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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

BGA opens

Views: 7359



BGA opens | 14 January, 2007

I have several boards built at various CM's. We have prototype builds with some JTAG failures on BGA's (1-3 failures with qty 3, 1500 pin BGA's). CM's always claim reflow profiles look good. X-ray does not show bridges and they claim it looks good, but cold solder joints and opens are difficult to see. I have had reflow temperature issues with other CM's (wrong profile loaded, etc). I have seen these issues with gold and OSP plating, but the latest boards were OSP. PCB's had 100% electrical check. Some have benefited from reflow. Some local reflow showed additional JTAG failures down stream. The failures in general are on different pins and different BGA�s during testing. Some failures can be corrected by pressure on BGA (colds solder joint?). I plan to do some destructive dye test, but I am looking for advice I may not be looking for. CM's generally focus on bad/old OSP or bad boards. In the past, I have not seen that on other prototype PCB's (sectioning, etc looked good). Reflow sometimes works. Generally these are lead free processes.

reply »


BGA opens | 15 January, 2007

Contractor mantra: * Profile is good * Boards are bad

On profiles, the key question is: where did you take the profile?

We�d guess cracked connections are giving the intermittent test results when pressing on the components. You use acoustic microscopy to assess this, which would save you from sacrificing the boards. Although, dye testing is fun to do.

BGA cracking is caused by mechanical stress. This may be the result of: * Rough handling of soldered board * Inadequate consideration during package design of CTE effects on BGA * Poor control of or selection of thermal ramp rates during soldering * Excessive force from ICT fixture

We talked about BGA cracking several times. Search the fine SMTnet Archives. For instance

reply »

Peter W.


BGA opens | 23 January, 2007


maybe you are experiencing the so called "head in pillow" defect. It is especially critical for larger BGA. The defect comes from the deformation during reflow of the BGA, so that some of the balls are lifted and have no contact during liquidus. When the temperature declines, the BGA-warpage changes, and may press down the still soft ball on a still soft pad, which is wetted with the printed solder paste. Most of the time they have electrically contact, as they touch each other, but the solder deposits are not melted together. So, especially during temperature changes (and thus induced changeing of the warpage of the BGA again) you may see sudden opens.

This defect is very hard to detect, mainly by manual XRAY from a 30� angle, or by speciallized AXI routines. When you do a cross-section, you will find this defect.

To prevent a "head in pillow" defect, try some or all of the following: - use Nitrogen - optimized your reflow profile - print thicker solder paste on the board - overprint the pads

and hust before the production: - make sure your BGA has a very low warpage/ a good co-planarity

reply »



BGA opens | 23 January, 2007

Well I am a CM, it has to be that OSP finsh!!! hahahaha

seriously I hate that finish. but this is big BGA and there could be a multitude of issues as you suggested.

Do not assume that the boards were 100% tested for all vias and lines unless yuou know for certain that they are included in test. I have had many boards with test stamps that had internal breakage at vias and what not.

Have the CM show you the profile and have them explain how it was performed as Dave mentioned.

reply »


BGA opens | 24 January, 2007

Seriously, OSP?, change the board finish, you are the OEM, correct? You have the power! Also, as another poster mentioned, high force applied at ICT ( possibily to cut through the OSP goop ?!? ) could be warping the board enough to crack the joints. Check the probe force uniformity, support pin heights and watch the operators. If you are using an overclamp fixture vs. Vacuum, try and watch the boards as they are being clamped and see if there is any deflection.

reply »

Phlip Kid


BGA opens | 24 January, 2007

CMs suck!

reply »



BGA opens | 24 January, 2007

I'm probably not telling you anything you don't already know, but a technique that's worked very well in the past for me is to methodically step through every part of the process and observe your boards being built, or take a scrap or simulated one and walk it through. See what you learn.

I do agree with the other poster on ICT. Also not mentioned is support pin position in addition to length. Another one for ICT is the fixture itself. Carefully examine the "pocketed" areas that are clearanced for components and see if there isn't an interference there, which could cause localized warp/flex.

Also check the ramp up/down rates for the reflow profile being used, as well as time above liquidus and peak temps. If it is a thick and dense board, they may be on the cool side, etc. If the BGA has an integrated spreader on it, you may need very high preheats and longer than normal above liquidus to anchor it to the board.

I have worked in both OEM and CM worlds, and had the same failure you mention; having good results with the two items above fixing the problem.

There is one time where we had this same problem, but it turned out to be the BGA supplier at fault. They changed their process of attaching the balls to the substrate; their new process was faulty.

Good luck.


reply »

Sweet Old Bob


BGA opens | 28 January, 2007

Okay you CM haters I don't see anything about the design of the board, other than OSP which is strike one. What is the board size? What is the board thickness? Where is the BGA located on the board? What is your ball to pad ratio? What is your aspect ratio on your vias? What is the the raw fab material? How's your layer stack and thermal relief? What's your minimum trace width? I will assume you have solder mask between yor pads and vias as stated in PCB101.

When you have made sure that all the above are looked at and correct, then blame your CM as all of the above can cause you issues.

Hint: Raw PCBs that have been processed with ENIG, OSP, Immersion Silver have not been subjected to any real thermal heat cycle, which makes electrical test almost a mute point.

After being in the CEM business for 25 years I can guarantee you 99.8% of BGA issues are design related.

If you don't belive this then continue to suffer.

Thank you for your time.

SOB (Sweet Old Bob)

reply »



BGA opens | 8 February, 2007


Can you please amplify each of your points - what limits do you recommend.

Yours, Very Ignorant

reply »


BGA opens | 8 February, 2007

Wow, what a bunch of crabby old men. BGA opens can occur from pre,during or post process. I never chimmed in cause the original poster said they would performa a dye test. Yet as usual, they never get back to this forum on what the cause was. So arguing the points of BGA opens is fruitless. I bet the real problem was a bad by pass cap.

reply »



BGA opens | 12 February, 2007

I think the best way is to the following: - to use the low temp solder paste, or - asks semi-con use low temp solder sphere to make BGA

reply »


BGA opens | 21 February, 2007

BGA open? Isn't some golf outing? Can't wait for the golf season so all the engineers here go away.

reply »



BGA opens | 2 May, 2007

Has anyone experienced deformed balls on BGAs - especially very low volume build to order devices? I'm looking into a possible hidden pillow defect and think I have eliminated most of the other factors I believe could cause it.



reply »


BGA opens | 2 May, 2007

I think all persons above are correct in a way, > to get a 0-defect layout, OEM and CM have to work together,

If you know what connections (balls) are the problem ones, let me know, and send me the gerber-layers + solderpaste layer, so we can see if the problem is design related or if you CM realy sucks.

If you blaim your solderprofile, maybe test VF-soldering for once.

regards Bart

reply »

IPC Training & Certification - Blackfox

Online IPC Training & Certification