Hi, Gabriela
You did not mention what type of circuit board it is, and what the finish is. But the defective BGAs with cracks can certainly be reworked.
The BGA, to me, is like any other component. They can be reworked, provided it is done correctly.
There is nothing wrong with simply adding flux and reflowing the existing solder spheres. Quite often it will fix the problem. I have done this in previous jobs for assemblies that failed test with opens under the BGA due to an improper reflow profile. I have also done it with BGAs that exhibited cracks caused by warpage stresses or by process and handling stresses.
I have noticed that reworking a BGA site on an assembly does not usually shorten its reliability life span. Historical failure data has indicated just the opposite, especially for boards with ENIG finish. I assume it is because when a BGA is reworked (removed/replaced) on an ENIG board there is a good coating of solder on the nickel on the pads after the site preparation is done (the pads are now freshly tinned, so to speak). Therefore a good intermetallic formation (or foundation) is created, and when a new BGA is mounted and reflowed, you are essentially soldering to pads where a good IMF is already established. This IMF is essential to a reliable solder joint. It is more difficult to achieve when soldering to nickel, as the dissolution rate of nickel into the IMF is much slower than copper. This is one of the reasons I have stated so often that when soldering BGAs, ENIG is not the finish of choice. However, the IMF is the weakest part of the solder joint. It's kind of a catch-22, you need the alloys to blend, but you do not want that blended area to be too large. Simply adding flux and reflowing also helps a good IMF to form or reform with the nickel base.
Warped BGAs are BGAs that are under stress. During reflow, the top side of the BGA substrate gets a little hotter or not as hot as the bottom of the BGA substrate. This is because of the insulative nature of two pieces of fiberglass stacked on top of each other, the BGA and the circuit board. Therefore one side of the BGA expands faster than the other, and this is what causes BGAs to warp, either up or down. When the reflow profile is too short in the Time Above Liquidus Temperature (TALT), and then cools down too rapidly, the BGA does not have the opportunity to get a full soak of the heat so it can fully relax back to its original state. Remember we are not talking about the amount of time elapsed from liquidus to room temperature, we are talking about the amount of time from peak temperature down to 183 deg. C., the point at which the solder solidifies, which is often only a few seconds. If a complete soak is not achieved by this time, when the balls solidify the substrate is "locked up" in its warped condition. As it continues to cool to room temperature it wants to go back to its equilibrium stage. It cannot, therefore the solder balls are under stress.
Some BGAs have a copper heat spreader built into them and this can cause the warpage to be even more pronounced. I know there are some BGAs of this type where warpage is practically inevitable, and there is little you can do to the process to avoid it.
So now you have a BGA package under stress, and it is soldered to a nickel base where the IMF is not the strongest. Additional stresses placed on the package such as from pressing on a heatsink with thermal grease, from clamping the pwb down when building up the box and flexing the board, from ICT testing when a vacuum is applied and the board is pulled down onto the bed-of-nails tester, from stresses seen during flexure when pressing in the press-fit connectors after SMT assembly, stresses applied from a thousand different possible sources can cause BGA solder ball fractures to begin. As the board undergoes thermal cycling in its normal service life in the field, the solder's crystalline structure or grain characteristics change over time, and this morphology typically becomes coarser and coarser, and these cracks propogate until a complete open is achieved and the assembly fails.
Compounding the problem, if ENIG is not properly done by the fabricator, conditions can exist where a nickel oxide layer is formed under the gold and on top of the nickel before the board is ever assembled and soldered. If the board is baked prior to assembly to remove moisture, this nickel oxide barrier continues to grow, because heat exacerbates it. It continues to grow during reflow, during rework, and during the time it is cycling in its service life. This condition can also cause catastrophic failures. I have seen boards come back from the field with this condition, and the BGAs fell off of the board. Remember, these are assemblies that passed ICT, passed full functional testing, passed HALT and/or HASS testing, and continued to work for some period of time while in service. I am not the only one, many others have seen this, and it is well documented.
Performing a flux-and-reflow as a rework method can fix the problem, perhaps for a long time. It depends on what caused the failure in the first place. But the rework profile should allow the BGA to achieve a full soak, and then a cooldown of not less than 2 degrees per second to solidus. I know of many companies that routinely rework BGAs with a simple reflow, followed by a good wash and re-test.
There is so much more I could add to this, but it would require an epic novel.
Happy Friday, all.
-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Gabriela Bogdan
Sent: Friday, October 28, 2005 5:05 AM
To: [log in to unmask]
Subject: Re: [TN] BGA failure and assembly repair
Thanks, Inge, I will try to do it.
I was thinking also about the design of the boards, like stress factors after soldering.I can not eliminate those- we are just assembling a wide variety of boards and we don't always know what happes after assembly.
I was thinking also about the residual stresses in the BGA-like warping.Each failure has its own footprint, which I recognize. The question is, what others are doing usually when they have to repair a BGA site- what is their experience in reliability and cost?
IPC speaks only about replacing.Is there someone who is doing additional reflow on the old component as a repair method? Is it considered reliable?
Gaby
----- Original Message -----
From: "Ingemar Hernefjord" <[log in to unmask]>
To: "'TechNet E-Mail Forum'" <[log in to unmask]>; "'Gabriela Bogdan'"
<[log in to unmask]>
Sent: Friday, October 28, 2005 12:41 PM
Subject: SV: [TN] BGA failure and assembly repair
> Gaby,
> If the BGAs are OK, and your is process OK, then I can't see better
> solution than search for a better board supplier.
>
> You'd better send some pics on the cross sections. You told us nothing
> about the failure freqency per position, per board, batch dependency,
> and variations by time. To change supplier is of course a question of
> what the failures cost you.
>
> Inge
>
> -----Ursprungligt meddelande-----
> Från: TechNet [mailto:[log in to unmask]] För Gabriela Bogdan
> Skickat: den 28 oktober 2005 09:37
> Till: [log in to unmask]
> Ämne: [TN] BGA failure and assembly repair
>
> Dear Technetters!
>
> We sometimes receive failed boards from the field with failures on
> BGA's for repair.
> Supposing there is no problem with the BGA as a component, supposing
> that the X-ray test (feinfocus tiger) does not reveal any problems in
> the solder joints (with exception to cracks which are very hard to
> detect), what procedure would you suggest?
> We are concerned with the following:
> Reliability
> Cost: the assemblies are still our responsability We are considering:
> Scrap
> Repair by reflow and electrical test
> Repair by installing a new component and electrical test- this
> includes the possibility of visual inspection of the area under the
> failed component and finding more out about the failure mechanism)
> Lately we have performed several destructive tests in order to see
> the nature of the failures which could not be repaired by reflow, and found:
> through hole cracks or corrosion in plugged holes pad lifting and
> conductor cracks solder joint cracks All these were not detected by
> X-ray.
> Could you suggest the best procedure?
>
> Gaby
>
> ---------------------------------------------------
> 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
-----------------------------------------------------
---------------------------------------------------
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
-----------------------------------------------------
|
