I used to make aircraft instrumentation for Boeing, Lockheed Martin, British Airways to name a few..
I had the exact same issue on 14 of our product lines with a MILSPEC-CECC 0805 Type part Manufacturer was Rohm.
Nearly every single board failed due to these devices (>65%), if one failed on the board i replaced all of them, and re-tested and thermally cycled them inline with our spec (-55/+80 Deg/C for 48hrs) this was a nightmare, especially when some cracks were not detected until several weeks later while flying installed on a jet over the Atlantic! Not good for the company or the pilots underwear.
I had to make this good in quick time, i created this project;-
1 Tested components from the reel - destructive, themal shock and cross-sectional analysis from a lab and produced a report, the sample showed that most of the components were already cracked before they began the testing!
2 Obtained new reel and different manufacturers batch, trialed them again as above - failed for the same reason.
3 Called Rohm Technical, was told its our processes! i told him to get a life and be onsite at the factory the next day or else.
4 He turned up, i presented the data and he took all but 60 secs for him to realise it was the parts, so i sent him packing with the info and the bogus reels for thier own analysis.
5 I got a report back from them (which is available off thier website, if you read it it blames everything but the component!) and he stated that they tested a sample from the reel and found no problems.
6 Could not change the part in the design due to incompability, banned the use of the parts for any new design and set our design engineers on the case.
7 To this day any part that is spec'd to be military or screened, i immediately ban from production - engineers are sometimes forced to us these parts, but tell them the issues from a manufacturing viewpoint and how much time THEY will spend hand placing devices and you might get them to select some other package.
Good Luck! Simon
reply »