BS88 max ZS table for fuses over 63A TT or not TT

[John-]

Hi

First time with this kind of a job for me.

I have a scenario where i am planning to install about 50m 25mm TPN from a TNCS supply to supply a small metal container housing TPN pump control gear to control two TP 24KW .8 PF pumps (to be fair one is a 15kw, but allowance to swap out one for a bigger one later). Design Current and Volt drop so far state 25mm will be fine.

Because of the distance i would have thought exporting PME would not be a good idea and was going to make it TT as there would be metal railings, the container itself etc. and we are in a very rural area, but...

Speaking with the Pump installers supplying the control gear they do not convert to TT, they always use the earth supplied - but to be fair they are also very close to the supply and have very low ZS.

I have not yet measured the ZS (nor prospective), but i expect ZS will be .2 or above and the supply fuses i am told are 100A, newly installed by DNO. The client took responsibility for this.

First Q, if i am to rely on the suppliers fuses (to protect only the cable between the meter room and the metal cabinet) and install only an isolator at the mains intake to isolate the supply, does anyone have any charts available please that cover Max Earth fault Loop impedance values for the BS88 fuses? Regs book only goes to 63A. BTW the pumps are individually OC protected in the control gear and internal wiring in the control gear to this point is also 25mm.

Also how do i tell what version fuses they are without breaking the seals?

I do not like the idea of leaving the suppliers fuse as the only means of protection for the 50m run of cable into the control gear - it just feels wrong?

I need to confirm but i am pretty sure there is a lockout on the control gear so that both pumps do not start at the same time. So even at 43A (one pump starting) there is a lot of load for about 10 seconds or so TBA.

So if i look at putting in an MCB at the meter end, i need to be sure that they will discriminate (something the pump guys said you will not guarantee even if you had scope to lower the rating of the MCB to 80A) between DNOs and mine and i tend to agree. The pumps, whilst having a slow feature start, can draw a heck of a lot of current on start up, meaning that a C Type MCB maybe the only reliable option, even assuming that the pumps do not start simultaneously. .22 Max ZS for a 100A Ctype, so an MCB may not be an option anyway, may already be too high for the length of riun and Ze. So if cannot discriminate or be selective, what is the point of having additional OC at the meter - that is my thinking. If worst happens and there are issues, i can leave BS88 fuses on site to replace any that go. If they do go though, that's going to be one hell of a fault or lightning strike most likely needing more than just a fuse...

So assuming the maz ZS for the 100A 88 fuses is greater than that measured at the end of the cable supplying the isolator for the control equipment, then TT would not be needed. If the max ZS was lower than the calculated value, and increasing the cable size was not practicable, then I will have to make it a TT and supply a 100mA S type RCD when it comes into the cabin, disconnecting PME and staking it.

Welcome yo ur thoughts please.

Thanks.

John

 

[John-]

I think i got to the same / similar answer:

r1 will always increase as you move away from the supply.
A will also increase as you move away from the supply
B will decrease as you move away from the supply
C will be a constant but is in series with B.
Both B+C are in parallel with A.

Therefore

r1+r2 = r1+ 1/((1/A)+(1/(B+C)))

Bugger, wont let me attach a spreadsheet.

But yea, around 50m i make it

 

[pc1966]

I think i got to the same / similar answer:

r1 will always increase as you move away from the supply.
A will also increase as you move away from the supply
B will decrease as you move away from the supply
C will be a constant but is in series with B.
Both B+C are in parallel with A.

Therefore

r1+r2 = r1+ 1/((1/A)+(1/(B+C)))

Basically that is how I calculate it, but using some python script so I get a pretty graph out of it, etc. Putting in your numbers I get:

Enter line R1 in mOhm/m 0.55
Enter CPC R2 in mOhm/m 0.55
Enter armour R2 in mOhm/m 1.6
Enter length in meters 107
Maximum fault R1+R2 = 0.104 at 96.7 meters

This forum won't allow spreadsheets, no doubt it won't allow python though I could just paste the text to allow others to use it / check it.

 

[pc1966]

Realistically you will struggle to meet disconnection for 107m under those values on the OCPD alone.

You might find it better to look at putting in a fancy adjustable MCCB style of RCD at the feed end set for 1 sec delay and 0.5A or more trip so you might get some selectivity with final circuit 50A BS88 fuses if they can clear a fault in under 1s (as well as no issue with double pole 30mA RCBO or RCD)

You might even get away then with 4-core and parallel 10mm together with the armour as CPC, saving a small bit to help cover the fancy RCD cost!

 

[pc1966]

Another option would be parallel runs of 16mm SWA or similar for each pump if that would allow 50A switch-fuses at supply to disconnect, but location for isolation might require another one closer, etc.

 

[pc1966]

Two 16mm 4-core in parallel have enough SWA to meet TN-C-S 10mm copper CPC equivalent and if putting in values it comes to:

Enter line R1 in mOhm/m 1.15
Enter CPC R2 in mOhm/m 3.1
Enter armour R2 in mOhm/m 3.1
Enter length in meters 107
Maximum fault R1+R2 = 0.289 at 107.0 meters

That along with your supply Ze would be fine for 50A fuse on 5s disconnection (0.27 + 0.29 = 0.56 < 0.79 ohm). You would need yet another SWA for the ancillary stuff to a small CU but if 10-20A then it could be smaller, even single-phase, or maybe just more 16mm.

Cost would need looking in to, but it might actually be easier to wrangle 3 * 16mm 4-core in one duct/route than 35mm 5 core.

 

[John-]

Another option would be parallel runs of 16mm SWA or similar for each pump if that would allow 50A switch-fuses at supply to disconnect, but location for isolation might require another one closer, etc.

Overloads are already provided in the controller.
The run should not be anywhere near 100m...
More worryingly for me right now is that our answers do not match, inclined to think that it is my maths not yours .

I would love to have a look at your code, but right now i need to get my head around what i am doing wrong here...

I found some bracketing errors and sorted those but still not matching. Then i realised that i may not have solved the parallel resistance values correctly for A, B and C?

I am now wondering my assumption that B and C are a simple series circuit in parallel with A is the issue:

This is the formula i used

r1 + (1/((1/(B + C)) + (1/A))) = r1+r2

r1 is in series with A, B and C so no special care needed for r1.

A has a parallel path back with B and C. And B and C are in series, but collectively in parallel with A.

Any ideas please?

This is the spreadsheet with the formulas revealed-: -

Cheers

John

 

[pc1966]

The effective R2 is A in parallel with B+C as you say, can look later at your numbers and another look at my code as well, out at the site so don't have it with me!

 

[John-]

Thank you , that would be great .

I didn't know if it was more complex, along the lines with this example - but it seem you agree with me and therefore it is not: -

If my math is correct i am scuppered anyway, because at just after 40m 5 core 35mm i exceed max ZS just after 40m away form the supply: -

 

[John-]

Spotted an error on the value for R1 for 35mm, i incorrectly stated r1+r2 = 1.1, but is should be 1.048 - not that it made any significant difference.

 

[pc1966]

Checking the numbers I used before getting 0.104 ohm at 96.7m of 107m we have:
R1 = 0.55 * 96.7 / 1000 = 0.053185
A = 1.6 * 96.7 / 1000 = 0.15472
B = 1.6 * (107 - 96.7) / 1000 = 0.01648
C = 0.55 * 107 / 1000 = 0.05885
B+C = 0.07533
R2 = A||(B+C) = 0.050663
R1+R2 = 0.103848 which is 0.104 ohm to 3 digits.

 

[pc1966]

Overloads are already provided in the controller.

It is for fault protection, and to prevent a fault taking out the DNO fuse (which is basically the same as protecting your own 100A fused-switch).

The run should not be anywhere near 100m...
More worryingly for me right now is that our answers do not match, inclined to think that it is my maths not yours .

Try it with some of the numbers I just did to see if they agree at all, I can post my script later as text if anyone wants to play with it. In the mean time I re-ran a manual calculation for the solution my script found and it seems to make sense.

 

[pc1966]

OK, got my script and edited to add end-of-cable result, ran with your values:
Enter line R1 in mOhm/m .55
Enter CPC R2 in mOhm/m .55
Enter armour R2 in mOhm/m 1.6
Enter length in meters 100
Maximum fault R1+R2 = 0.097 at 90.4 meters
End fault R1+R2 = 0.096 at 100.0 meters

 

[John-]

OK, got my script and edited to add end-of-cable result, ran with your values:
Enter line R1 in mOhm/m .55
Enter CPC R2 in mOhm/m .55
Enter armour R2 in mOhm/m 1.6
Enter length in meters 100
Maximum fault R1+R2 = 0.097 at 90.4 meters
End fault R1+R2 = 0.096 at 100.0 m