The Wrong Inductors
The pilot assembly began last week. As is typical with a new design, the process started slowly. After setting up the pick and place machine and reflow oven, a small batch of five units was assembled by the automated process. These units were tested in order to validate the process before proceeding with the remaining units.
Unfortunately the first five did not test well. A few minor soldering problems were corrected, and they still didn't test well. The units functioned, but they had terrible RF performance. The engineer looked at the units and compared them to the hand-assembled samples I had sent, and he noticed that the inductors in the RF section looked different. Since that component selection was a fairly late design change on my part, he suspected that my sample units were built with a different inductor than the one I specified.
He tried replacing the inductors with a vaguely similar part he had lying around, and it improved the performance. It still wasn't as good as it should be, but dropping in a randomly selected part made things better. The production was stopped, and he shipped me two of the new units so that I could help diagnose the problem.
When they arrived, I looked at the inductors and immediately noticed that they were unfamiliar. The manufacturer even sent me spare inductors on tape cut from the reels purchased for the assembly, and they had already emailed me photographs of the reels. The labels in the photos looked correct, but the part was not correct. It didn't look anything like the genuine component that I have previously purchased from multiple sources.
Was this an intentional counterfeit? Was it an honest error? We'll probably never know. The phony part has the same inductance and is the same package size as the genuine part. It might even be made by the same manufacturer. My guess is that it costs about 5 cents, half of the cost of the genuine part. Is that enough of a profit to motivate intentional relabeling for an order of 20,000 units? I don't know.
Fortunately the distributor is taking quick action to correct the problem. They are using a known good source to supply replacement units for the pilot production as quickly as possible. I expect that we will be able to resume production next week.
The good news is that this problem allowed me to inspect and test a couple of early sample units from the factory. The PCBs are beautiful, and I am confident that the assembly will proceed more smoothly once the correct inductors arrive. I will make the shipments to the PROTOTYPE backers as quickly as possible, probably by mid-April.