Offseting header footprints

[slateraptor]

https://www.eevblog.com/forum/index.php?topic=5650.0

Second comment by joelby links to a Digilent CPLD breakout. The offset JTAG header holes immediately caught my attention.

First Google search resulted in: http://www.sparkfun.com/tutorials/114

Based on Sparkfun's board house tolerances, 5 mil alternating shimmy results in a self-retaining fit. Not too sure how I feel about the author's solution. Admittedly, it's simple yet clever.

Just thought I'd share.

[ejeffrey]

The first place I ran into this solution was sparkfun's eagle library.  I wouldn't use it for a prototype board: there are plenty of ways to get the headers aligned good enough, especially if you don't care about cosmetics.  Without experimentation on a particular PCB house and design rules, I wouldn't count on the manufacturing tolerances being sufficiently predictable to get the offset right.  I would hate for a prototype board to be ruined because the drills were rounded from imperial to metric sizes.

For production where you are making lots of identical boards always manufactured by the same board house, this sounds like a great idea.

[slateraptor]

I wouldn't use it for a prototype board...Without experimentation on a particular PCB house and design rules, I wouldn't count on the manufacturing tolerances being sufficiently predictable to get the offset right.

Agreed, hence the "based on Sparkfun's board house tolerances" and "not sure how I feel about the author's solution" comments.

Soldering with the PCB laid upside down on perf board held by a panavise works well for vertical headers. For right-angle headers, a pair of mechanical fingers works well since they're mounted towards the edge of the board.

[joelby]

It also seems like an nice solution where you don't necessarily want to permanently solder the header onto the board. I do this with my PICkit programmer - just put a 1x6 header into the end. The staggered pins would make it less susceptible to just falling out.