Kang Zhang wrote:
>
> Hello all
>
> Two questions:
>
> 1. Is there any big difference in reliability prediction results between
> using MIL-HDBK-217F notice2 and using Bellcore? I just try MIL. Using
> Bellcore can get "better" MTBF or "worse"?
>
> 2. Except the above two, are there any other methods to predict the
> reliability of electronic components, such as FET, Bipolar?
>
> Any inputs are very welcome. Thanks in advance!
>
> Kang

For #1 ...
My opinion only, I've no experience with the MIL-217
(what I'm comparing Bellcore with MIL-217 is what a
 colleague I saw go through with 217), but I've been
working with Bellcore 332 for a couple of years now.

Bellcore will give you better "numbers".  The assumption
with Bellcore 332 is 50% electrical stress and uniform
temp stress on the part (usually 40 degrees C).  That's
a global assumption, i.e. all parts throughout the entire
system.  The guy I knew who worked extensively with 217
made individual electrical and temp stress measurements
for the MIL contracts he worked on.  Very tedious work.
I'm not sure if that's really necessary for 217, but
that's what the guy I knew did.

For #2 ...
Yes.  The reliability numbers the semiconductor
industry produces comes from using the Arrhenius
equation in conjunction with accelerated temp
testing of the chip. Emailed back and forth with
one of the reliability engineers over at Intel
about it a couple of years ago.  Since, they use
so many different types of material in the making
of a chip, the activation energy (which is needed
for the Arrhenius equation) is guessed at.  But
this is all used for the accelerated temp testing
they do for chips.  But the bottom line is that
they develop a lot of empirical data over time.
So they're pretty accurate.

Just an aside.  I was corrected by a vendor once rather
sternly for using MTBF for a part.  They were adamant
that "MTBF" was to be used for something that was to
be repaired such as a printed circuit board.  The correct
acronym for a part is MTTF or Mean Time To Failure. The
key words being "Between" versus "To".  You don't
"repair" a chip or a fuse.  It's from this type of
database and field repair reports the come into Bellcore
that the FITs (Failure In Time) data tables in the back
of 332 are generated.  I believe 332 edition 4 had
some corrections to some FITs values in edition 5
to make them more realistic.

Regards, Doug McKean

##############################################################
TechNet Mail List provided as a free service by IPC using LISTSERV 1.8c
##############################################################
To subscribe/unsubscribe, send a message to [log in to unmask] with following text in
the body:
To subscribe:   SUBSCRIBE TECHNET <your full name>
To unsubscribe:   SIGNOFF TECHNET
##############################################################
Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional
information.
If you need assistance - contact Keach Sasamori at [log in to unmask] or
847-509-9700 ext.5315
##############################################################