This blog is divided into two parts. The first deals with the general principles behind process optimization, the second gives some practical examples of the approaches Palomar Assembly Services has successfully negotiated for its customers, saving them both time and money.
A good metaphor to better understanding the challenges in approaching prototyping and development in semiconductor packaging, as well as other changes in process improvement, is to think in terms of building a mechanical clock. You know exactly what the outcome needs to be—a device that accurately and consistently keeps time—it is just a question of how you get there. In other words, identifying the right cogs, and placing them in the needed relationship to one another, needs, in a ruthlessly competitive world, to be accomplished as swiftly as possible.
The skills needed to do this, are a combination of experience, familiarity, understanding the principles of physics, and intelligent patience in trial and error. It is often when unforeseen errors arise, despite the most meticulous planning, that the greatest challenges to efficient solution appear their most remote.
Figure 1. Cost Benefit Analysis
This is particularly an acute problem for SMEs working in our sector. Taking time to physically try out various potential solutions, means diverting people, machines, or both from more immediate, revenue generating activities. As a result, the longer it takes to resolve challenges, the more pressures begin to build elsewhere. If the package is totally or largely new, the engineer will have the additional need to learn and experiment simultaneously. This means less money coming in from elsewhere.
Fortunately, many of these technical factors can be distilled into financial calculations, what in Economics is known as “Cost Benefit Analysis”. Uniformly part of business practice for well over 150 years, we set out the principles in diagrammatic form in Figure 1. The reader will immediately note that CBA (cost benefit analysis) explicitly labels, then includes, all the holistic elements contributing to our “clock building” exercise: the estimable (man hours, system hours, materials and so on), contingencies, business contingencies such as the extra time and cost when unforeseen problems are incurred, technical contingencies, such as changing parameters, material interactions, tooling quality, software, hardware irregularities, yield variations, and so on. Finally, there are intangibles, such as customer or employee frustration. The chart also makes it clear that the goal of any process development work must be to reduce the potential size of contingencies and intangibles as much as is humanly possible.
At Palomar our Assembly Service managers are trained both in using this principle internally and in taking potential customers step by step in providing quotes for process work they are considering outsourcing to us. These conversations usually center on an appropriate level of investigation. Armed with data, this then allows the customer to see if it pays them to use these services. Quite simply, our collective experience totaling many decades allows us to have a much firmer grasp on intangibles and contingencies than many of our customers. This is further enhanced by the fact that we operate in a dedicated development environment using dedicated, specialist personnel. This is what makes our quotes both competitive, accurate, and—above all—the outcomes so successful. If they were not, it would mean us losing money, not the customer.
Figure 2. Diagrammatic Representation of Improving Yield in Large, Complex Packaging
In what follows below, we give an example of a process we undertook requiring both die and wire bonding. More explicitly: we review the product, the particular challenge, the approach taken in technical negotiation, and finally, some generalizations that may help the reader in honing their approach.
Figure 2 represents the demands placed on us to populate a single PCB with over 200 LEDs set out in a matrix format of seven rows and 36 columns at a pitch of 400 microns. Such stipulations put severe burden on process accuracy, repeatability, and discreet orientation. We utilized a system with 1.5 micron accuracy at 3 sigma. Then with eight arrays, we set reference fiducial markers for axis and origin, supplemented by the use of radar to locate the LEDs, as well as standard then high magnification cameras to orient, place, and confirm the LED. The pathway of contingency to be negotiated related to planarity, speed, and force of placement, together with daubing (epoxy dots), a process suited to repeatability and accuracy within epoxy tolerances of less than 25 microns. On account of the small pad size and the complexity of the layout, we reference every single LED twice. This underscores the importance of getting it right so as not frustrate the next step.
Figure 3. Wedge-bond Loop Challenges
That next stage is sequential bonding and looping, which is undertaken within very tight pitch parameters. For this reason, consistency from die to die is essential, as is the ability to quickly spot irregularities and assess the degree of deviance. The perpendicular and horizontal loops cannot touch either each other or anything else in the package, such as a lens or lid. Figure 3 sets out in more detail some of the constraints that need to be negotiated.
It is important to state this is part of a wider, holistic approach. Such elements include a close inspection of components and materials to ensure they are of the standard required. We then consider diagrams and statement of work against the backdrop of our experience. After calibrating our system, we will then create a test vehicle to hasten development of a stable, reliable and robust process, simultaneously avoiding wasting the customer’s materials. Finally, we consult the customer one more time by way of ensuring the proposed solution meets all requirements before we begin actual prototyping or small-scale manufacturing.
If this blog has sparked some interest, you may wish to follow up by viewing our recent webinar. Alternatively, if you have more detailed and specific questions arising from what we have said here, please contact us. Remember, optimizing packaging solutions is what we enjoy doing!