Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Low Silver Solder Problems

Views: 6472

We are using a vapor phase oven which we have used successfu... - Aug 21, 2009 by ross  

#59600

Low Silver Solder Problems | 21 August, 2009

We are using a vapor phase oven which we have used successfully for a number of years with both leaded and unleaded solders.

Recently we experienced problems with a no lead BGA that has low silver balls. The problem is that the soldered boards do not hold up to shock (even just shipping the product can result in failures).

We have tried increasing our dwell to 85 seconds at 245C and this helped but did fully correct the problem. We increased the temperature to 253C and still the problem persists. We have changed solder paste to one that has a more active flux, still no help. We have verified profiles and have tried to get as close as possible.

If anyone can offer any suggestions I would be very greatful.

Eric

reply »

#59610

Low Silver Solder Problems | 22 August, 2009

Dear Eric, in my opinion this is a very complex problem, that require to know many details for answer. For example : 1) what about the PCB finishing ? Have you verify the quality of they ? 2) what about the solder paste (the exactly composition)? 3) The crack at T=0 after reflow, where is located ? At the corner or at the center of BGA ? And always on the same BGA or random position ? 4) Have you performed a cross section for verify the failure mode of crack ? It is : - at the pkg side IMC/solder interface ? or - at the PCB side IMC/solder interface ? or - at the PCB metal/IMC interface ?

I think before is necessary to have the exactly map of situation, for found the solution.

And finally, try to use a SACX Sn-0.3Ag-0.7Cu+Bi composition (but only after tests).

Good work

Hi ghepo

www.analysispcb.com

reply »

#59612

Low Silver Solder Problems | 22 August, 2009

Thank you for replying. 1) the finish is Gold Immersion (ENIG) 2) paste type EM907 SN96.5/AG3.0/CU0.5 3) it is always the same BGA but its not on the same ball every time. It is always at the joint on the PCB pad like the solder is not sticking properly to the PCB Pad. When we just put down the paste with no BGA the solder paste seems to adhere properly to the pad.

You can sometimes cause failure by flexing the board slightly and in some cases you can hear an audible pop noise when flexed.

reply »

#59613

Low Silver Solder Problems | 23 August, 2009

OK Eric, is difficult answer at distance without examine the PCB..., anyway in my opinion the problem is the PCB quality.

There are two possibility : the crack is between solder and nickel or is between nickel and pad copper. In both situation the problem is almost always due to the presence of chemical contamination and must be verify using a SEM EDX analysis. Due to your answer I understand that the crack is between solder and nickel, right ? So maybe the problem it can caused by high concentration of phosporus or carbon. If for example after SEM you found 70%nickel/10%phosphorus/20%carbon, the carbon is the contaminant.

Another factor of joint breaking is the adoption at the project step of solder mask as restricting the solder area of PAD. Maybe the PADs are contaminated by solder mask.

If the raw card same vendor and week code are still present, you can before perform a tape test in the same position for verify the good gold attachement of the nickel.

I hope to have given some good ideas to solve this problem.

Hi ghepo

WWW.analysispcb.com

reply »

#59624

Low Silver Solder Problems | 24 August, 2009

I don't think its an issue with the PCB. When we just put the paste down and not the BGA the solder paste seems to adhere to the PCB pads very well (not brittle).

Also we have other very similar PCBs that we run different normal silver content BGAs that we do not have problems with.

Additionally, we have done several runs at different times with PCBs that were fabricated for the run and they all behaved similarly.

reply »

#59627

Low Silver Solder Problems | 24 August, 2009

OK Eric, maybe you're right. My suggestion is to proceed by exclusion. First perform the tape test. Is easy and without costs.

You have used many similar PCB without problem, but from the same vendor ?

Is my experience that in similar cases the quality of PCB is always the first contributor. Moreover, if the nickel is contaminated by phosphorus or carbon, is normal that the joint after solderability test, "seems" good.

Don't forget that often the distribution of contaminants on the PCB is random.

Anyway, I think is impossible to solve this problem without perform a serious failure analysis program (using a SEM/EDX on the PAD fail, cross section on BGA that seems well soldered for check the IMC growth, VOID inside, etc

Good luck

reply »

#59632

Low Silver Solder Problems | 24 August, 2009

Thank You for all your help. The PCBs are from the same vendor.

You mention the tape test. How exactly do we perform this test?

Eric

reply »

#59636

Low Silver Solder Problems | 25 August, 2009

Eric, use the TM 650 test method annexed.

Don't forget to perform many test on many PCB (TOP and BOTTOM)on PADs in several position (at 4 corners and at the center) and obviously in the faulty position. Don't forget to change every time the tape.

Verify PCB and tape under microscope. No portions of gold must be detached from the PCB.

Hi

Attachments:

reply »

#59638

Low Silver Solder Problems | 25 August, 2009

It sounds like the BGA balls are not melting. You're probably getting no collapse. Put a bare BGA on a hot plate. Heat it up until you get collapse. Adjust your reflow profile until you have that temperature on the BGA pads on the board.

What is the material of the BGA balls?

reply »

#59642

Low Silver Solder Problems | 25 August, 2009

Do you know the exact BGA sphere alloy make-up and the alloys melt temp ? Then you can judge the ideal reflow temperature. Also, does the package warp in the reflow zone ? This may be causing the solder spheres on the BGA to lift off the pcb and hence you don't get a true solder joint, but you tend to get some bond between the melted solder paste and bottom of the sphere.

reply »

#59644

Low Silver Solder Problems | 25 August, 2009

Eric, the problem is a crack of the nickel PAD, or is a PILLOW defect ?

See the photos annexed.

Attachments:

reply »

#59646

Low Silver Solder Problems | 25 August, 2009

It appears that the crack is between the paste and the ball. We physically pulled off the BGA by force and almost all the balls broke at the PCB. There was a thin layer of solder left on the pad. Maybe 1% of the balls stuck to the pad.

I have attached a picture of the balls after soldering.

Attachments:

reply »

#59651

Low Silver Solder Problems | 25 August, 2009

The solder balls on this part are SAC105 (98.5 Sn, 1.0 Ag, 0.5 Cu). Though the manufacturer doesn't specify, I believe that the melting point of the alloy is 228-230 C according to the research that I've done. The manufacturer recommends a reflow temp of > 250C. Neither of the package or PCB appear to be warping during reflow.

reply »

#59656

Low Silver Solder Problems | 26 August, 2009

Hi, as you are using Vapour phase reflow what temperature is the boiling point of the solvent you are using ? If it is @ 250 C then you should be getting everything hot enough to melt the ball and solder paste together. If this is the case then your defect seems to be caused by poor solderability of the BGA sphere. In the vapour reflow there will be no further oxidation so soldering conditions should be good unless the balls are so heavily oxidised in the first place. This is a cause of 'head in pillow' defect and in normal forced air convection reflow it can be worse as there is lots of hot air to cause more oxidation.

reply »

#59658

Low Silver Solder Problems | 26 August, 2009

Ok Eric, finally we know that the problem should be the PILLOW (sincerely, your photo is not very clear) and not the crack of the joint. There is a big difference on the failure mode of the two defects...

If the problem is realy a pillow I only can suggest you to use an adequate solder paste "anti-pillow" (as KOKI S3X48-M406-3, for example). Ask to your solder paste supplier. Then use a convection Ni reflow oven instead of vapor phase. Studies suggests that vapor phase may result in more pillow defects than does convection reflow, but the reasons are still unknown.

Obviously, use an adequate profile reflow as per suggested by the BGA application note.

Hi ghepo

www.analysispcb.com

reply »

#59674

Low Silver Solder Problems | 26 August, 2009

I am not sure that it is the pillow problem. With the pillow problem you tend to see residual solder on the pad and voids where this solder pulled from on the ball.

What we are seeing is a very thin flat layer of solder on the pad and a flat surface where the ball was joined to the Pad.

What I think maybe happening is that the majority solder paste is wicking up the ball leaving the thin layer underneath. And similarly to the pillow problem the two solders are not melding.

Also we used a new paste with a more aggressive flux and did not see an appreciable difference.

Have you ever heard of anything like this? Let me know your thoughts. Eric

reply »

#59689

Low Silver Solder Problems | 27 August, 2009

I have already told you that is not possible solve this problem by e-mail, without perform a detailed failure analysis. Also I have told you to procede by exclusion. Are you sure, by tests, regarding the PCB quality ? Are you sure regarding the solder paste printing process ? Are you sure regarding the BGA coplanarity ? Because if not, this can result in open solder joints if the amount of solder available is insufficient to wet to the pad or the pad solderability is not optimum.

So, if all these parameters are OK, and you cannot change the reflow method, you can only modify the vapor phase parameters and the solder paste chemistry.

Take a note that the wicking defect in the VPS method, occurs when heating is carried out too rapidly. So, is important that the PCB, fluxes and parts are preheated evenly and gradually and that the temperature be maintained at 150°C or above until the assembly reaches the vapor.

At the end of this, is mandatory verify the results, performing some cross section (minimum 1000 magnifications).

For an help, I suggest you to search a skilled consulting in your country.

Hi

reply »

#59690

Low Silver Solder Problems | 27 August, 2009

Thank You for all your help. Eric

reply »

#59691

Low Silver Solder Problems | 27 August, 2009

We recently had a very similar issue with BGA breaking off at the pad. We sent samples to a lab for SEM / EDX and they found that there was no intermatalic with the nickel, or very little do to high phosphorus. I recommend you get a board to a lab ASAP.

reply »

#59692

Low Silver Solder Problems | 27 August, 2009

Thanks Jim.

Can you tell me how you resolved/corrected the issue? Eric

reply »

#59718

Low Silver Solder Problems | 31 August, 2009

What is the boiling point of your vapour phase solvent ?

reply »

#59727

Low Silver Solder Problems | 31 August, 2009

We have tried 2. The first is 244C and the 2nd is 252C. The processor BGA document recommends a solder temperature above 250C. Eric

reply »

PCB Soldering Tools

PCB Assembly Equipment Auctions