BGA Pull test

First, any results of pull or shear tests are unscientific at best. [We pop our BGA from boards with, appropriately enough, a beverage can opener.]

Second, we have no have problems with your ENIG specification.

Third, as with your customer, we'd expect foil tearing and board-based distruction when popping BGA [or any other part] from the board. Breaking at the nickel tin intermetallic is not necessarily a bad thing. While we have no ENIG boards, in that case, we: * Don't expect to see a flat nickel plane for nothing. * Expect a rough solder surface attached to the pad.

So, let's assume that your customer is seeing nickel pads with no [or little] solder. Failure of this type does not have to be hypercorroded nickel [black pad]. From a board fab stand-point, oxidized nickel could result in similar problems. * Could the gold be pourous or unevenly plated for any reason? * Could the nickel have sat around for a while prior to immersion gold? * What are the results of the customer's in-bound solderability test of the this lot of bare boards?

Finaly, OK. OK. Let's give the poor fab a break. [Everyone blames the fab. They're just such easy targets.] You only mentioned the BGA popping from the board. * If this is a fab problem, other components should be just as poorly attached. * What if your customer [or their BGA suppler] changed their process? Maybe the alloy of solder between the paste and the BGA solder balls has changed and it's not as strong as it used to be.

reply »

When we first got into BGAs on ENIG we were getting joint cracking at in-circuit test. The joints were breaking in the Sn-Ni intermetallic. We did a lot of pull testing while working with our board suppliers and our conclusion was that pull testing is not a good indication of board or solder joint quality. If you pull test, for example, a 324 ball PBGA on ENIG you'll see 300+ balls break in the intermetallic and most of the rest will tear the copper from the board. Repeat the test on HASL, OSP, or silver and you'll get 300+ copper pads tearing from the board and the rest breaking at the top of the ball and leaving solder on the pad. Having the board supplier mess with the ENIG process will have minor effects, but the bottom line is that Sn-Ni is brittle so it is the weak spot, especially at high strain rates. What your customer is seeing (breaking in the Sn-Ni) doesn't surprise me and isn't an indication of a process or board problem - that's just ENIG.

Are the joints breaking in the field or are you catching them at a test in your factory? If you're catching some then I'd get out the strain gages and start working on reducing the loads in your processes - especially those where the loads are applied quickly such as ICT. If you really want a test to evaluate joint strength then set up a 3 or 4 point bend test. Don't use an Instron to drive it because they can't reach the strain rates you'll need to simulate ICT. Build a fixture with an air cylinder so you can apply the loads quickly (~ 20,000 ustrain/sec).

reply »

Hi Guy's thank you for the Info,

This is a New Assembly house that our OEM has Moved to and they have not been very open in discussing this issue other than they are telling us this is board related.

I have ask for samples of faild product with the chips so we can see what their seeing and they have not supplied us with that as of yet. So for know all I have to go on is what I see on a bare board and what the Enig looked like, We checked things like Solder mask Reg and cure and both looked good.

So for know I'm in the dark to what they are seeing, until they can supplie us with a Sample so I can send it out to a Lab to have it check..

Thanks again for your Idea's,

Mike

reply »