The design community has a love/hate relationship with interface logic. Many designers would like to see it incorporated in ASICs or replaced by some type of programmable gate array at the start of every design. This desire to eliminate interface logic is a noble effort. But logic is needed to resolve many unforeseen design issues.
As long as designers need to interface two ICs with different types of I/Os or different voltage levels, or maintain signal integrity in a system, interface logic will never die. As long as a designer must find a quick, low-cost solution to fix a mistake in an ASIC, or cut costs, logic will never die. As long as a designer needs a system solution that requires a combination of high power and drive found only in bipolar/MOSFET circuit structures along with low-power CMOS performance, logic will never die.
Interface logic grows and prospers because every application requires some kind of power and logic interface. The hot spot for interface logic is still bridging the gap between the multitude of ASICs used in everything from complex high-end systems to low-cost consumer products. Increasing microprocessor speeds translate into a need for faster communication buses and memories, with more information to be bundled, accessed, shared, and displayed. The result is a migration to smaller process geometries that require lower core operating voltages. The biggest burden associated with this migration to next-generation performance is the need to be backward-compatible with legacy system ICs and architectures. This coexistence of high-performance microprocessor cores and ASICs with legacy peripheral ICs requires logic Interface solutions.
A BRIGHT FUTURE FOR LOGIC
Interface logic helps designers overcome a variety of signal-integrity, signal-distribution, and chip-to-chip interface problems.
Logic designers drive heavily loaded signal pathways by extending the signal's quality with a logic driver or buffer. Some designs require a way to switch data among different signal paths without introducing additional jitter to mission critical datapaths. A zero-delay bus switch solves this challenge. Distribution of clocking signals is accomplished in fan-out buffers and via multiplexer/demultiplexer logic. Numerous systems today require these types of logic interface ICs.
Also, many system platforms today require designers to resolve chip-to-chip signal-compatibility issues associated with I/O and signal-level translation. For instance, many designs need interfaces between 5- and 3-V ICs or 3- and 1.8-V or below ICs.
Interface logic is necessary for memory modules and bus switching as well. Memory modules come in a number of configurations, from planar to stacked solutions. These solutions require high-speed logic capable of high drive to support the demands of current and next-generation DDR-I and DDR-II memories. In addition, bus-switch logic provides ways to support low-cost PCI-x expansion. A bus switch lets designers expand the number of memory slots or PCI-x slots in their application at a fraction of the cost of traditional ASIC expansion solutions.
On the far end of the spectrum are the SOTiny product solutions. This workhorse class of logic has emerged as the champion of ASIC logic fixes as well as a way to minimize material costs in consumer products. The SOTiny gate continues to provide the smallest and easiest logic solution used in the industry.
In the end, interface logic provides a number of great characteristics and features to help the design community overcome interface issues between next-generation and legacy systems, ensure that signal integrity is maintained, and provide cost-effective extensions to next-generation memory and PCI-x slot expansions. Logic continues to have a bright future and will always be regarded as the stopgap solution when advanced ASICs or FPGAs fail to supply the best performance-to-cost ratio. As future logic innovations emerge, designers can rest assured that logic will never die.