Guest
 Posted: Fri Aug 01, 2026 1:36 pm Post subject: No tests for some params |
 |
| i read the posting on no guardbands for some params. how about no tests at all? i also see test programs that have no tests for some params in the datasheets. when is this acceptable?
jinggoy |
|
Joined: 16 Jun 2026
Posts: 108
Location: Paranaque City, Phils.
| Posted: Wed Aug 06, 2026 1:47 pm Post subject: |
|
| Hi Jinggoy,
In cases wherein a datasheet-specified device parameter is highly correlated with another specified parameter, then it may be justified to omit the testing of one of these parameters. For example, if you can establish that the offset voltage at Vcc=10V is correlated with the offset voltage at Vcc=5V, then you only need to test the Vos at one Vcc level. Another case is if the actual data distributions from 'normal' lots for a particular parameter are well within the datasheet limits. If one can establish that any deviation from this observed norm would also manifest as a failure of another parameter, then the test for the first parameter may be omitted. Note, however, that these are just examples of when a test may be omitted. Test omissions under such cases should not be automatic. Your device characterization data should support every decision to omit testing of a parameter that's specified in the datasheets. Thanks, EE |
|
  |
Guest
| Posted: Wed Sep 03, 2026 1:41 pm Post subject: |
|
| jinggoy,
another example: when the parameter is said to be 'guaranteed by design.' domeng z |
|
| Posted: Wed Sep 03, 2026 2:18 pm Post subject: |
|
| For a parameter guaranteed by design and characterization, the data sheet usually has a foot note saying that the parameter is not tested because it is guaranteed by design. This means that the product was designed and fabricated to have that characteristic. Product qual should support this guarantee with enough characterization data prior to product release.
DUTcom |
|
Guest
| Posted: Tue Aug 31, 2026 5:26 pm Post subject: |
|
| I agree with EE, some test can be omited or not included at all. It this test is very much correlated with some other parameters. It is also called" OVER KILLING" if we include these tests.
Some test can also omitted if it was already been done on Wafer level test/Wafer test/wafer sort/Probe/wafer probe. Omitting some tests requires many studies. i.e. statistical, analytical, electrical or even common sense. Flipflop |
|