Reliability of Selective Solder Paste on BGAs
Components with BGA packaging have become denser and have gained higher ball counts over time. The ball counts and pitches have gotten so high that some simple components now require an HDI approach with filled via-in-pad due to the fine pitches between components. These components take special considerations in assembly to ensure yield, as well as targeted X-ray inspection to ensure successful soldering.
When scaling to high volume manufacturing, there is always a pressure to reduce costs, and one area where it can be tempting to eliminate some cost is in solder printing on high ball count components. One of the most common cost reduction measures after material swaps is elimination of consumables, and solder materials are the largest consumer material in PCB assembly.
Yield in BGA Assembly
Soldering of BGAs generally uses some additional solder paste that is applied to the land pattern in the bare PCB. Some assembly houses will offer differing guidelines on this point, but the fast remains: solder paste application is the only way to provide sufficient solder on standard pad sizes. In the default paste mask data applied in your CAD tools, the solder paste will be applied everywhere in the BGA footprint. This will include both functional and non-functional BGA pads.
Functional vs. Non-functional BGA Pads
In this case with solder paste printing on a BGA, we’re not using “non-functional pads” to refer to pads on a through-hole via. Instead, we’re referring to the pads (or balls) on a BGA that do not have any function. These could be reserved pins or explicitly unconnected pins, or pins that have some function but are unused. Soldering would normally be performed by printing solder paste (selective printing) or deposition through a stencil on all pads in the land pattern, followed by passage through a reflow oven.
The non-functional pads on a BGA component can be identified in a datasheet. For many components like application processors, most of the pads will be functional, and probably less than 10% of the pads will be reserved or non-connect. For FPGAs with BGA footprints, where the developer has freedom to assign pins and interfaces, it’s possible there are many more non-functional pads that are not broken out in a production environment.
Paste mask openings on a 0.4 mm pitch BGA.
Will Eliminating Solder Paste Affect Cost?
Any time some material is removed from assembly, it will reduce costs, but the cost reduction could be insignificant in terms of the total product development and production costs. This is a matter of volume and rework, rather than a matter of simply costing the unused solder paste. Only a small amount of solder paste will be used to print pads in a land pattern, but this scales up to non-negligible costs when manufacturing something like a consumer product at high volume.
To reduce costs in high-volume applications, there is a temptation to eliminate the solder paste from the non-functional pads. To remove solder from these pads, the apertures for these pads would be eliminated from the solder paste stencil. This would be done in the PCB footprint for the component in question, or by modifying the Gerber data to remove mask openings from the paste mask layer.
Don’t Eliminate Paste on BGA Pads
Although there is a cost reduction motivation to remove solder paste on some pads at high volume, the non-functional pads should still be printed with solder paste just like functional pads. The additional solder cost does add up, but the investment in additional solder paste is insignificant compared to the cost of rework if the assembly fails during reflow cycles. Even worse, the product might fail in the field, leading to a very expensive recall and additional liabilities for the business.
There are two major reason to keep solder paste on all pads in the BGA footprint:
1. Electrical vs. Thermal Function
Although a pad in a BGA package may not be electrically functional, it could still be thermally functional. Depending on the packaging, the pads may be attached directly to the die inside the package (e.g., wafer-level chip-scale package). These pads could still be thermally functional and they can draw heat from the die into the PCB substrate. Sufficient solder application with solder paste ensures there is a low-thermal resistance path for heat pulled from the die and into the PCB substrate.
2. Stress Concentration
When the component is stressed, either from heating the board or if the component is impacted, stress will concentrate in the solder joints. Insufficient solder from elimination of paste mask leads to physically smaller joints or weak joints on non-functional BGA pads. The result is greater stress concentration and higher likelihood of failure under repeated vibrational stress, mechanical shock, or temperature cycling. Applying solder on all pads ensures applied stress is maximally distributed across all solder joints in the BGA.
Stress concentration in the thinnest region of a solder ball on a BGA. If the solder ball is thin in the neck region, stress concentration can lead to fatigue and fracture. [Image source]
Summary
In short, if you’ve applied IPC-standardized pad size guidelines in the PCB footprint for the BGA, some solder paste will be needed on all pads to ensure reliable assembly. Make sure your footprints include sufficient paste mask opening so that strong solder joints are formed in the BGA during PCB assembly.
When you need to define assembly requirements in your PCB footprints when preparing for high-volume production, use the complete set of design tools in OrCAD from Cadence to build your circuit board. OrCAD includes the industry’s best PCB design and analysis software, complete with a set of schematic capture features, mixed-signal simulations in PSpice, and powerful CAD features, and much more.
Subscribe to our newsletter for the latest updates. If you’re looking to learn more about how Cadence has the solution for you, talk to our team of experts.