It is getting hard to find BGAs that come in Tin-lead solder. Most of the newer components are only available in SAC305 lead-free solder. So what are we supposed to do?
One expensive way of doing things is to re-ball the BGAs. This process involves scraping off the SAC305 from the BGA and attaching SN63Pb37 balls. The user can then use the new package as a regular Tin-Lead component.
Source:https://www.micross.com/component-modification-services/bga-reballing/
Some applications will use special stencils where the whole PCB has a Tin-Lead stencil and only the BGA will have an individual stencil of SAC305. This process is similar to rework of a single BGA.
Some PCB designers are using SAC305 BGA components in a Tin-Lead PCB. The solder paste on the PCB is SN63Pb37 and it is applied in the regular stencil. The result that they are hoping for is a full mixing of the solder to a new alloy
But what actually happens looks more like a lead-rich volume near the stencil was and a lead-free volume on the package side.
Source: Kinyanjui, Robert, et al. "Challenges in reflow profiling large and high density ball grid array (BGA) packages using backward compatible assembly processes." Proceedings of IPC APEX 2010 (2010): S35-03.
This lead-rich area will be different in each individual solder ball. The resulting package will have hundreds of different solder balls and each one will have a slightly different mechanical performance. These differences may be small enough that they will not matter overall but they are practically impossible to predict using regular reliability physics analysis tools.
Let's assume that the reflow profile is tweaked in such a way that the collapsed ball will result in a complete mixing of the solder. The alloys for each ball will be slightly different due to manufacturing variances in stencil thickness and ball size.
This example shows some "back of the envelope" calculations for a SAC305 BGA with a Tin-Lead Solder paste on the PCB
The resulting alloy is 86% Tin and 10% Lead with some bits of Silver and Copper. There should be some test data for this alloy before using it in high-reliability applications.
The resulting alloy for a different BGA ball size will be 90% Tin and 7% Lead with some bits of Silver and Copper. The test data for the BGA with 300 micron balls is not valid for the 500 micron balls. The testing and validation needs to be repeated
It is getting hard to find BGAs that come in Tin-lead solder. Most of the newer components are only available in SAC305 lead-free solder.
The resulting solder ball will not automatically create a perfectly mixed solder alloy without changes to the reflow profile. And even if the mixing can be sorted out then the alloys created will change for each geometry.
Some applications will be fine with these mixed solder balls. High reliability applications will need to have some test data for validation.
Copyright Gil Sharon June 16, 2021. All rights reserved.