Hi team! I have been just sitting in the background watching the
conversation but Joyce got me to finally hit the keyboard. There is a
ginormous amount of information on BGA voids in the public domain. Some of
it is very good and some is, well, less good. Rockwell Collins and several
other OEMs (Lockheed, HP, Blackberry, etc.) looked at voiding (both
tin/lead and lead-free) for the JSTD-001 specification committee as part of
an activity to change the BGA voiding requirement. It was an effort that
took 4 years and several co-ops to finish but the outcome was the following:
1) BGA voiding is driven by the reflow profile and the specific solderpaste
formulation you are using. If you try to intentionally make voids for an
investigation, it takes extreme efforts. Voids are a result of poor reflow
soldering process so if you are getting voids (more that 1%-5% is typical
depending on the paste you are using) then you need to track down the root
cause.
2) There are still two "void theories" that are still running around the
industry: (1) voids are beneficial to solder joints because the "blunt"
crack growth and/or provide compliancy to the solder joint; (2) voids are
detrimental to solder joints because the accelerate cracking by shortening
the crack path length necessary for failure. Two really seemingly logical
ideas that have very very very little published, data supported rationale.
IMHO I don't think either idea has significant merit and I have looked
really hard for the supporting data.
3) It would be really cool if simple Force per Area calculations would
provide the industry the answer but when you look at the published data,
that idea falls apart very fast. The primary reason is that the overall
problem is complex as the forces involved (shear, bending, etc.), the BGA
construction, the PCB construction, and the nonlinear solder alloy
characteristics over temperature play a role on the solder joint integrity.
There are BGA solder joints that are nearly "hollow" that haven't cracked
after 2000 thermal cycles from -55C to +125C.
4) The JSTD-001 specification has the BGA void requirement criteria and the
IPC-7095 standard contains an enormous amount of information on
void/process control and Xray assessment methodologies. Don't confuse the
two documents as they both fulfill important and significant needs that are
not the same - lots of folks do and it creates industry confusion.
5) It would be awesome if everyone had an Xray system that would provide 3D
imaging but that isn't a realistic approach for the industry or the
specifications. The industry data does support that if your voiding stacks
up at the shear interface at either the component pad/solder joint or board
pad/solder joint then you will lose solder joint integrity quickly. But -
given that BGA voiding should be/is nontypical then the use of transmission
Xray assessment should be adequate. Again, the IPC-7095 standard contains
significant information on how to address void control/assessment.
BGA voiding is one of those topics that really gets discussions going!
Dave Hillman
Rockwell Collins
[log in to unmask]
On Mon, May 11, 2015 at 5:33 PM, Yuan-chia Joyce Koo <[log in to unmask]>
wrote:
> minor voiding actually are good for solder joints (a old old story -
> provide it is uniform and small and not that many and round). not sure if
> there is any digital record... early 90 and late 80 stuff. Am I dated
> myself? OMG...
> jk
> On May 11, 2015, at 6:23 PM, Wayne Thayer wrote:
>
> Practically, XRAYS are used to measure voiding since that is
>> non-destructive. The XRAYS used are uni-directional, so they basically
>> have
>> no clue as to vertical position of a void: They only allow you to estimate
>> areas where there is less mass for the XRAYS to interact with. Hence we
>> use
>> area. (And this seems adequate also!) Since the XRAYS are good detectors
>> of
>> missing material, what is called "area" is really something about missing
>> mass. In fact, if the actual void is 2 microns tall and occupies 95% of
>> the
>> area on a joint, the XRAYS just plain don't see it at all.
>>
>> Wayne Thayer
>>
>> -----Original Message-----
>> From: TechNet [mailto:[log in to unmask]] On Behalf Of Joyce Koo
>> Sent: Monday, May 11, 2015 4:32 PM
>> To: [log in to unmask]
>> Subject: Re: [TN] Voiding evaluation
>>
>> it is all depend upon what kind of stress - temp cycle most likely you
>> have
>> experience shear load, vib and impact are different, pending on direction
>> (6
>> faces, x-edge, y corner, etc.). strain rate also place the role. voiding
>> can not be a uniform acceptable criteria... it is all depend upon how much
>> design margin allow. (we can see the cow go home...)
>> jk
>>
>>> If the criterion was the evaluation of the ratio load/area, one should
>>> sum only the void areas that lie on a same plane (perpendicular to the
>>> load),
>>>
>>> For example, if the direction of the load is vertical, the areas of
>>> voids placed on higher or lower planes shouldn't be summed, because
>>> the solder joint area which bears the load in each plane depends on
>>> the voids intersected by that plane, not on the voids placed higher or
>>> lower in the solder joints.
>>>
>>> Since the sum is extended to the entire volume of the solder joint,
>>> the criterion would seem another one.
>>>
>>> Enrico
>>>
>>> Il 08/05/2015 15.55, Ed Hare ha scritto:
>>>
>>>> I would suggest that the area criterion is appropriate since stress =
>>>> load/area. It is not a missing mas issue in my opinion, it is a
>>>> reduction in load bearing area that is of concern.
>>>>
>>>> Ed Hare
>>>> VP SEM Lab, Inc.
>>>> www.semlab.com <http://www.semlab.com>
>>>>
>>>> On Fri, May 8, 2015 at 5:15 AM, Enrico Galbiati
>>>> <[log in to unmask] <mailto:[log in to unmask]>>
>>>> wrote:
>>>>
>>>> I would like to ask anyone if there are any reliability data
>>>> regarding the evaluation of voiding in a solder ball (BGA).
>>>>
>>>> In both the standards IPC-7095C and IPC-A-610F, the voiding in the
>>>> solder balls is evaluated measuring the area of the voids.
>>>> However, the weakening of balls caused by voiding should depend on
>>>> the amount of the missing material caused by the presence of
>>>> voids. If this is true, the amount of the missing material should
>>>> be measured by the total *volume* of voids, not by the area.
>>>> Consequently, the limit should be set on the volume, instead of
>>>> the area.
>>>>
>>>> For example, with the present rule based on the percentage of area
>>>> of the voids, a solder ball of 0,85 mm diameter, with a single
>>>> void of 0,45 mm diameter, is acceptable, since the percentage of
>>>> voiding is 28%, thus less than the maximum limit of 30% (ref.
>>>> IPC-A-610F). In this case, the missing volume of the material is
>>>> 15%.
>>>>
>>>> Considering another example, if a solder ball has 6 voids of a
>>>> 0,20 mm diameter each, giving 33% of the area of voiding, would be
>>>> rejected. However, in this last case the percentage of the missing
>>>> volume is only 8%, i.e. less than the previous case (about 53% of
>>>> the previous case!).
>>>>
>>>> So, the ball of the second case is rejected, even if it stronger
>>>> that the one of the first case. On the contrary, it is the solder
>>>> ball of the first case that should be rejected.
>>>>
>>>> Enrico
>>>>
>>>> --
>>>> Enrico Galbiati
>>>> Consulenza Affidabilità e Normative
>>>> Via Kennedy Ingresso 2, 20871 Vimercate (MB) - Italy
>>>> Desk: +39.039.8908.4547 <tel:%2B39.039.8908.4547> - Fax:
>>>> +39.039.8908.5051 <tel:%2B39.039.8908.5051> - Mobile: +39.335
>>>> 6833616 <tel:%2B39.335%206833616>
>>>> E-Mail:[log in to unmask]
>>>>
>>>>
>>>>
>>>>
>>>> ______________________________________________________________________
>>
>>> This email has been scanned by the Symantec Email Security.cloud
>>>> service.
>>>> For more information please contact helpdesk at x2960 or
>>>> [log in to unmask] <mailto:[log in to unmask]>
>>>>
>>>> _____________________________________________________________________
>>>> _
>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Ed Hare
>>>> gmail - [log in to unmask] <mailto:[log in to unmask]> gvoice -
>>>> 360-453-7550
>>>>
>>>
>>> --
>>> Enrico Galbiati
>>> Consulenza Affidabilità e Normative
>>> Via Kennedy Ingresso 2, 20871 Vimercate (MB) - Italy
>>> Desk: +39.039.8908.4547 - Fax: +39.039.8908.5051 - Mobile: +39.335
>>> 6833616 E-Mail:[log in to unmask]
>>>
>>>
>>>
>>> ______________________________________________________________________
>>> This email has been scanned by the Symantec Email Security.cloud service.
>>> For more information please contact helpdesk at x2960 or
>>> [log in to unmask]
>>> ______________________________________________________________________
>>>
>>>
>>
>> ______________________________________________________________________
>> This email has been scanned by the Symantec Email Security.cloud service.
>> For more information please contact helpdesk at x2960 or [log in to unmask]
>> ______________________________________________________________________
>>
>
> ______________________________________________________________________
> This email has been scanned by the Symantec Email Security.cloud service.
> For more information please contact helpdesk at x2960 or
> [log in to unmask]
>
______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please contact helpdesk at x2960 or [log in to unmask]
______________________________________________________________________
|
