I blog for a leading manufacturer of circuit protection devices, so I keep my eye out for articles on fuses, ESD diodes and the like. Recently, I came across the article, “Fuse selection factors critical to circuit design.” Among the factors discussed is:
11. Application testing/verification prior to production.
Request samples for testing in the actual circuit to verify the selection. Before evaluating the samples, make sure the fuse is properly mounted with good electrical connections, using adequately sized wires or traces. The testing should include life tests under normal conditions and overload tests under fault conditions to ensure that the fuse will operate properly in the circuit.
Being a former test engineer, this got me to thinking about how one would actually do this. For example, how would you test that a fuse will actually protect a circuit board? Would you inject faults, i.e. deliberately short-circuit nodes? If so, which ones?
What measurements would you make to ensure that the fuse was working as you hope? Would you measure the time elapsed between the time you injected the fault until the time the fused actually blew? How about measuring the current profile over that period of time? That might be important and/or tell you something about the failure.
What kind of fault analysis should you perform after the fuse has blown? I suppose at the very least you’d want to replace the fuse and ensure that the circuit is functioning again. I would say that you should also run a full performance test to ensure that the fault didn’t adversely affect the board’s performance. Also, I’d think that you’d want to visually inspect the board to ensure that the fault current didn’t damage the board or traces at all.
I’m curious if any of you have had any experience with this kind of development testing. If so, please e-mail me or comment here.