1
\$\begingroup\$

I'm considering a redesign involving moving from maintenance-unfriendly mezzanine connectors for daughter cards to a real backplane with card edge connectors.

I like how easy and physically robust card edge connectors are. I'm a bit worried about reliability, though.

My environment should be generally sealed, but with high temperature cycling. (inside a NEMA 4 enclosure, but located anywhere from hot deserts to frozen tundras)

I've read that a thicker gold plating on the gold fingers isn't that much more expensive and adds considerable life. Is that good enough? These units need to live in the field for a long time. (10+ years)

How about upgrades? Can I expect to be able to leave a unit in the field for years and then install an expansion card in a slot that's been unused that long? How about when replacing a card with another one?

\$\endgroup\$
5
\$\begingroup\$

Card edge connectors can be ordered ruggedized, and with several platings of choice, check for instance Sullinscorp

There are ISA slots with a cap for protection.

Some vendors are quite willing to help you with selection for parts, if you describe your problem to them.

You may want to insert an empty daughterboard into the connectors when they are not in use, if you cannot use a cap.

Make your design so you can easily exchange the backplane too.

Usually, gold plated contacts last a very long time. When they didnt, we used a piece of cardboard with a very fine sandpaper (1200 or higher) laid over one edge, and inserted into the slot once. But only as a last resort, back when industrial PCs where still expensive. Sadly, I found no documentation about the connector lifetime.

Also, for extended outdoor use, you might consider a fully sealed enclosure. Filters tend to clog, nasty ants will move in, or even the dreaded micropenguin might build a nest inside the enclosure, if the climate is cold enough.

\$\endgroup\$
  • \$\begingroup\$ They will be fully sealed, except when they're not. :) So far, they're all NEMA 4 enclosed units. They can pull the system out of my enclosure and install it in their own enclosure, but although people have talked about that I don't think anyone's done it yet. \$\endgroup\$ – darron Feb 20 '12 at 18:51
  • \$\begingroup\$ That reminds me of a customers experience with lowtech parts in equipment in the field: Once something was lowtech enough to be used in the field, the field would use it. Like, locals using batteries etc. Hightech or unuseable parts would be safe, though. \$\endgroup\$ – posipiet Feb 20 '12 at 18:57
2
\$\begingroup\$

A good reference is Connector Design/Materials and Connector Reliability.

According to one study:

  1. Gold-gold is the best performing pair. Insist on mating same-to-same. That means gold-to-gold, or tin-to-tin. Tin-gold combination performs worse than tin-tin.

  2. An extra mating cycle after a period of environmental exposure can turn well-performing contacts into not-so-well-performing ones. This causes grief since "benign" service work can turn a reliable system into a mess.

If you ever anticipate disconnection while the current is still flowing, gold-gold is subject to self-destruction for currents above 50mA or so. Cycled power contacts must be silver-silver. Cycled signal contacts must be gold-gold.

Vibration environment is also of consideration. You want minimal cyclic stresses on the mated pair - nothing will kill connectors faster than repetitive micromotion, or even just simple high contact stress cycling. This requires serious mechanical design by someone who actually knows mechanical design. You must verify relative displacements of both ends of a pair on a test article, exposed to vibration spectrum similar to what's in the field.

\$\endgroup\$
  • \$\begingroup\$ How does silver compare with tin? I've only heard of silver contacts being used for electro-pneumatic pipe organs. \$\endgroup\$ – supercat Feb 12 '15 at 19:02
  • \$\begingroup\$ @supercat Any high-performance power contact pair will be silver-silver. Say, any power pins from Harting are silver-plated. Tin-tin is OK if you want it cheap. It doesn't belong in any sort of professional gear, or products that are intended to be not necessarily disposable. \$\endgroup\$ – Reinstate Monica Feb 12 '15 at 19:03
  • \$\begingroup\$ Why is silver singled out only for high-current applications? If it's better than tin for high-current applications, would it not also work well at lower-current ones, or do higher currents remove silver oxide in a way lower currents would not? \$\endgroup\$ – supercat Feb 12 '15 at 19:08
  • \$\begingroup\$ @supercat Silver-silver is an open circuit in low-current circuits. It may not be so when you assemble the product, but it will be by the time your customer starts to care :) That it's better than tin-tin is indicated by major connector vendors either not offering tin-tin for power use, or deprecating it over silver-silver. \$\endgroup\$ – Reinstate Monica Feb 12 '15 at 19:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.