With the recent announcement of Avago's agreement to acquire Broadcom, and other M&A activity in the semiconductor market, my first thought is focused on the increased test capacity management challenges these new organizations most assuredly will face. I could be biased in emphasizing this point, because I'm always thinking about test capacity management at, Chip Nexus, but I'm also certain that the expected acquisition synergies related to manufacturing and operations may prove more difficult to obtain when it comes to test.

One reason for this reality is the nature of test capacity. While two previously separate entities may share a common foundry and the same EDA software, the test platforms and configurations (e.g., options and licensing) used are all but guaranteed to be different. That's because test capacity is specified by the chip designer (e.g., the "test specifier") and not the test provider.

Internal test providers might have more influence over such platforms and configurations than that of external test providers—such as OSAT (Outsourced Semiconductor Assembly and Test) companies—but both largely install test capacity that is specified to them. Even when two entities use the same tester platform and model, lower-level configuration and test-cell differences can lead to gaps and surplus in capacity that can result in costly interrupted production and idle equipment, respectively. For example, Company A and Company B might use the same model XYX having 512 pins, but Company A requires 100 Mbps speed and 8 Mbytes vector memory, and Company B requires 200 Mbps speed and 16 Mbytes vector memory. Without a way to discern and manage (e.g., specify, match, plan, etc.) these differences, the combined Company AB can experience issues related to overbuying, underutilization, and more. As suggested in this example, such issues aren't always more visible system level, but can be at the option or licensing level, yet still substantially impact margins over multiple test cells.

Merging companies often find the blending of test capacity at such low levels difficult because of a lack of common or compatible test capacity management tools. In my experience, companies indeed have some way of representing their test capacity, but largely using internally developed tools. Such tools are often built on familiar and widely available platforms such as Microsoft Excel, yet are limited in addressing the challenges related to the nuances of test capacity. Perhaps more relevant to this topic, however, no two implementations of an in-house Excel-based "configuration tracker" are the same. Different worksheet structures, formatting, taxonomies, and more are inevitable. Table 1 two real implementations of configuration trackers in Excel that might represent the Company A config (left) and the Company B config (right) mentioned earlier.

Table 1. No two implementations of in-house tester configuration trackers are the same.
Table 1. No two implementations of in-house tester configuration trackers are the same.
When two or more such disparate implementations intermingle, these tools no longer seem familiar and comfortable. This, of course, can happen when combining companies and operations in an acquisition, but unfortunately, also occurs every day within companies, and certainly between different companies in the test ecosystem (e.g., the fabless specifier and the OSAT provider).

Awareness of at least these two test-capacity issues can better prepare M&A parties for a more successful manufacturing-test integration. The parties must work together to not only deploy a common test capacity management platform, but can manage capacity at the lowest levels of the test-cell configuration. Including these ideas in the M&A integration plan can ensure that the expected synergies in test operations can be readily obtained.