PCB defects - problems under two BGA's

[senilicus]

PCB defects

Hello All,

lately we received a batch of assembled pcb's from our assembler and it turned out that from a 100 pcb's only 20 pieces are working oke.

There are problems under two BGA's, pitch =1mm, with courtyard via's

X-raying the pcb's doesn't explain why the yield is so low. One other possibility is that we are facing a moisture problem at the pcb level.
The pcb's are assembled in a lead-free environment and I know that the higher temperatures can cause defects on the pcb. Anyone has some experience/information on this issue.

 

[FvM]

Re: PCB defects

You didn't tell, why you know or assume that "There are problems under two BGA's", so it's even more difficult for others to form an opinion on it. The first question is, what do you actually observe?

Connection failure, caused by PCB delamination or whatever should be pretty clear detectable, by in circuit test or boundary scan.

 

[majnoon]

Re: PCB defects

Does you Assembly House bakes these boards? Do they bake BGAs prior to assembly?
As FvM mentioned have you done any ICT if not possible any other test?

M

 

[ezpcb]

PCB defects

There're many courses of problems under BGA.
1. the chip must be baked.
2. The finish of PCB should be immersion gold. HASL isn't flat
3. The vias under the BGA must covered by soldermask. This is designer's problem.
4. High-Tg materials must be used to manufacture the PCB for lead-free application because the lead-free soldering temperature is much higher.

Mike
https://www.ezpcb.com

 

[senilicus]

Re: PCB defects

Hi guy's

Thanks for you'r reactions and sorry not to respond earlier, but my compagny has moved from one location to another and it didn't go as planned.

anyway, a status update.

The pcb's did fail the functional tests, with this tests we were able to localize the problems with the two bga components. The boards were x-rayed but we couldn't see any problem.
After this, the two bga's were removed from the board and now we were able to find out what the problem was.

it turned out that there was a soldermask mis-registration, causing some trace being not covered by soldermask. This caused some solderbriding under the BGA's and this is why the tests failed. These bridges could not not be detected with the used X-ray system.

For new designs we will add some 'chevron solder mask registration feature' as described in the pcd&f magazine from jan. 2009.

Thanks for your reactions