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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Need Expert Support

blair

#5537

Need Expert Support | 13 March, 2001

Hi: We are about to build a batch of highly critical control cards with 5 BGA on them. We just recently had a significantly bad yield at our subcontractor in the San Jose area. We feel it is in the reflow profile that was used but are not 100% sure. Ideally we would like to get an expert to go with us prior to and during the next build to support us and give us hot fuzzy. We need to know the right questions to ask. It seems that BGA reflow should not be a problem but...we got less than 30% yield last batch.

Any ideas for me?

reply »

Dave G

#5538

Need Expert Support | 13 March, 2001

Here is some general things I usually ask myself when having trouble with BGA's

What type of BGA ?

Pitch/Bump Dia etc. Smaller bumps sizes & tighter pitches require more attention. This also impacts stencil aperature design & paste selection.

Was the paste inspected in these areas after printing ?

Look for adequate coverage/general paste deposit quality.

What kind of paste was used ?

The aqueous cleaning process (For water soluble pastes) can cause problems with residues left under BGA's. This is a component of a low stand-off height. No-clean flux chemistry is usually not a problem.

Did you get X-ray's of bad & good one's ?

This helps to determine placement quality and finds gross opens/shorts. Also helps in finding voiding in the joints.

Did they run a profile with a thermocouple attached near the BGA ?

Placing the thermocouple underneath the device would be ideal but, this is not always possible.

reply »

blair

#5545

Need Expert Support | 13 March, 2001

Thanks a lot for the response 1) there are 8 BGA on this board 6 PBGA, 1 TBGA and 1 ESBGA (metal lid). They are dispersed over the board (approx 11"* 14") 2) Is there a tool to use to understand pitch/bump size and stencil aperture (optimized)? 3) Yes paste was inspected - I imagine paste height only - anything else? 4) Will investigate type of paste 5) Will investigate cleaning process 6) They did take XRAY under duress. Did not see much - some evidence of voiding and poor wetting. They blamed voiding on moisture. Are there any other causes for voiding? 7) I will find out exactly their reflow profiling methods

Do you know of any expert in this area (Santa Barbara) we can learn from?

reply »

#5550

Need Expert Support | 13 March, 2001

Hi, can you definitely attribute the 30%yield to poor bga soldering? What is your high flyer defect encountered on the failed boards.Is this ICT ,flying probe ,functional test yield? Can you determine a defect pattern? Did your subcontractor provided you with quality data collected @every step of the process?what process controls do they have in place? most people will agree that 90 %of defects could be attributted to solder printing process and the variables involved (stencil design ,solder paste used ,environment,etc)I think the stencil design would definitly be critical.Without seing the actual files I can only tell you generalities .did anyone do a DFM/DFT analisys?Are the bga's on topside only?Is it a mixed technology board?(SMTtop/smtbot/th/wave?)how many reflow cycles?In what order?Are the vias under the bga's tented ? (on both sides if boards are going over the wave)Are the bga pads isolated with solder mask? If this is a subcontractor you worked with before with good results do a quality run to find out where the trouble spot is. The bga's pitch and other mechanical caracteristics are available on line from most manufacturers .Look under data sheet or specifications.About stencil design:Intel and Tessera have extensive info on their web sites.For example on a 5 mils stencil for a topside board with FP and BGA's (not micro or mini)appertures would equal the bga pads size (no apperture reduction).In certain cases one would make the stencil apperture square and equal in size with the bga pad diameter(oversized appertures).Your subcontractor should be able to give you exact resons why a stencil is designed in a certain way.If the copper distribution is fairly equal across the board and the boards didn't dryout before they made it into the oven and they use any decent watersoluble solder paste and they have succesfully reflowed qfp's in a similar board configuration you shouldn't have any problems.Look at their quality data during the run and board and stencil design.I totally agree that cleaning is extremely important but probably your boards would failed later in the field if the cleaning was inadequate.Did they measured paste heights at all bga locations on all boards .Are the failed boards bga x-rays looking worse than the ones that passed? Solder voids could be attributted to various factors and one of them is bad profile.If you have a similar weight component in a qfp package in the close vicinity of a bga somewhere in the middle of the board ask them to attach some thermocouple to it and this would give you a fairly accurate picture regarding the bga solder joints temp.I don't remember exacly but I think it is in the +/- 5 degrees celsius range. Good luck. Rob

reply »

blair

#5551

Need Expert Support | 13 March, 2001

1) We could press on BGAs and the boards would boot 2) This is basic boot up yield 3) We are now getting stpe by step inspection data going fwd. 4) So, Stencil design could be a critical factor. Any documented guidelines? 5) Yes DFX was done - mostly focussed on the PCB Fab itself. 6) BGAs are only on top side. How much could warpage affect this or PCB stackup? 7) I think process was bottom side/ reflow/top side/reflow. We had some BGA (unfortunately) added by rework. 8) I will find out about tenting and SM over via to pad under BGA. 9) We have got good results out of these guys before. 10) We usually give stencil file with our release. Typically do CMs adjust the stencil to optimize? 11) will find out if paste heights were measured before reflow 12) I will find out how they arrived at the profile.

reply »


JAX

#5554

Need Expert Support | 13 March, 2001

Are the boards in question immersion gold? This could be a process problem at the board shop. If this subcontractor has done good work for you in the past it might be an outside variable causing the problem.

reply »

#5557

Need Expert Support | 14 March, 2001

Have You considered popcorning?

reply »

blair

#5565

Need Expert Support | 14 March, 2001

Boards were not immersion gold. They were HASL but I would like to change. What is the prefered surface fininsh for BGA and fine pitch stuff these days? I am starting to ask for Gold/Nickel on FAB Dwgs due to PCB house and assy house discussions.

What is popcorning?

reply »

#5566

Need Expert Support - Popcorning | 14 March, 2001

Popcorning is the common designation for defects related to the moiture sensitivity of plastic molded components, like PBGAs.

The plastic body of these parts will absorb moisture from the ambiant atmosphere on your factory floor. If a critical level is exceeded prior to reflow then the trapped moisture will create internal cracks and delaminations when it vaporizes during reflow. This is why the moisture sensitive components are dry packed by the manufacturers and they have a maximum floor life prior to reflow specified on a label on the bag.

Popcorning is the tip of the iceberg since it is an extreme case of complete delamination. Partial cracks and delaminations are insidious latent defects which can easily escape to the field because they cannot be detected by visual inspection, AOI, electrical test or X-ray. When PBGAs are popcorned, you will notice a dome-shape deformation underneath the substrate which will cause shorts near the center of the package and opens on the perimeter.

Another side effect of moiture absorption on PBGAs is that they will tend to warp, even prior to reflow, after they have absorbed too much moisture. This will also cause shorts and opens during reflow. You can verify this by baking the parts prior to reflow and notice if the problem goes away.

Pay special attention to rework because if you heat the body of the components above 200C with localized reflow you might induce the same type of defects.

For more details on the proper guidelines to handle moisture sensitive devices consult the joint IPC/JEDEC standard J-STD-033 (free download at http://www.jedec.org).

reply »

Glenn Dody

#5586

Need Expert Support | 15 March, 2001

I am currently in San Jose working on a BGA assembly problem in a customers subcontract house. I do consulting in this exact kind of work with years of experience trouble shooting PC board assembly problems especially with BGAs. I am booked up until late next week. You can call me on my cell phone at 512-636-7986.

reply »

blair

#5675

Need Expert Support | 21 March, 2001

Thanks all for your help. I have found some one to support

reply »

HeatShield Gel- thermal PCB shield during reflow

Manufacturing Software