Ramon,
Good point. I know that AVX has made many major improvements. However, I
also know that most people run into the cracked cap problem through
improper pad design and improper processes. I have data to back this up,
as I do have experience from doing failure analysis work on this
specific problem for several companies. In 4 different incidents at
different companies, the issue was exactly the same. The root cause was
as Werner explained last week, large flat pads cause the endcaps
(termination caps) on the MLCCC to be anchored into the solder, flat to
the pwb pad, with little compliancy from the solder perspective.
Subsequent thermal expansions of the pwb cause the expansion/shrinkage
CTE forces and also from flexural stresses such as from de-paneling,
screwing down frames, heatsink attach, ICT fixture vacuum flex, etc. to
be placed directly onto the ceramic body of the MLCCC, either causing
premature failure to the endcap connection to the dielectric plate
inside of the capacitor body, or delayed failures caused by moisture
entering the cracks formed in the ceramic (sic) body. The larger and
flatter the capacitor, the more likely the problem will occur.
By chance, at one site the same exact part number was used at different
locations but had no failures, and this turned out to be due to a
different pad layout from the design library being used. The reason for
the two different pad layouts was to able to use either of two slightly
different capacitor sizes, dependent on availability.
In all cases but one, changing the pad to a smaller size that was only
.005" to .010" larger than the endcap termination itself, and going from
a .006" stencil to a .007" stencil completely eliminated the problem.
This put more solder under the endcap, providing a small amount of
compliancy. I owe this solution to earlier documentation of this problem
and the cure in Werner's earlier publications awhile back.
Thank you, Werner!
In the exceptional case, the issue was compounded by four other things;
the pick and place machine was set to place with 300 grams of force (to
make sure the "darn things were seated firmly into the paste so they
don't tombstone during reflow") and because the circuit board used an
OSP finish, and due to improper wave solder setup. The ICT test
technician would literally place both hands on the assembly and jump up
and down during the test in order to get good test pin contact on her
bed-of-nails ICT fixture through the OSP coating. Imagine the stresses
placed on the capacitors! In this case, the caps were also wave-soldered
on the bottom side of the board. Those failed more often than the
topside caps. The operators had been told that there were problems with
the capacitor solderability, so they were touching up all of the solder
joints.
The causes were collective; the thermal shock caused by insufficient
pre-heat temps, the rapid cooldown of the board as it exited the wave
(they had an extra fan mounted on the exit conveyor blowing directly
onto the board as it came over the wave), and the other issues combined
to muddy the waters, making it difficult to pinpoint the 4 causes. It
took almost a week to go through each process and identify the causes.
Once they were fixed, no more cracked caps were seen.
I wish to clarify that none of this happened here at GD.
-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Dehoyos, Ramon
Sent: Wednesday, June 14, 2006 8:08 AM
To: [log in to unmask]
Subject: Re: [TN] Finding cracked capacitors
It seems to me that it is time for the manufacturers of ceramic
caps to get back to the drawing board and redesign the caps so that this
problems is resolved by getting to the root cause. Some kind of strain
relief termination so that the cap is not susceptible to stresses. I can
think of more than one way to do that but the manufacturers have the
resources in place to make the changes and come out with the best
resolution.
Ramon
-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Camille Good
Sent: Tuesday, June 13, 2006 3:03 PM
To: [log in to unmask]
Subject: Re: [TN] Finding cracked capacitors
Using the hot soldering iron approach can cause a number of problems,
many of which have been described in previous posts. I've also seen a
case where a hot soldering iron caused the cracked capacitors to
self-heal (temporarily) - the heat drove out any water that had worked
its way into the cap due to a high-humidity operating environment, and
that in turn got rid of the leakage current problems we were having. At
the time, we didn't know that the issue was a cracked capacitor, so it
was maddening to troubleshoot the product. The product clearly was not
working correctly, but the instant you touched that area of the PCB with
a soldering iron (and there were a number of components all in parallel
with the capacitor) the problem went away and the product started
working fine.
-Camille
I saw a product a couple years ago that had problems due to cracked
capacitors. The problem only showed up in high humidity situations,
where the presence of high humidity would cause an increase in the
leakage current of the capacitor. This increased leakage current sucked
away enough power from other components that the part wouldn't work
correctly in certain conditions.
The worst part was that as soon as we touched that particular PCB
no
Dale Ritzen <[log in to unmask]> wrote:
Quite a lot has been done in regard to this in the design of
electronic assemblies. There are some pretty good white papers out on
part placement changes during board design that minimize this risk,
especially during depanelization. I used to have one saved aside after
an incident with a customer of ours. I'll see if I can find it and post
same.
The only solid way we were ever successful in finding these was with a
quick touch with a hot soldering iron. The cracked terminus usually pops
off the pad if it has cracked away from the body of the part. Yes, it
does expose the cap to another shot of high temp, but it does ensure
that you detect all cracked caps and get them replaced. Since caps with
cracked terminus typically don't fail consistently during ICT or
functional testing (as they can be in parallel or be making partial
contact to pogo pins which still see a device), the soldering iron
method was our only recourse. But, this is only practical if you have a
consistent problematic area of the board. In a case where the caps may
be located in many areas, trying the soldering iron technique on an
entire board is probably not very timely or cost effective.
Good luck with this. It's a tough one...
Dale Ritzen
Quality Manager
Austin Manufacturing Services
-----Original Message-----
From: TechNet [mailto:[log in to unmask]]On Behalf Of John Maxwell
Sent: Monday, June 05, 2006 10:46 AM
To: [log in to unmask]
Subject: Re: [TN] Finding cracked capacitors
Daniel,
Most flex cracks are hidden under the termination at the capacitor
termination/PWA interrface making it very difficult to detect using SLAM
or C-SAM. A quick way to uncover cracks is exposure to humidity & bias
accelerating failure. A few temp cycles ahead of humidity testing will
open the cracks a bit easing humidity entering the capacitor body
accelerating failure.
John Maxwell
>A good acoustic microscope should be able to find these cracks quickly
>and non-destructively. It is not a cheap tool, though, at $100k-$200k
depending
>on the options.
>
>On Mon, 5 Jun 2006 10:11:04 -0400, Tempea, Ioan
wrote:
>
>
>
>>Hi Technos,
>>
>>a never ending story, ceramic caps that crack, possibly due to
>>
>>
>depanelizing. The damage is not visible, unless the cap is cross
sectioned.
>
>
>>Is there any testing, thermal cycling or something like that, that
>>could
>>
>>
>show the damage real fast?
>
>
>>Thanks,
>>
>>Ioan
>>
>>
---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 1.8e To
unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet To temporarily halt or
(re-start) delivery of Technet send e-mail to
[log in to unmask]: SET Technet NOMAIL or (MAIL) To receive ONE mailing
per day of all the posts: send e-mail to
[log in to unmask]: SET Technet Digest
Search the archives of previous posts at:
http://listserv.ipc.org/archives Please visit IPC web site
http://www.ipc.org/contentpage.asp?Pageid=4.3.16
for additional information, or contact Keach Sasamori at [log in to unmask]
or 847-615-7100 ext.2815
-----------------------------------------------------
---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 1.8e To
unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet To temporarily halt or
(re-start) delivery of Technet send e-mail to [log in to unmask]: SET
Technet NOMAIL or (MAIL) To receive ONE mailing per day of all the
posts: send e-mail to [log in to unmask]: SET Technet Digest Search the
archives of previous posts at: http://listserv.ipc.org/archives Please
visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for
additional information, or contact Keach Sasamori at [log in to unmask] or
847-615-7100 ext.2815
-----------------------------------------------------
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 1.8e To
unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet To temporarily halt or
(re-start) delivery of Technet send e-mail to [log in to unmask]: SET
Technet NOMAIL or (MAIL) To receive ONE mailing per day of all the
posts: send e-mail to [log in to unmask]: SET Technet Digest Search the
archives of previous posts at: http://listserv.ipc.org/archives Please
visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for
additional information, or contact Keach Sasamori at [log in to unmask] or
847-615-7100 ext.2815
-----------------------------------------------------
http://mail.yahoo.com
---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 1.8e To
unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet To temporarily halt or
(re-start) delivery of Technet send e-mail to [log in to unmask]: SET
Technet NOMAIL or (MAIL) To receive ONE mailing per day of all the
posts: send e-mail to [log in to unmask]: SET Technet Digest Search the
archives of previous posts at: http://listserv.ipc.org/archives Please
visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for
additional information, or contact Keach Sasamori at [log in to unmask] or
847-615-7100 ext.2815
-----------------------------------------------------
---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 1.8e
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to [log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to [log in to unmask]: SET Technet Digest
Search the archives of previous posts at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for additional information, or contact Keach Sasamori at [log in to unmask] or 847-615-7100 ext.2815
-----------------------------------------------------
|
