BGA CORNER WARP

Robert,

Were your parts in a sealed bag before use or were they exposed to the environment? If the parts have absorbed moisture, you might be seeing "popcorning"

Phil

reply »

If you heat a BGA too quickly, the corners of the device will curl towards the ceiling [potato chip], because they come to temperature faster that other areas. Adding to this, thinner substrates tend to potato chip more and faster than thicker substrates.

I wonder if it�s possible that a high ramp caused the corners curl; then the solder reflowed, expanding enough to make contact with the pad; then the solder cooled while the corners were still raised, standing the corners further off the board than the center of the device.

reply »

Are you also getting shorts further into the package? Or is the failure simply the appearance of the package?

If you send me the profile, I'd be happy to offer some suggestions. I have seen something similar.

pbourgelais@us.vitronics-soltec.com

reply »

Robert,

Someone already mentioned the popcorning phenomena. This would show up as a dome-shaped deformation underneath the component, due to internal delamination around the die area.

However what you describe sounds like another moisture-related issue that is unique to PBGAs. They tend to warp after absorbing too much moisture from the ambiant air, even prior to reflow. If you are unsure how long these parts have been out of their dry bag, this is a very likely cause for your problem. You can verify this by measuring the planarity of your parts prior to reflow. If that seems to be the problem, simply bake your components and everything should be back to normal. The proper bake conditions are defined in the joint IPC/JEDEC standard J-STD-033, available for free download at http://www.jedec.org.

If you suspect moisture-related warping you might want to check the parts previously reflowed for other internal damage (cracks and delamintations), using acoustic microspcopy.

reply »

We have seen the same thing. Our vendor recently change the way the BGA is fabricated and the new package warps. We are currently trying to adjust our process paramentes to fix this problem.

I will try the "bake out" suggestion and see if it corrects the paroblem.

reply »

BGA warpage combined with PCB warpage can lead significant process yield loss.

First, it is possible that major portion of BGA's apparent warpage is, in fact, coming from the PCB warpage. Assuming that the warpage is solely from BGA itself, I think unfortunately there wouldn't be any root-cause solution other than changing the BGA package design.

Your observation on the relative thinness of the BGA substrate suggests that the warpage is most likely due to the inadequate package design.

reply »

I think the answers to the problem have been given but if anyone wants to see this happening and solder shorts forming during corner warp there is a video clips on BGA shorts to download. Its available free on http://www.bobwillis.co.uk

It should amuse you all.

reply »

Does this device have a integral heat sink / metal slug on top, like a TI TM320C6XXX er a Altera EP20K4XX?

Compare the size of the pads on the interposer and the board.

Compare the type solder mask [NSMD vs SMD] on the interposer and the board.

When you talk about the profile, what is your point of reference [TC on the board, TC on the defective connection, over TC, erwat]?

reply »