Although many analog fault simulation papers have been published in the last 20 years, no general-purpose analog fault simulator is presently available commercially. The two dominant reasons for this are that there is no industry-accepted parametric fault model, and the time to simulate all possible defects in an industry analog circuit is prohibitive (years!).
Lets have this discussion with the hope of having some good outcomes from it!
If the defect is permanent (called harddefects like capacitor or resistor being shortcircuited or is open), tests can be designed to know this by dc measurements. If there are problems with deviation of compnents from the ideal (desired) value, they can not be found.
With active components like opamp, OTA etc , it may be more complicated (e.g. bandwidth changed from 1MHz to 700KHz) , it is quite difficult to point out where is the problem.
In modern designs like filters, facilities like automatic tuning loops exist to bring the critical values to desired values if there is no defect in interconnections, and if all components are working but their values are ouside the range.
Yes, the testing of analog circuits is a very interesting issue. Guaranteed parameters by manufacturers should suffice for conventional electronics. It is assumed, of course - design, functional sample, prototype series, etc. Testing method ("sensitive sites") must already be part of the design - is elected "by nature" of the circuit! And in modern (smart) structures it may be some tests repeated "in operation" - and thus continuously evaluate the situation.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering