# Minimum size of CPC

When using the Adiabatic equation, only to calculate the minimum size CPC, I am thinking that using the fault currents given in Appendix 3 of BS7671 rather than that calculated (UO/ZS) would give a more realistic outcome.

For example: A D6 circuit breaker is protecting a circuit where the calculated fault current is 328A. The table associated with Fig 3A6 in Appendix 3 of BS7671 shows that an overcurrent of 120A will operate the 6A breaker in 0.4s. Using these values and a K value of 115 results in a minimum CPC of 0.66mm2 (1.0mm2), whereas if you use the 328A the calculation results in a minimum CPC of 1.80mm2 (2.5mm2).

I cannot see a problem with using the lower value, but would appreciate confirmation and guidance from others.

• I have just checked and the PFC at the origin of the circuit is 1.38kA.

• So, yes , this would be the relevant calculation for you:

Ipf ≤ 3 kA, S ≥ {√(5900)}/115}, or S ≥ 0.67 sq mm, so minimum S = 0.75 sq mm (flex from connection unit etc) or 1.0 sq mm (fixed wiring)

And 1.0 sq mm would therefore seem to be OK. At the 3 kA fault level, you'll see that the 0.67 sq mm value is very close to your original calculation of 0.66 sq mm ... not so when you get to 10 kA fault current where the minimum S comes out at 1.1 sq mm.

• I guess you could consider if the cable is running at a temperature less than 70c, in which case the k value could be more than 115. If the formula is plotted it provides a straight line with k at about 163 at 0C and 115 at 70C. Run at a conservative 50C would give k at about 129.

The question of minimum csa of cpc to meet thermal constraints is a very common one on the 2396 Design course. There would be no expectation to adjust the k value, but I guess the various methods indicate some calculations in installation work are what Mapj referred to as measuring with a micrometer and cutting with an axe!

Using the 5900 figure that gives a csa of about 0.59mm2. As a matter of interest, you would be marked correct in the 2396 Design course using the simple method and the 1.38KA and a fixed time to disconnection of 0.1s, which would lead to a minimum csa of  around 3.8mm2.

• It is rare to see a 1mm2 circuit protected by a 6A breaker running at anything like 70C unless the ambient is unusually hot. Possible in parts of the middle east  - so if 1,1mm seems a squeeze remember the assumption about cable core temp before fault is probably most pessimistic.

Mike.

• False precision indeed - I bet most 1mm2 cables are not made with copper of  exactly 17.241 milliohms per metre length(*) nor for that matter are they actually 1mm2 to within on part in 100,000 either. The world of mass manufacture is not that good, nor does it need to be.  Expecting to be within a few percent may be more credible, and new paper designs that squeak a pass only by that much should probably be re-thought, and any existing  design that fails by that much is no great risk for continued operation.

Mike.

(I tend to use a rule of 16 for  walking around looking at a cable type,  pacing and muttering about voltage drops - allows one to  identify the 'no problem' and the 'clearly a problem' cases very quickly, and the few that are a bit suck teeth and 'that looks borderline' need calculating or measuring to be sure anyway.)

• As a matter of interest, you would be marked correct in the 2396 Design course using the simple method and the 1.38KA and a fixed time to disconnection of 0.1s, which would lead to a minimum csa of  around 3.8mm2.

But if the disconnection time is less than 0.1 s, which is all that the tripping characteristic curves for circuit-breakers illustrates, then para 2 of 434.5.2 is invoked because it's a "shall" requirement:

For a fault of very short duration (less than 0.1 sec), for current limiting devices k2S2 shall be greater than the value of let-through energy (I2t) quoted for the Class of protective device to BS EN 60898-1, BS EN 60898-2 or BS EN 61009-1, or as quoted by the manufacturer.

Why, then, would the answer in a C&G 2396 design course be marked correct when the method used is incorrect according to BS 7671?

The assumption that the disconnection time is "fixed" at 0.1 s is fallacious in the first place.

I accept that the csa that it leads to is greater than necessary, and this has the effect of "safer design", but 4 sq mm csa on a 6 A lighting circuit would be really pushing capability to terminate into devices if nothing else. However, there may be cases with circuit-breakers where the let-through energy for a given prospective fault current is unusually high.

I accept that 543.1.3 doesn't invoke 434.5.2 directly, but it doesn't exclude it either. However, we further ought to note that in the adiabatic formula in 541.1.3, para 2 of 434.5.2 is effectively brought into the description of "I" in the formula in 543.1.3, which differs from the definition of I in 434.5.2, and tells us to take account of let-through energy. The application of the two Regulations, whether for cpc or for live conductors, is the same where circuit-breakers are used for fault protection and the prospective fault currents lead to operation in < 0.1 s.

• GK said; "Why, then, would the answer in a C&G 2396 design course be marked correct when the method used is incorrect according to BS 7671?"

Indeed! It is not me who is marking it correct. In fact, I would be at pains to point out the issue when disconnection is clearly less than 0.1s.

To be honest, my claim about the 2396 marking is only based on candidate feedback after they have sat the exam. I am not permitted to see the paper and the candidate has to sign their individual scripts to declare they will not divulge the questions. Some time later a Chief Examiners Report is provided which is a bit like getting an EICR but not knowing the building it relates to. Absolute balls really, but there you go.

So, despite their solemn vow,  if a candidate divulges to me the nature of a question in their eagerness for a post mortem on the paper they just sat, I am listening to something that may not be accurately relayed.

There are no sample papers and no exemplar answers to allow a better understanding of expectations. I cannot be 100% confident in my assertion but despite efforts in quality assurance, the moderators may well get things wrong. The problem is, we will never know whilst the cloak of secrecy prevails!