Hi Mike,
You wrote to me:
MB>MB>The principal benefit to gold plate is immunity from
MB>MB>corrosion.  This is essentially unimportant in networking,
MB>MB>where currents are negligible.  In audio, for example, you
MB>MB>might commonly pass tens of amps to a speaker, and
MB>MB>corrosion would form a high-resistance spot which would
MB>MB>heat up.
MB> GW> From my experience in designing high reliability
MB> GW> engine/transmission control systems for military
MB> GW> applications (everything from milivolt to several amps on
MB> GW> the connectors), the advantage of gold is indeed immunity to
MB> GW> corrosion, but only during storage! Once a properly designed
MB> GW> connector has been mated the existance of the gold plate
MB> GW> becomes irrelevant.
MB>Keep in mind that we are not necessarily talking about something that 
weighs
MB>only a few grams and which is protected by a plastic shell,
MB>which is the usual model in automotive electronics.
In a military application it is usually a circular metal shelled
connector with a significant weight. They bear no relationship with the
connectors we used for commercial automotive applications.
MB>               In a
MB>BNC connector used for networking, the parts are often
MB>press-fit and can wobble or rotate freely, often each time
MB>the computer or cable is moved or bumped.  There's no
MB>strain relief on BNC, and even the bayonet locking pins
MB>comprise part of the live electrical connection! The
MB>practical effect is that the connector is always exposed to
MB>corrosion as if it was still in storage.
I agree with this, and they are the real weaknesses of the design. The
RJ45 connector used in 10BaseT networks is less problematical because
although the strain relief of the connector is not particularly good
the fact that all the circuits go through the contacts, which distort
when mated to give a good electrical connection, means that there are
relatively far fewer problems in the field than with the BNC connectors
of 10Base2 networks.
MB> GW> You have the benefit of gold against current the wrong way
MB> GW> round! The higher the current, the less relevant the gold
MB> GW> plate becomes (when a high power connector is mated the two
MB> GW> mateing parts should distort and cut through the gold plate
MB> GW> with the working contact being between the underlying
MB> GW> connector materials)
MB>Well, yes and no.  Yes, an impedance discontinuity can do more damage to a
MB>milliamp signal in terms of noise.  No, it is usually a
MB>much bigger problem if you pass tens of amps through an
MB>impedance discontinuity -- because you might start a fire.
I agree on both counts, but in my experience high power connectors are
better designed/manufactured than low power ones, and the pins/sockets
are designed to distort on connection so that a good contact is made.
The distortion normally results in the gold film being broken as it is
usually only very thin.
Off topic aside: In fact the commonest design of high power connector is
also about the poorest for electrical contact :-( It's the standard
mains connector (despite the differences between countries nearly all
of them suffer similar problems)!
MB>The gold plate is not cut away on BNC (or most other)
MB>systems.  Gold is an excellent conductor, and there is no
MB>reason to let it be cut away.
They are low power systems (BNC should be, I shudder at the thought of
using BNC connectors with significant power going through them) and the
cost cutting in the production does indeed make the gold essential,
especially on the centre (or, for inhabitants of the USA, center)
contact and indeed it is usually unaffected by mateing BNC connectors.
George
 * SLMR 2.1a * KEYBOARD - Used to enter errors int a computer.
--- Maximus/2 3.01
---------------