BGA open joint

Subscribe to Thread

Customer returned a brd claim that one of the FPGA is having... - Aug 04, 2004 by

Did you record the thermal recipe on the ball that your cust... - Aug 05, 2004 by davef

We get this from time to time and I have found that a lot of... - Aug 05, 2004 by RDR

Agreed. Joint failure can happen when un-bagging or durig b... - Aug 05, 2004 by

You don't mention the board surface finish, but BGA's on NiA... - Aug 06, 2004 by

C.W

#29848

BGA open joint | 4 August, 2004

Customer returned a brd claim that one of the FPGA is having "cold solder joint". I inspected the the BGA location with X-ray and Ersascope, balls' shape look fine, voids are not detected, perimeter joints show shinny and smooth appearance, i have seen no evidence of "cold solder", secondly, this brd actually passed Functional Test before, i just can't believe it is a cold solder problem. But customer insist the joints are bad. We reball the BGA and reinstall it back to the brd using the same paste, profile etc., and the brd passed test now. Is that really mean that the failure was really a process related issue?

note: i've already verified the profile against the paste supplier recommendation, everything looks good tho.

How do you normally deal with your customer if you encounter situation like this, simply admit this is a process issue? or there is something we can share with the customer?

Thanks' chester

reply »

davef

#29888

BGA open joint | 5 August, 2004

Did you record the thermal recipe on the ball that your customer complained as open or some other ball?

reply »

RDR

#29896

BGA open joint | 5 August, 2004

We get this from time to time and I have found that a lot of times the customer or somebody has broken the joints from flexing or other handling. Is this component in the center of the PCBA?

Russ

reply »

KEN

#29908

BGA open joint | 5 August, 2004

Agreed. Joint failure can happen when un-bagging or durig box-build (at customer site). Careful handling should be used. In one case I had boards returned to me for evaluation and failure analysis. Too bad our site never built the boards!!! (jackass customer gave me hell about our quality then sends us boards built by another supplier)

Also, don't count out black or "dark" pad. We have seen this time and time again. Passes ATE, but fails shortly during test at customer site. Try pressing on the device during test...sometimes it will pass....sometimes not. This is a good indication of a joint separation.

reply »

Bob R.

#29918

BGA open joint | 6 August, 2004

You don't mention the board surface finish, but BGA's on NiAu are prone to brittle fracture in the Sn-Ni intermetallic on the board side of the joint. You may not be able to see it visually, but it would show up in a cross-section. In-circuit test was breaking a lot of our BGA joints on NiAu boards. We measured the strain at about 1200 microstrain and the strain rate a 20,000 microstrain/second which means the boards are getting hit pretty fast. We modified the tooling in the fixture to get the strain below 650 and the problem went away. I did a lot of bend testing at various rates and lowering the strain rate would have eliminated the problem also. The in-circuit test engineers wouldn't let us reduce the strain rate because they said the probes wouldn't be able to break through the flux (test was before clean in this case).

We don't see the same failure mechanism on OSP or HASL boards. Our board suppliers have had black pad under control for several years and I think a lot of what gets blamed on black pad is actually just brittle fracture due to high strain rates.

reply »