Can large and small devices be processed in harmony?

It never fails. Every time I deliver a presentation about technologies and developments our company is working on, I get the same question: What is your solution for heterogeneous assembly? Clearly, this is a top-of-mind issue for the majority of manufacturers today; while some less-than-optimum methods are in use, some alternative technologies are delivering very good results, with even more promising developments on the horizon.

Considering the pace at which component footprints have become smaller, device pitches more dense, and the differential between the smallest and largest components wider, it’s no wonder heterogeneous assembly is the hot topic at many assemblers. The issue wasn’t so troublesome when the smallest component on the board was an 0402, but with 0201s and 0.4 mm CSPs entering the mainstream, and 01005s and 0.3 mm CSPs close behind, accommodating the various requirements and sizes has become a real challenge. The most common approaches today will not be the most cost-effective or robust solutions going forward.

Currently, manufacturers dealing with the extremes of heterogeneous assembly are following one of two routes: a stencil compromise or two printers to accommodate the variation in component sizes and paste volume requirements. With the former, manufacturers try to select a stencil thickness somewhere in between the optimum thickness for the smallest and largest components. So, generally speaking, one would require a metal thickness of approximately 0.003" to 0.004" for the smaller components and somewhere in the range of 0.006" or 0.007" for larger ones. The midrange that many firms use is a stencil thickness of about 0.005" for standard stainless steel or 0.0055" for electroform.

Unfortunately, this so-called solution really doesn’t work. You aren’t getting the best for either. The process isn’t in control, and the outcome is poor quality assemblies and less-than-ideal yields. Inferior results such as drop-test failures already are being seen with the current component mix, but imagine what will happen when we pair an 0201 – or smaller still, an 01005 – with a very large connector and attempt a stencil compromise. Not a good scenario, to be sure.

Connectors need a big, thick stencil to get a lot of material down; provided the openings are large, the result will be a good deposit and the paste volume required to hold that large component in place. But attempt to put smaller apertures on a really thick stencil, and the area ratio will be pushed way out of control, and suddenly the deposits of solder paste on the small pads are quite unrepeatable. Logically, the reverse is true if a thin stencil is used; there will not be enough material to hold the larger components in place.

The second option is to use two printers for each set of components. The smaller device’s pads are printed with a thin stencil – approximately 0.003" thick – and a second printer is set up with a 0.006" or 0.007" stencil with a step etch underneath that permits the stencil to go over the already printed paste. Again, this is less than ideal, as it results in two tool sets to manufacture a single product.

Since neither of the current approaches is the most robust or cost-effective, our company has been researching stencil technologies that will permit the smallest and largest of devices to be processed with a single stroke, thus eliminating yield issues associated with the stencil compromise, and the cost and throughput drawbacks inherent with a second tool set. To date, we have had great success applying the expertise learned with plastic stencils for adhesive printing to heterogeneous assembly print processes (see “How Mass Imaging is Changing Dispensing,” Circuits Assembly, January 2005).

More recently, we’ve learned that using a similar stencil to print solder paste can deliver excellent results: appropriate paste volumes for both large and small components. With this stencil technology, apertures as small as 200 µm and as large as 3 mm can be used. The process is repeatable, extremely efficient and in control.

It also warrants mentioning ongoing research and development of alternative metal stencil technologies for heterogeneous assembly. These include a range of options regarding nano-coated stencils that enable thicker stencils to be used without sacrificing deposit volumes or repeatability for the smaller devices. For now, however, using a modified plastic stencil is the most viable and proven solution.

So, it is certainly possible for small and large devices to be processed harmoniously and without compromising quality, UPH or expense.

Clive Ashmore is global applied process engineering manager at DEK (dek.com);cashmore@dek.com. His column appears bimonthly.