I was sitting in the tech area chatting to one of the guys one day when a bunch of boards came back on RMA. I thought it odd that a whole pile came back from one customer. So I asked him what's the story. Here is what he told me.
Somewhere between 3/4 and 9/10 of all products returned on RMA as defective are tested and nothing is found wrong with them. My first hunch was this might be caused by differences in the way a board manufacturer tests a board and the way a customer uses the board. Boy was I in for a shock. First of all, some basics. Boards are typically assembled by an assembly shop. They usually run some basic static tests (bed of nails tests) to ensure that all the components have been installed on the board and that most of the static components (resistors, diodes) are the right values and installed the right way. They then ship to the test shop. We were a small company at the time and did our own in-house testing so they came to us. The board was first run through the test suite. Those that pass are placed in a burn-in chamber for a couple of days. They are then retested. Those that pass are put in finished goods to go to customers.
The batch of boards that came in that day were from our number one customer. They bought boards by the thousand. I was told that their RMA rate was close to 8%. The techs take these boards, retest them, and most are diagnosed as NPF. These boards are tagged as "replacement boards" and put back in finished goods. They can't be sold as new, but can be shipped as waranty returns. All boards are tracked and if they are returned more than 2 or 3 times, they are given to us poor saps in software development where we never have a stitch of trouble with them. Anyway. Let's follow the boards. A batch of boards came in on RMA as DOA. We count them and ship an equal number from the "replacement stock". We test the boards and most go into "replacement stock". The big customer orders another batch of boards. Returns 8% of them, and guess what. He gets the boards he returned last time back as replacements. And guess what? He returns 8% of those.
NPF RMA's are a fact of life. They do not necessarily reflect badly on the person who couldn't get the board to work. They also have little reflection on the quality of the manufacturer. One thing everyone has to realize is when you return a product to a manufacturer, odds are VERY HIGH that the product you get back in replacement was returned to the manufacturer by someone else.
Now, why do I harp on people to make sure they do everything right before returning a product? Well, I will be honest. I am beling selfish. I realize that it costs the manufacturer money to test these NPF boards. It costs the manufacturer money to ship NPF boards back to the customer who repetedly can not get a board to work. The manufacturer then turns around and passes this cost along to ... me ... a purchaser.
So, when I write a help page with detailed instructions how to self test a board. I am doing so for selfish reasons. I don't want to pay for your mistakes. Far too often we throw everything togather, make one very simple, tiny, mistake that we miss during our inspection. This one simple, tiny mistake prevents the product from working. A second set of eyes often find these mistakes in no time at all. More often than not the process of explaining what you did will reveal the problem to you. Other times doing the same test a slightly different way will indicate the mistake. I hope that those of you who follow my instructions find your mistake in the process and therefore save yourself time and frustration having to return a product over and over, and save me money when I want to buy a product.