Gunther Schadow <gunther@aurora.regenstrief.org> wrote:

>What I would probably do is use thicker red wires. I don't
>really understand why the shorter length should be a problem
>in itself. I don't even trust each wire actually carrying
>current because of the way the concats are simply crimped
>over the insulation.

Resistors in parallel. If you have two resistors in parallel and one of them
has slighly less resistance, it will carry more current than the other one.

If you have 6 wires all capable of safely carrying 5A, and you assume that
together they can safely pass 30A, but one of them is shorter, then that one
will take more than 1/6 of the current, exceeding it's specification.

The resistance will differ by the ratio of the lengths, just how much shorter
is the short wire? You would think that there would some overengineering?

If we assume that there is one short wire and the others are the same length
(while probably not true, the math is easier) then we have

r1 = c * r2 where c is the ratio of the wire lengths.

Since the voltage across the wires is the same:

p1 = v^2 / (c * r2)
p2 = v^2 / r2

p1 = p2 / c

The ratio of power lost in each wire is the inverse of the wire length ratio.

With a simple linear relationship between increased power loss and wire length
in the bundle, surely they weren't that keen to save copper? They can't have
forseen the current draw of the add on cards, surely?

--
Tom Parker - tom@carrott.org
           - http://www.carrott.org