Anyone have done transition DFT testing?

[feel_on_on]

how much coverage have u gain with transition fault?

Do u need to generate 2 kinds of pattern ? Stuck pattern and transition pattern?

How about your test coverage on stuck pattern and transition pattern?

 

[dft_guy]

The stuck-at patterns and at-speed patterns are definitely different, and whether you run both is up to you. If you catch a fault with a transition delay pattern, it is caught regardless, so you could run the transition patterns, and then the stuck-at patterns to make up for the lack of coverage achieved with the transition patterns. However, you do lose diagnostic information that way. If a device fails transition test, is it because it's too slow? or is it actually stuck? No way to tell unless you run both sets complete.

John
DFT Digest

 

[rameshsuthapalli]

Hi,

we have to genrate the transhition faults frist since we can use TFT pattrens to detect S@faults. but we cannot use S@faults to detect transhition faults.

Regards,
Ramesh.S

 

[kiranks9]

Usually u will get the coverage 90% for transition DFT. here faults are detected only transition level i.e 0 to 1 or 1 or 0. After generating the Transition faults, then fault grade those faults, after words u can load these faults to Stuckt patterns.