Protection When Changing CSA

I have always went by the rule " when changing (lowering) csa of cable/wire then insert protection device",

The regs have traditionally worded it something like that - although I prefer to think about it the other way around - i.e. insert a protective device any any point where the cable would no longer be protected by the upstream protective device. These days it's more about the cable's current carrying capacity rather than its c.s.a. - as installations methods (esp. thermal insulation or ambient temperature) can make a significant difference.

 

and further to that the second rule was "if its only a small run perhaps <2M then first rule need not apply".

Regs say 3m rather than 2m - but it's not a blanket exception - you also need to ensure that the chances of a short circuit are minimised, the consequences of any short circuit are acceptable and overload protection is provided (if downstream).

This particular application in design phase has an incoming 95mm cable in through a a switch disconnector which requires distribution to many smaller branch circuits. usually this application for me in control panels is simple as small currents (<30A) are in use but this larger requirement means the terminations get tricky having many different wire sizes.

Sounds like the sort of problem the distribution board was invented for.


   - Andy.

Agree, the simple answer is a fuse board or MCB panel.  If the proposed switch disconnector is nearby then this need not incorporate a main switch.

If the switch disconnector is remote then you will need a panel that has a main switch.


Alternatively if this equipment is to be installed within a large control panel, then you probably dont need a dedicated enclosure as the control panel Is the enclosure. Connect the incoming cable to large DIN rail mounted terminal blocks. From these terminal blocks run suitable sized wire to each DIN rail mounted MCB or bank of fuses.

The short conductors between the terminal blocks and the fuses or MCBs can be rated according to the load current.

Yes of course, sorry I have not explained that very well, this is within one control panel, therefore no DB board within the panel.


The DB board for the project splits the sizes into many different control panels, and within these control panels these would be various motor sizes requiring the different wiring sizes.

Ideally the cable needs to be projected against 2 distinct cases 

1) an overload  that over heats the thin cable but not the fat one supplying it - that protection may be at the load end of the thin cable, perhaps the motor oveload relay.or outbound breaker of some sort.

2) damage to the thin cable should operate a protective device (fuse/breaker MCCB etc)  to remove power to it, and perhaps the whole lot, without failing spectacularly and showering the user or the inside of the box with hot metal.

This depends on the let-through energy of the next MCB or fuse upstream, which may be the thing covering the 90mm incoming cable, and rated at 400A or whatever. Here the adiabatic approach is needed to tell you if you need a local fuse down or not.

You may decide not to do that 2nd one if the link is short (3m is the normal distance), and the chance of such damage is improbable / contained . (cable well supported and enclosed)

The classic must be the wire to the light switch in the substation - it is not normally done in 185mm2  in the hop from the busbars to the 5A fuse holder ?

Mike.

Hi 


Thanks Mike, great comments:

- Yes the thin cable running for about 1.5m downstream will then have a motor circuit protection (overload and short circuit).

- I had to look Adiabatic up. ?

The cost of 1.5 M of cable is trivial and I would therefore be inclined to size the smaller cable as the maximum that readily fits the terminals.


Example, 95 mm cable into DIN rail terminal block or bus bar, then a smaller cable to a 20 amp fuse. Whilst 2.5mm should suffice for 20 amps, but if the terminals will accept 6mm I would use that rather than 2.5mm.

This increases the chances of the smaller cable surviving a short circuit.


Or taking the example of the light in the substation, indeed one would not use 185mm into the 5 amp fuse holder. I might use 16mm though if the fuse holder will take that size.

therefore no DB board within the panel.

Depending on the size, you might be able to take a similar approach as a DB - e.g. DIN rail mounted devices, bus-bars and distribution (commoning) terminals, if within your own enclosure.

    - Andy.

You really do need to know the detail of the incoming protection to allow you to make an informed decision, especially if slow devices are used.

Sometimes a set of 3  "death or glory" fuses at the input even at some rating way above normal operation (800 or 1000A perhaps!!) limits the let-through energy in a way that allows the downstream hardware to be safely  sized much more economically than the 'do not know' case, by setting an absolute upper limit on the i²t.


As an example of what I mean, consider this  fuse

and this explanation of how to use the let-through graphs.

Then stare at this curve where I have highighted the 800A curve,





If we use this fuse fed from a supply of  anything up to about 15kA RMS PSSC, the peak current on the load side could be as high as 2.8 times that, if the fault comes on at the wrong moment in the cycle. But on a higher pssc supply, he curve kinks as the fuse energy limits, so even by the edge of the paper and 100kA supplu side PSSC, the downstream is protected to a level more like twice the peak fault current of the 15kA case, not 8 times.