD with a TT supply as the earth fault disconnection device and say you can't use it with a TN supply? The only difference is with TT supply it is more likely to detect a neutral - earth fault.

With a 45A supply to shower or cooker only looking for 0.97Ω total so minus 0.8 that's just 0.17Ω which means one is very limited to cable length. This is not allowing for the impedance of the MCB or fuse. I think we have to rely on RCD's we really have no option.

 

[JohnW2]

I can't see how one can accept using a RCD with a TT supply as the earth fault disconnection device and say you can't use it with a TN supply?

One can but presume that 'they' have more confidence in an OPD than an RCD as a means of proviing fault protection. I suppose it's not all that unreasonable to think that magnetically detecting a current of 'dozens or hundreds of amps' might be more reliable than magnetically detecting a current imbalance of less than one-thousandth of that current.

Hence, I imagine that, if they could, they would demand an OPD as the primary fault protection device with TT as well as with TN. However, since that simply is not technologically possible, they have no choice but to accept an RCD as the primary (only) fault protection with TT.

With a 45A supply to shower or cooker only looking for 0.97Ω total so minus 0.8 that's just 0.17Ω which means one is very limited to cable length. This is not allowing for the impedance of the MCB or fuse. I think we have to rely on RCD's we really have no option.

Whilst I don't disagree with what you're saying, it is not necessarily straightforward to argue that reliance on an RCD in that situation is reg-compliant. In any event, even with your figures (most TN-S installations probably have a Ze lower than 0.8), 10mm² T+E would allow about 22m (without any 'correction' of cable impedance etc.) - probably enough for many/most shower circuits.

Kind Regards, John

 

[JohnW2]

I took the impedance at centre point as 0.94Ω and worked from that which actually calculated out at 106.773644469205 meters but that is close enough to 106 meters to show my calculations agree with what was told to me at the IET meeting ...

Fair enough. We've discussed this many times before and, although the 'assumption' seems pretty reasonable (20A at midpoint of ring and 12A uniformly distributed over rest of ring, mathematically equivalent to 26A at mid-point), other than your having been told it at an IET meeting, I have not seen that documented (or even mentioned) anywhere as being an acceptable way of calculating/estimating VD. In the absence of what you were told at that meeting, one could not be blamed for thinking that one had to make the 'worst case' assumption - i.e. 32A at the midpoint of the ring.

The reality, of course, is that hardly any real-world domestic ring final circuits are going to ever experience anything like the VDs that result from such calculations. That, coupled with the fact that (as you've said) it is very hard to think of any likely safety problems resulting from VDs >5%, is the basis of my inability to get very excited about VD!

In any event, it is basically silly talking about absolute limits of VD when the permitted supply voltage variation is so wide. If, as is the implication, it is deemed acceptable to have a VD of 11.5V ("5%") when the supply voltage is only 216.2V (i.e. assuming that loads are happy down to 204.7V), then the 'average installation' with a supply voltage of around 240V could have VD of 35.3V ("15.3%") before loads got any less voltage, and even those receiving 'nominal voltage' could have a VD of 25.3V ("11%").

Kind Regards, John

 

[EFLImpudence]

Line - earth needs to be 1.37Ω which at 230 volt = 167 amps using ohms law. To ensure a type B 32A MCB.

Whilst that may be the correct maths, the relevant figure is 160A - 218.5/160=1.366.

What about the temperature correction?

Line - neutral needs to be 0.94Ω (assuming incomer at 0.35Ω)

0.94+0.35=1.29
Do you allow for 0.35 regardless or would it be better to say "needs to be 1.29-Ze"?

which at 230 volt = 244 amps using ohms law.

230/1.29=178A

To keep volt drop to 5%.

Not sure which amperage you are using so haven't worked it out.



There are too many variables to state blanket figures.

 

[JohnW2]

0.94+0.35=1.29

Quite. It appears from subsequent posts that eric selected 0.94Ω (and 'assumed' that Ze=0.35Ω) because that was the figure he had to use to get the maximum ring length of 106m, per the assumptions regarding current distribution which he had been "told at an IET meeting".

Kind Regards, John

 

[EFLImpudence]

Ah. The 0.94 is (for 2.5/1.5) the L-N loop when the L-E loop is 1.02 (1.37-0.35).

Does that relate to volt drop?

 

[JohnW2]

Ah. The 0.94 is (for 2.5/1.5) the L-N loop when the L-E loop is 1.02 (1.37-0.35).

What temperature/resistivity are you assuming? If one uses the regs' (70° C) resistivity figures, a "L-E loop" (R1+R2) of 1.02Ω would, by my reckoning, correspond to a "L-N loop" (R1+Rn) of about 0.78Ω, not 0.94Ω. Have I got my sums wrong?

Kind Regards, John

 

[EFLImpudence]

Oops. my mistake. Yes, you're right.

 

[EFLImpudence]

0.78 x 1.2 = 0.936

He could have mentioned it.

 

[JohnW2]

0.78 x 1.2 = 0.936

Ah yes - maybe that's how he got it!

He could have mentioned it.

Assuming your guess is correct, he certainly could/should have!

Kind Regards, John

 

[ericmark]

Yes I used Ct = 230 + Tp - (Ca² x Cg² - Ib²/It²) x (tp - 30) / 230 + Tp in my Java Script program to get the correction factor. When I wrote the program I had no real idea what the results would be. However after writing it I can easy swap all the values and get different results and that is when it became apparent there is really no point.

The meters we use are not accurate enough to be able to show one has 80 meters or 130 meters once it goes silly over the top yes one can show it is wrong but if it passes the 5 x current rating so 32 x 5 = 160A now times an extra 5% so 168A or 1.369Ω then really that is good enough. Even if for volt drop it needs to in theroy be lower.

We can argue about the technical merits but questions like is a RCD an over current device may be good fun in a collage class room but in real terms does the DIY man need to know or care. To me it is an over current device can't see how it can be classed as anything else. The fact that the current measured is the differential rather than current used does not change the fact it measures current.

 

[ericmark]

For volt drop it does not matter if the earth is 1.5mm or 2.5mm only worried about line and neutral not earth loop impedance.

 

[ericmark]

I see my calculator is still on line at http://gw7mgw.co.nf/Electric-Cal.html so you can play yourself. Just press calculate to enter standard ring final values.