|PCB, gold plating facts|
Message #2 Posted by Ellis Easley on 14 Aug 2002, 4:35 a.m.,
in response to message #1 by Michael Meyer
In the mid-1980's I had occasion to have some printed circuit boards made and I learned that the options available for gold plating on card edge connector fingers were 30 microinches, 15 microinches or "gold flash" which I think must be a contact plating, rather than electroplating. Also, gold isn't plated directly on copper, there is a layer of nickel plating in between.
One other thing - tin plating on connector surfaces (such as lower cost card edge connector fingers) is much harder than gold, and since tin develops an oxide coating like most metals (other than gold), to make a reliable connection it is necessary for the spring force pushing the two contact surfaces together to be quite high, so the oxide coating is scraped off during mating and a "gas tight", metal to metal connection can be made. Therefore, one of the worst things you can do with connectors is to mate a tin plated contact with a gold plated one - with the high spring force and greater hardness of the tin plated contact, the gold is scraped off the gold plated contact. Mating tin plated contacts together makes reliable connections, only the allowable number of mating cycles is much less than with gold plated contacts.
All PCB traces are copper, with or without other plating. The original way to make the traces was to start with a board covered with the full thickness of copper and print the desired pattern, in a positive image, with a coating resistant to the chemicals used to etch away the unwanted copper. The wider the traces, the longer you could leave the board in the etchant to make sure any stubborn bits of copper were dissolved, since there was some etching of the edges of the traces (between the resist and the substrate). As traces became narrower, the board had to be removed from the etchant as soon as possible, so copper shorts between traces became more of a problem. One solution was the "additive" process: starting with a board covered with a very thin layer of copper, the pattern was printed in a negative image and then additional copper was plated onto the board, making the exposed trace patterns thicker. Then the resist was removed and the board was etched until the original, unplated, thin copper between the traces was dissolved.
Many PCBs have their traces covered with solder to make it easier to solder components to their pads, but this step can cause shorts between traces. Also, solder shorts can be caused when the components are soldered to the board. "Solder mask" is the (usually green) paint which is applied to the majority of the surface of the board (except for the component pads) to prevent solder shorts between traces when the components are solderd in place, but it can lead to another problem: a board with solder plated on all the traces and then covered with solder mask can end up with solder shorts UNDER the solder mask, where they are harder to deal with. One solution to this problem is "SMOBC" - solder mask over bare copper. Sometimes with this kind of board the bare copper traces look black under the green solder mask (I imagine this is because the copper reflects mostly red light).
Some lighter colored solder masks with shiny copper or solder plated traces underneath make it look like the traces are gold plated when they aren't.