CPC calculation

[ETradesmen]

Just got back from looking at a potential job. What a bodge up!

I was called around to be asked to install an isolator for a sub main feeding a consumer unit which a car charger company will come off feeding their charger.

New mains upgrade recently done and the previous electrician has bodged a bit of 16mm t and e straight into a henly block from the meter to feed secondary consumer unit.

I’m thinking the cpc within the 16mm may be undersized but just wanted to clarify if my calculation is correct (it’s been a while)...

The size of fuse I’m thinking of installing is a 50A type b (either RCBO or mcb with rcd main switch).

S = 250A x 11 (squared) / 115 = 7.2mm??

CPC is 6mm.

The sub main in question has no bonding attached to it and is approximately 15mtrs away from main board running underneath the floor.

Please correct me if I’m wrong, it’s been a while!

thanks

 

[Vortigern]

Think you forgot the square root, and the t I don't think would be 11, more 1 or 0.1.
S = square root of I squared x t / K, can't be bothered to look it up or work it out. And as I understand it this equation if for the main earth cable to to marshalling point.

 

[ETradesmen]

Hi Vort,

I'm getting the below on my calc?

(250x250) x 0.1 = 6250

6250 squared is 79.05, 79.05 / 115 = 0.68mm?

 

[Vortigern]

Say I = 50 for arguments sake.
50 x 50 x t (=0.1 look up time curve) = 2500 x 0.1 = 250
Square root of 250 = 15.8 purely as an example. However you want the pfc as I, what is it?

 

[pc1966]

The size of fuse I’m thinking of installing is a 50A type b (either RCBO or mcb with rcd main switch).

So are you proposing to change from a fused-switch feeding the sub-main to it coming off a MCB/RCBO in the main CU?

For your I2t for a MCB best value to take is the values from the OSG table B7 and for 40A and up to 6kA PFC the requirement on CPC size is 2.5mm

If, however, you are meaning a 50A fuse feeding the sub-main, then Table B3 has limiting values on Zs but note that here 2.5mm requires a lower Zs to be met as the I2t for 5s disconnection time is too much! You need to go to 4mm CPC size to be OK for 50A BS88-2 fuse and the full permitted 5s disconnection time for a sub-main.

 

[buzzlightyear]

What size of tails going to the new board?

 

[pc1966]

So are you proposing to change from a fused-switch feeding the sub-main to it coming off a MCB/RCBO in the main CU?

Reading the OP again it seems there is currently no protection for the T&E feeding the sub-board?

If so then your best plan is a fused-switch (if the cable is not at risk of damage/nail penetration) as then any MCB/RCBO on the sub-board has a sporting change of selectivity with the feed OCPD. If it comes of a MCB/RCBO in the main CU you won't get that. Of course, if it is less than 50mm from the surface, etc, then it should be on a RCD of some sort and a RCBO would tick the box.

However, if the sub-board is only for an EV changer then probably you don't care, and it could have been a junction box at the end of a 40A or so RCBO circuit (depending on the requirements of the EV changer).

 

[buzzlightyear]

4mm CPC size to be OK

the begging question what is the size of the charger.

 

[pc1966]

I guess it is important to add here this is the CPC size only for allowing safe operation of the OCPD.

If there is an issue of bonding extraneous conductive parts to a PME supply, etc, then the other IET tables need to be consulted to determine what is actually an acceptable minimum for the CPC.

 

[ETradesmen]

Hi guys,

At the moment there is a 16mm t and e straight off the incoming supply (henly block) which goes to a 12 way 17th edition consumer unit in a garage (Now converted into a living space).

There is currently no fuse protecting the 16mm apart from the 100A cartridge fuse.

I don’t need to know the load at the consumer unit end, I just want to make the sub main safe which is why I’m going to add a fuse making sure it’s the biggest / safest one.

My thoughts were to add a 63a double pole RCD controlling a 50A MCB, the part I’m stuck on is the CPC calculation,


Thanks, Michael
[automerge]1592858732[/automerge]

What size of tails going to the new board?

it'll be 25mm tails off the henly block to new separate CU controlling submain

 

[buzzlightyear]

You could use a 50 amp rcbo with a main isolator.

 

[ETradesmen]

You could use a 50 amp rcbo with a main isolator.

Ok, that’s good news.

any chance you can solve the adiabatic formula for future reference?

 

[pc1966]

any chance you can solve the adiabatic formula for future reference?

You need the I2t value, and while simple in principle, it is hard to get for many OCPD. The "accurate" approach is:

1. Measure Zs and compute PFC as Umin/Zs which is the 'I' in your fault I2t (usually, but not always for MCB/MCCB, the lowest PFC has biggest I2t due to the much longer disconnection time).
2. Look up the disconnection time 't' for your PFC. For a fuse this is usually simple as they go down to under 100ms in manufacturer's data sheets (even if the BS regs don't for "generic" fuses). For a MCB it is complicated as once you get in to the magnetic "instant" trip they just show less than 100ms.
3. Select the 'k' value for your CPC and insulation temperature, etc, from BS table 54.2-54/5, but often 115 as you have is fine.
4. Compute minimum CAS from S = sqrt(I*I*t) / k
5. Round up to the nearest standard CPC size.

In reality getting 't' for the I2t is hard and if you don't have measurements of PFC/Zs then safest is to look at the worst-case let-through of the type of OCPD. For fuses this is at long disconnection times and so you can look at the 'I' for 't' = 0.4s / 5s depending on the situation. For MCBs you either use the worst-case generic value from, say, the OSG table B7, or look up the manufacturers data (assuming they even provide it). For example, here is the let-through energy from the Hager catalogue for their commercial B-curve MCBs:

[automerge]1592861371[/automerge]
From the Hager graph, lets assume a bad design where you have a 63A MCB but your Zs only results in 200A PFC. We can simply look up 200A on the 63A curve and the let-through is about 210k A2s, so we get:

S = sqrt(210,000) / 115 = 3.98 => 4mm CAS needed.

If we had a PFC of 300A then we see the let-through would be 1k so:

S = sqrt(1000) / 115 = 0.275 => 1mm CAS minimum (but not sane here!)

So rather oddly increasing the fault current leads to a smaller CPC requirement due to the vastly quicker disconnection time. Finally, if we had the worst-case PFC the MCB is rated at of 10kA fault current then we see the let-through is approximately 70k so:

S = sqrt(70,000) / 115 = 2.301 => 2.5mm needed

Generally once you have disconnect times much longer than 5s the adiabatic assumption of no heat escaping is not so valid and eventually you have to size the CPC as for a live conductor carrying the fault current continuously (as smaller than the adiabatic formulae would predict), but long earth faults are bad design in the first place!

 

[Baddegg]

I was literally 3/4 of the at through typing that out when you posted @pc1966 ?......?

 

[pc1966]

Reading it again I realised I have typed CAS instead of CSA = cross sectional area.