Research Snapshot: Semiconductor Programming
In addition to our analytical marketing skills, clients also value the ability to see and interpret for themselves the unfiltered comments discovered during research efforts with both existing and potential customers. Product developers and engineering managers often find new insights – little nuggets of qualitative gold – in our reports.
Here's one example from the more than 250 projects we’ve done in technical B2B markets:
Here's one example from the more than 250 projects we’ve done in technical B2B markets:
Managers of semiconductor programming centers on new trends they saw coming in late-1990’s.
(Context: phone interviews)
(Context: phone interviews)
|
Upcoming Trend: “For us, it’ll be parallel programming of FPGAs, where we do 4 or 8 at a time, in an automated way. And we want something that can do logic, not just memory.” Driver Behind This Trend> “Logic devices are most of what we do. Logic pins are so sensitive on fine pitch devices – mostly QFPs, PGAs and FPGAs. They’re not rugged packages. Plus, logic parts are so much more expensive, we can’t afford to damage them, compared to memory. Most memories are TSOPs, which are fairly rugged packages, so we can use a manual load-on memory and do parallel programming because we don’t destroy them with manual handling.” Semiconductor Programming Manager #1
Suggested Product Improvements: “One of the biggest things to me is to increase the longevity of the sockets. They have an end of useful life – where you start to lose signals when programming – so you need to track how many insertions you’ve made for each socket to determine its overall usage. And it’d be nice to have a timer or counter or something that tracks the socket usage to know when to replace the sockets, rather than watch for bad product that comes out due to weak socket signals. And we’d like all programmers to come with software that can run in a Windows environment. Everything will be networked on Windows.” Semiconductor Programming Manager #2 Changes in Near Term: “We will be doing on-board programming only of fine pitch devices. We’re in the process of migrating over now. We plan to use HP 3170 in-circuit programmers to perform those functions. It will do both test and programming for us.” Main reason for rejecting conventional programmer/handler set-up and instead switching to HP 3170s is “co-planarity concerns. The sensitivity of the fine pitch devices are driving the decision. We’re real worried about the leads getting bent even with the best handlers.” This site currently owns seven HP 3170s that are underutilized, but plans to purchase another seven (for total of 14) to meet fine pitch device volumes of 500,000 annually by 1997. This site currently programs approximately 20,000 fine pitch devices annually. Only in-circuit programmers and testers were considered for programming fine pitch devices. Brands that were considered and rejected: Teradyne, Lawrence, Schlumberger. Semiconductor Programming Manager #3 |
