Posted by Rivs: Wed Dec 12, 2007 10:32 am    Post subject: Cratering Issue

guys I need you help on what reliability test to perform on cratering affected units, thanks

Posted by FARel Engr: Thu Dec 13, 2007 11:17 pm    Post subject:

As a minimum you should do preconditioned TCT on an adequate number of samples. Aside from the usual readpoints, you must perform a thorough inspection of the Si under the bond pads after the precon, and after the TCT. Thus, you need to have enough samples to account for sample 'losses' after each 'crater check' since this procedure is destructive.

Posted by Paula: Fri Dec 14, 2007 5:23 pm    Post subject:

Sa tingin ko wala appropriate rel test for bondpad cratering. Usually, scrap na yung affected lot.

Try mo mag-crater test sa mga good units, medyo malaking sample size, pag meron cratering ibig sabihin- test can't weed out units with cratering. I don't see a reason bakit kailangan pa ng rel test.

My thoughts.

Posted by Rivs: Sat Dec 15, 2007 7:58 am    Post subject:

thanks guys.

yes paula even good units have cratering. reliability test purpose is to check if the device (good affected with cratering) will still pass.

Rejected parts failed leakage parameters werein cratering is not the main suspect, as experienced device affected with cratering fails breakdown voltage (shorted).

Also rejected parts have undergone microthermography (liquid crystal technique) and hotspot was seen on the gate runners while none seen on the good units but after crater test we found out that cratering is evident on both good and rejected parts.

Anyway thanks for the help guys I highly appreciated the responses [/b]

Posted by FARel Engr: Thu Dec 20, 2007 3:55 am    Post subject:

The rel tests that I earlier recommended is to confirm if other lots are affected by the same cratering issue (containment of affected lots) - not to determine if affected lots will survive in the field. I agree with Paula on the point that lots already known to be affected by cratering need to be scrapped.

However, lots suspected to be affected, but not yet confirmed to be affected, will benefit from rel testing, since it would also not be wise to just scrap lots on the mere suspicion of being affected, unless the process that caused the cratering has been identified and the suspected lots underwent the same process.

Posted by Rivs: Thu Dec 20, 2007 8:11 am    Post subject:

However, lots suspected to be affected, but not yet confirmed to be affected, will benefit from rel testing, since it would also not be wise to just scrap lots on the mere suspicion of being affected, unless the process that caused the cratering has been identified and the suspected lots underwent the same process.

Hi farel engr,

thank you very much for the response.