FFT: Q1. Problem with Multiple path sensitization is that we cannot find a specific path that causes faulty output; rather, we will have multiple paths. correct me, I am confused
At 10:10 In the normal mode, there is only capture function and hence whatever the fault is stuck at, the flip flop is going to capture only that value. Hence it is difficult to detect stuck at fault in normal mode.
In 13:30 is it possible to test with 16 patterns in total (instead of 20 , we can parallel lines 17-20 to 13-16 , because those segment are independent)?
This is by far the best DFT/ATPG lesson in RU-vid. I hope you keep these videos in RU-vid for all the future IC engineers to watch. Thank you very much.
In normal mode, the input signal is an absolute means designer give standard combination of inputs and expects the output but never assumes one of the input stuck to either 0 or 1. But this stuck part happens during manufacturing...so it can be detected in test mode only. I hope this is the correct answer
Q2 27:54 For XOR gate There is no equivalent faults , EFC is cannot be done (10),(11), 0:00 0:00 (01)/(00) these three test vectors are essential test vectors of XOR gate So, DFC also cannot be done 27:54
Thank you professor for a great lecture. It makes clear what the scan test is than any other articles or videos. Does anyone know the answer for the question? Why can't stuck at zero fault be testable in normal mode?
On the last question, the FF2 will be stuck at 0 since after reset, output of FF2 will be cleared. This output then gets fed as an input to the only OR gate with a SA0 issue. As a result, FF2 output will never change; therefore, prevent the circuit from normal operation.