IC socket reliability

[George Farris]

Greetings,

Lee just made a point about IC sockets. What is the groups experience
about using IC sockets with gold contacts as opposed to not gold. All
my sockets were not using gold contacts and so far seem to work quite
well but...

Boards I have that are 30 years old and don't use gold contacts seem to
work well but...

Thoughts?

George

[Peter Shkabara]

I think that Lee was referring to the QUALITY of the socket rather than
suggesting gold. In my "former life" as an electronics engineer, it was
understood that gold is pretty buy tin could actually be better. The problem
is that cheap sockets may tarnish and not have a good contact surface. With
aging, IC pins in a tin socket could corrode where gold would be resistant
to corrosion. Upon insertion of the IC (or moving it) in a tin socket would
scrape the tin and produce a reliable contact.

Bottom line - if you don't have a socket, you will not have a contact
problem. Otherwise, a good quality socket whether it is gold or tin should
be almost as reliable.

Peter Shkabara
kol...@gmail.com

-----Original Message-----

Lee just made a point about IC sockets. What is the groups experience about
using IC sockets with gold contacts as opposed to not gold. All my sockets
were not using gold contacts and so far seem to work quite well but...

George

[Lee Hart]

On 10/17/2012 11:52 AM, George Farris wrote:
> Lee just made a point about IC sockets. What is the groups experience
> about using IC sockets with gold contacts as opposed to not gold.

The advantage of gold is that it does not corrode. It remains a good
electrical contact "forever". Gold plating will not "fix" a crappy
connector; but if used with a good quality connector it will
considerably improve the long-term reliability.

But it is expensive, so almost all connectors use some cheaper metal
(usually tin). Tin is initially good, but it slowly oxidizes over time,
due to the oxygen, moisture, and air pollution (mainly sulfur
compounds). The worse the air quality (damp basement, big city air
pollution, etc.) the faster it corrodes.

Fresh tin is bright and shiny. Tin oxides are dull gray, and tin
sulfides are dark; even black. They are lousy conductors. As the tin
corrodes, you get intermittent connections.

Tin's one virtue (besides being cheap) is that its oxides and sulfides
are mechanically weak -- they can be "rubbed off" by wiggling
connectors, or removing and replacing ICs in their sockets. But that
assumes the connectors are good enough to survive plugging/unplugging.
The ones in the H9 were bad in this regard, too.

Note: Real gold plating needs to be think enough so it's not porous.
This means 30 to 100 microinches of gold. A lot of "gold plated" parts
just have a gold flash perhaps 5-10 microinches thick. This is just for
appearance. It is so thin that there are numerous "holes" that expose
the underlying metal, which corrodes as usual and you still get bad
connections after a while.
--
Never doubt that the work of a small group of thoughtful, committed
citizens can change the world. Indeed, it's the only thing that ever
has! -- Margaret Mead
--
Lee A. Hart http://www.sunrise-ev.com/LeesEVs leea...@earthlink.net

[Chris Elmquist]

My experience (with a fair number of vintage platforms from mid-70's
and on) is that the environment they were used in (or left to rot in)
has more to do with their reliability than the sockets.

Damp, humid basements or sheds and barns with extreme temperature
deviations take the greatest toll on edge connectors and sockets.
Rodent and insect infestation is next and if you take the two together,
there's a good chance the thing is just totally shot even if it was full
of gold.

I have some of my own machines, that were built in '74, '75, '76-- full
of tin plated sockets and connector pins and they all still work great
with nothing whacky or intermittant going on. But, I've kept those
machines in dry, cool, "human compatible" conditions for nearly 40 yrs.

cje

--
Chris Elmquist

[Lee Hart]

On 10/17/2012 1:17 PM, Peter Shkabara wrote:
> I think that Lee was referring to the QUALITY of the socket rather than
> suggesting gold.

I agree, Peter. The quality of a socket or connector has more influence
on its reliability than whether it is gold plated or not.

Tin can be used to make a high reliability connection if it is a HIGH
PRESSURE contact. The classic 0.25" quick-connect type blade connector
is an example of this. With the good ones, you need pliers to get it on
or off. The very high contact pressure squeezes the metal so tightly
together that it blocks air and moisture from reaching the actual point
of contact. If you pull off an old cruddy looking quick-connect, you'll
find bright shiny metal at the actual points of contact.

High pressure IC sockets are rare, but they do work. The force needed to
plug in or remove the IC is so large that you are likely to crumple or
bend a pin. There is usually a special tool needed to support the pins
and hold them in exactly the right position to insert the IC.

The best IC sockets in my experience have been the screw machined type
(Augat, etc.). Each pin has a round tube with a cylindrical spring
inside it that contacts the IC pin on 4 sides. The insertion and removal
force is high; but not so high that you can't deal with it. The spring
has 4 redundant contact points, so even if 1-3 of them get bad, you
*still* make contact with the pin.

The Heath H9 sockets were just about the worst I've ever seen. They were
tin, low-pressure, and only had a single point of contact. The plastic
body would cold-flow over time, relieving what little contact pressure
it had in the beginning.

> Bottom line - if you don't have a socket, you will not have a contact
> problem. Otherwise, a good quality socket whether it is gold or tin should
> be almost as reliable.

That is always good advice. A part that's not there can't fail! If you
don't NEED a socket, don't use one. ICs are almost always more reliable
than their socket!

--
First they ignore you; then they mock you; then they fight you; then you
win. -- Mahatma Gandhi

[Rob Doyle]

On 10/17/2012 12:21 PM, Lee Hart wrote:
> On 10/17/2012 1:17 PM, Peter Shkabara wrote:
>> I think that Lee was referring to the QUALITY of the socket rather than
>> suggesting gold.
>
> I agree, Peter. The quality of a socket or connector has more influence
> on its reliability than whether it is gold plated or not.
>

I was taught that gold-on-gold (i.e., gold socket and gold pins) was
a very good solution. Gold-on-tin (i.e., gold socket and tin pins)
was worse than tin-on-tin (i.e., tin socket and tin pins) -
assuming constant connection quality.

Dissimilar metals in a socket is bad.

Rob.

[Lee Hart]

On 10/18/2012 2:14 AM, Rob Doyle wrote:
> I was taught that gold-on-gold (i.e., gold socket and gold pins) was
> a very good solution. Gold-on-tin (i.e., gold socket and tin pins)
> was worse than tin-on-tin (i.e., tin socket and tin pins) -
> assuming constant connection quality.
>
> Dissimilar metals in a socket is bad.

That's true, but the severity depends on the circumstances.

Several things happen when two different metals are in contact. First,
they form corrosion couple. If it gets wet, the more noble metal (more
positive) is protected, and the less noble metal corrodes. In the case
of gold-tin, gold is more noble than anything, so the tin loses and
corrodes.

Thermoelectric effects are another problem when two different metals are
in contact. A temperature difference creates a small voltage difference
between the two. This doesn't matter in digital circuits, but it can
play hob with precision analog circuits that are trying to measure
millivolts.

Neither of these problems should bother an H8. The connectors should
never get wet, and none of them are depended on for millivolt accuracy.
So in the long run, gold on tin is still better than tin on tin.
--
A designer knows he has achieved perfection not when there is
nothing left to add, but when there is nothing left to take away.
-- Antoine de Saint Exup�ry