John Fields said:
In the first place, even if the resistances of the two contacts were
different the difference would be in milliohms [...] In the second place, the
temperature coefficient of resistance of metals is generally positive,
so even if one set of contacts did conduct substantially more current
than the other one, and got hotter, [...].
[...] a blanket condemnation of
paralling contacts to increase the current switched isn't warranted.
I wouldn't say I gave him a blanket condemnation (though others may have); I
just said that in this instance, 0.5 + 0.5 doesn't equal 1.0. I'll stand by
that math
But John, I'm not sure you're attending to the right aspects. You're
talking about overheating during normal operation. I don't think that's the
main thing relays are current-rated for, is it? I mean, a piece of 22-gauge
wire will carry 3A comfortably, and I've seen relays with contacts and
internal wiring that are physically larger than that rated for less.
I am *NOT* an expert on relays, so read this with considerable skepticism,
but:
I think the current rating of a relay has more to do with the amount of
current that it can safely interrupt, without damage to the terminals from
arcing. That is, the rating is concerned with the situation where the relay
is closed, current is flowing, and then the relay opens. The two contacts
of a DPDT relay must open at slightly different times (perhaps 0.1msec
apart); so, if they're paralleled, one contact is always getting opened with
the full current on it. It is precisely that brief period of opening that
matters, because if there's too much current to interrupt, the contacts will
bond and fail to open.
More realistically, since we're only talking about a factor of two overload,
the contacts will still successfully open; but over time, the effect will be
exactly the same as if the relay were being run at twice its rating, because
it will always be the case that _at the most critical moment_, all the
current is flowing through one contact. Little micro-welds will form and be
physically pulled apart by the relay spring, but they'll be twice as big as
they're supposed to be.