cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Newbie
Newbie
712 Views
Registered: ‎06-06-2019

HTOL and device reliabilty

Jump to solution

We are looking about reliability performance on XC7A100T-1FGG484I FPGA.

In last UG116 document, HTOL result is 11 FIT for 7 series FPGAs and Zynq-7000 SoCs.

Does this result corresponds to the die ? or does it correspond to the whole component (die and package) ?

In second case, is there reliability difference between these two serie 7 FPGAs: XC7A100T-1CSG324I and XC7A100T-1FGG484I ?

Thanks

 

 

 

0 Kudos
Reply
1 Solution

Accepted Solutions
702 Views
Registered: ‎09-17-2018

np,

ug116 device hard failure  FIT is the result of many processes, predictions, models, procedures, tests.  It is the device (die and package and environment - recommended voltages, temperature less than 85C with > a few thousand on/off power cycles from cold to hot).

Actual field failure data may be requested (under NDA), so you can find out what the actual FIT is, as opposed to the predicted FIT.  Actual FIT is interesting as it includes all the devices not being stressed past their limits, and those that are.  Predicted failure mechanisms and models are also available under NDA (service request).

The devices generally exceed a 15 year life for 85C C grade, 10 year life for 100C I grade, and 3 year life for 125C M grade.  Longer life is the result of less operation at the maximum junction temperature.  So, M grade below 85C is > 15 years...

What I always found remarkable was just how long they do last, and how infrequently they fail.  The only failures I have ever seen were test escapes (bad to begin with), power supply failure causing over-voltage, ESD damage (improper handling), and, lightning strikes (which nothing is able to survive).  In that order.  So if you have a good manufacturing flow and test, you weed out any test escapes, leaving the remaining factors.

l.e.o.

 

View solution in original post

Tags (2)
1 Reply
703 Views
Registered: ‎09-17-2018

np,

ug116 device hard failure  FIT is the result of many processes, predictions, models, procedures, tests.  It is the device (die and package and environment - recommended voltages, temperature less than 85C with > a few thousand on/off power cycles from cold to hot).

Actual field failure data may be requested (under NDA), so you can find out what the actual FIT is, as opposed to the predicted FIT.  Actual FIT is interesting as it includes all the devices not being stressed past their limits, and those that are.  Predicted failure mechanisms and models are also available under NDA (service request).

The devices generally exceed a 15 year life for 85C C grade, 10 year life for 100C I grade, and 3 year life for 125C M grade.  Longer life is the result of less operation at the maximum junction temperature.  So, M grade below 85C is > 15 years...

What I always found remarkable was just how long they do last, and how infrequently they fail.  The only failures I have ever seen were test escapes (bad to begin with), power supply failure causing over-voltage, ESD damage (improper handling), and, lightning strikes (which nothing is able to survive).  In that order.  So if you have a good manufacturing flow and test, you weed out any test escapes, leaving the remaining factors.

l.e.o.

 

View solution in original post

Tags (2)