Hi Bob.

You wrote:

<Are there requirements or guidelines around what a power supply input
breaker should be rated for a given nominal input current assuming the
power cord is rated appropriately?>

I don't recall seeing any protection choosing rules in any standards,
though I stand to be corrected.

Circuit breakers, like fuses, should have their characteristics chosen
according to the circuit they are protecting.  I.e.:

If your protected circuit is a resistive heating load you would have no
'inrush' current, and therefore your circuit breaker would be chosen with
an 'instant to constant' ratio of say 2:1, where a 10A circuit breaker
would allow 20A for , say, 60mS.

If your circuit is a motor or a small transformer input power supply, then
your ratio would need to be something like 6:1

If your circuit is a large transformer or a tungsten lamp load, you may
need 10:1 or even 20:1.

The circuit breaker (CB) constant (continuous) rating should be chosen to
protect the wiring downstream of it, and the wiring upstream of it should
be sized to cope with the CB continuous rating regardless of the actual
current used.

There isn't much future is trying to protect downstream components with a
circuit breaker, with the notable exception of transformers, as they are
often caused to trip by the failure of the component being protected.  The
CB tripping being an indication that the component has failed!

And the range of CB trip values available gives limited scope for finding
one which will be exactly 125% of your actual demand current.

A rule of thumb I use is to measure the actual current, double it, and
choose the nearest CB/fuse value upwards,  choose the 'Instant' ratio
according to anticipated inrush, and make sure that the wires are big
enough.

You have to be happy that short circuits anywhere will pull enough current
to trip the CB, but not allow the CB to trip on the switch on event.  If
you do this the equipment will be both functional and safe.

That's another twopence worth.

Chris Dupres
Surrey, UK. 

Reply via email to