I am designing a board and have found that a BGA component is cheaper than the larger footprint. I am curious if I should be reluctant to implement a BGA with regard to extra steps during manufacturing.

From a manufacturing standpoint I don't think a pick and place would have an issue, but I am sure it comes with some added steps.

  • If I use BGA does this increase the amount of bad boards?
    • If so, by what percentage?
  • Is BGA "easy" for most cheap manufacturing houses equipment (hopefully not too broad)?
  • Should I aim to avoid BGA even though there is a significant reason to switch?
  • \$\begingroup\$ At what production quantity (total, or per batch or per year) are you aiming? I am no expert, but my guess is that using something more demanding like a BGA will mainly increase the setup costs. \$\endgroup\$ – Wouter van Ooijen Mar 3 '15 at 16:30
  • \$\begingroup\$ 1k plus... So your thought is that beyond initial setup its not an actually increase in production steps? \$\endgroup\$ – tman Mar 3 '15 at 16:32
  • \$\begingroup\$ when they assemble the boards in the factory, BGA is just as easy as standard style packages, except there is a higher % failure depending on how you designed the pads and solder paste layers etc. \$\endgroup\$ – KyranF Mar 3 '15 at 16:39
  • \$\begingroup\$ Any idea of how much the failure actually increase? I mean I can calculate if its worth it if I had any actual data. \$\endgroup\$ – tman Mar 3 '15 at 16:41

It depends on quantity and your ability/desire to pay for additional steps in the manufacturing process. The only extra step that is typical in BGA assembly is X-Ray in QA to ensure that reflow occurred under the part as it cannot be visually inspected. You can choose whether you want this step or not. If you don't X-Ray and the assembly house has poor flow control or a bad reflow profile, it could decrease your yield. We have typically X-rayed early in prototype development, but once you have a line turned on and the reflow profile is known, then we usually don't as it becomes expensive and time consuming. However, if you are only doing small runs, I would absolutely pay the extra for X-ray as debugging opens/shorts that you can't get to is not a very good use of time. Of course, if you can break out all of your signals to test points, you may be able to find any problems that way and get around this limitation.

As to whether it is "easy" for assembly, this is going to be primarily based on the BGA pitch. This is no different than say a discrete R or C. Everyone is going to be okay placing 0603, you will have some assembly house fallout at 0402, and you are going to need higher end capabilities beyond that.

Also, keep in mind, that depending upon the pitch/ball count of the BGA, you may need to go to higher layer count just to break out of the part, or spend more money on the PCB fabrication to use things like via-in-pad or uVia to escape the BGA.

Are you limited in board space in the design? Do you mind sharing the IC you are referring to? It may help with a better answer.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.