I hate to argue a point with you Erik but you are a way off when you say that a properly designed board will have no failures. Plus the soldering is only a part of the issue.

You also have to consider ALL of the other failure mechanisms that can come into play during a product life cycle. And many of these factors can be so far removed from the board's initial design process that they cannot be factored into the design. Some factors can be intellectually removed from the list if you presume that the product is used correctly but then that is like telling people to never take a spill on a bicycle becasue it is hard on the tires and frame.

Think about thermal expansion and contraction that can crack surface mount resistors.

Think about mishandling by the equipment installer that bends the board which can break perfectly good solder joints and crack surface mount components.

Think about latent stresses put into a board because a reflow solder process may be off its profile by 5% because they started the line one morning before the line monitor dude got his cup of coffee. Latent stresses will crack things at any point in a product life cycle.

Think about the ever finer part geometries of the latest state of the art microcontrollers and other chips that happen to have a small hole in the SiO₂ just under an aluminum run on the chip. Voltage bias on the chip can and will lead to metal migration that can lead to part failure eventually. When....who knows...every chip is on its own piece of unique silicon.

Think about the connector that is attached to external cables and the equipment is found dangling by the cable. It can damage the connector, particularly now that nobody can afford to fill their back panel up with MIL STD flange mounted circular connectors that you used to be able to drive a tank over.

Think about the lightning that came into my house last June and took out 3 telephones, a monitor, 2 keyboards, my Ethernet router, 3 PCI NIC cards and two PC-Card NICs. Design in as much protection as you want and I can still show you a stronger lightning bolt.

Think about the vibration in a control board in a chassis that was designed for use on a plush Cadellac but has found application in a rural gravel truck. Long term conditions can eventually make things break.

So I suggest you revise your ideas and percentages and realize that products break for a multitude of reasons beyond crappy designs.

Michael Karas