Turning the Page on Co-Planarity
The industry is turning the page on co-planarity and what it’s finding is a host of new issues. Some are brought on by the fact that BGA technology is now at the 0.3 and 0.25mm ultra-fine pitch stage and requires a renewed and critical look at surface finishes. Other co-planarity issues deal with BGA placement, stencil design, and PCB design.
There’s no great secret that the BGA world is constantly evolving. However, at this point in time, it’s rapidly accelerating, bringing the industry new PCB assembly challenges with 0.3mm and 0.25mm ultra-fine pitch devices. When the technology gets to this stage, the balls on a micro BGA are extremely tiny.
What this says is you’ve got to have a remarkably perfect board surface to effectively assemble BGA and micro BGA devices on a PCB. Starting with BGA placement, of utmost importance is the fact you’ve got to have the most accurate pick and place system on your SMT line. If the PCB surface is not 100 percent uniform or non-planar, a perfect BGA ball collapse won’t occur, opening the door to such placement errors as open solder joints or cold solders.Some of those tiny BGA balls will have more solder compared to others, as shown in Fig. 1. Those with more solder will likely not properly collapse, thereby creating dullness of the joint and/or a cold solder at that particular joint. This can cause elusive BGA intermittent connections and can cause latent field defects and failures, which can be extremely costly.
The industry may be already facing co-planarity issues like this in many instances largely because we’ve gotten comfortable using HASL as the surface finish of choice. HASL has a number of important benefits both to the contract manufacturer/EMS Provider and to the OEMs. Those benefits played especially well at the BGA’s 1.0- and 0.88mm pitch stage.
But now at 0.3- and 0.25mm ultra-fine pitch, it’s a different story. Keep in mind that at 1.0 and 0.8mm pitch, there’s a bit more real estate to work with and HASL or hot air solder leveling PCB surface finish worked perfectly fine. That process involves immersing the PCBs in a tin/lead alloy and then forced hot air sweeps across the board.
Solder gets deposited on the surface mount pads, but there’s no uniform process. Some of the pads as well as some of the balls on the BGA have a bit more solder on them compared to others, hence no uniformity on pad surfaces, but since BGA pads are bigger at the 1.0 and 0.8mm pitch, there is sufficient room to get a good ball collapse and BGA assembly. Even if conditions are not 100 percent perfect, BGA balls still align properly because of surface tension, and they get soldered accurately.
But the luxury of all that extra real estate is no longer available with 0.3 and 0.25mm ultra-fine pitch. That means it becomes critical to take a closer look at optional surface finishes that provide greater co-planarity to safeguard against probable assembly issues.
The other three more common options are electroless-nickel immersion gold (ENIG), immersion silver, and organic solder preservative (OSP). All three are much friendlier and co-planar compared to HASL. But of that trio, ENIG is the favorite candidate. It makes the BGA ball array to properly collapse due to its surface uniformity and because it’s electroless in nature. This means gold is uniformly deposited on the SMT and BGA pads, thus completely removing the co-planarity challenge.
ENIG makes BGA rework easy as well. Surface uniformity and proper ball collapse can be verified by viewing x-ray images of the collapsed balls. Also, by using a side-viewing device such as an Ersascope, you can ensure that the first two, three or four columns are properly collapsed. This methodology doesn’t give 100% assurance of complete ball collapse across the entire BGA site, but it could be used a good process indicator to verify the integrity of the BGA connection.
On the other hand, OSP is not easy to rework. It can only withstand a couple of thermal profiles or heat cycles. Afterward, this surface finish starts peeling off. The downside to immersion silver is it poses an oxidation issue if it’s on the shelf too long without proper storage and moisture sensitivity controls.
Meanwhile, the other areas presenting co-planarity problems associated with BGAs relate to stencil, and PCB designs. Let’s first look at board design. Warpage or bow and twist can adversely affect co-planarity with the board as well as with a BGA device itself if the BGA package is plastic. Board warpage can occur for a multitude of reasons. However, a couple of the more prevalent are poorly designed ground and power plane stack up and uneven copper distribution on the surface of the board.
Also, warpage generated co-planarity issues can result in improperly soldered BGA balls. Some of those balls may be tied to critical power or ground signals, for example. In some extreme cases, especially with bow and twist, reflow heat may not transfer uniformly across all the balls in a section of the BGA, resulting in shorts between the balls.
Cutting a perfect stencil is another key step in curtailing co-planarity problems. Three aspects involved here are accurate stencil thickness, accurate aperture definition, and stencil material. All this means that the contract manufacturer or EMS Provider must have a tribal knowledge of how to cut a perfect stencil. Plus, it’s important to have a solid handle on what triggers board warpage or bow and twist, as well as a top notch familiarity with thermal profiling, BGA device placement, and reflow process.
In most cases, BGA placement passes the x-ray text when co-planarity issues are properly dealt with ahead of time. But in the event cold solders are found, re-balling an entire device is performed. At the same time, caution must be taken to assure components like decoupling capacitors around the BGA aren’t adversely affected since considerable heat is applied to de-populate BGA device.
At this point in the evolution of BGA technology, it’s a good idea to not get complacent with HASL and other co-planarity targets of possible misfortune. It’s best to re-evaluate current practices to avoid those issues en route to successful PCB assembly.