Board Thickness Effect on ATC Reliability

The purpose of this project is to determine the effect of varying board thickness on the solder joint reliability (SJR) of a variety of devices that are attached to those boards. Opinions, modeling results, and data on this topic are conflicting. There have been no definitive studies published.

Project stage: 
Lead company: 
i3 Electronics

If you are interested in participating in this project:

Members - Go to the "Subscribe Here" section to the right and select the "Subscribe to space" key.

Non Members - Go to the "Contact" section to the right and select the "Contact project facilitator" key.

Idea Information

Background: 

There are a lot of differing opinions on the effect of board thickness on solder joint reliability.  Modeling indicates that thicker (stiffer) boards afford less mechanical compliance and shorter ATC life. Other work contradicts these findings. No definitive published work on larger components and BGAs. This project will seek to resolve this conflict for both small and large components.

Problem: 

Much of the industry tests component solder joint reliability on thin PCBs per
IPC/JEDEC 9701. This data is difficult to translate to thicker PCB’s more typical of high reliability applications. The stiffness and the resulting increased stress in solder joints on thicker boards may affect the results of ATC testing differently from thinner, more flexible boards.

Published data on the effect of board thickness on component ATC is limited to a few papers on smaller CSP components.

 

Definition Information

Goals / Benefits: 

To resolve the issue of ATC variation between thin boards and thick boards by:

  • Using an existing Test Vehicle (The HDPUG VIP test board), fabricate it in thicknesses of 1mm (.040”), 1.58mm (.062”), 2.36mm (.093”) and 3.175mm (.125”) thicknesses
  • Fabricate this board with the microvias either Cu filled for epoxy filled and plated over.
  • Use 4 different component types;
    • Large CBGA 483 (used on multiple prior projects, donated by Oracle)
    • CBGA 192, CTBGA 84 – available from Practical components and used on various HDPUG and iNEMI projects
    • QFN72 – available from Practical components.
  • ATC at 0-100C until 63% failure. Also test at -40 to +125C
  • Solder pastes: Sn/Pb and SAC 305
  • Plot and graph performance of these different component types vs. board thickness. 
  • Compare SnPb solder ball and assembly leg to published literature.

 

High level objectives: 

Provide data to make it possible to translate tests on thinner boards to thicker PCB’s.

Determine the effect of board thickness and stiffness on ATC solder joint reliability results.

Include larger BGA packages as well as the smaller packages that can be used for comparison to previous published work.

 

Approach: 

  • Use the existing HDPUG VIP test board)
  • Thicknesses of 1mm (.040”), 1.58mm (.062”), 2.36mm (.093”) and 3.175mm (.125”) 
  • To eliminate the via-in-pad variable, fabricate this board with the microvias either Cu filled or epoxy filled and plated over.
  • Components: Large CBGA 483, CBGA 192, CTBGA 84, QFN72
  • ATC at 0-100C until 63% failure. Option to also test at -40 to +125C
  • SnPb solder ball and assembly option to compare to published literature
Public