Posted by Brad Quinton, Tektronix

Signal Processing Techniques

Article

Modern ASICs and FPGAs are tedious and time-consuming to verify and validate. Adding small, highly efficient on-chip capture infrastructure to the design makes this job much easier by providing 10x the visibility of instrumentation points for a given area. In addition, by using compression algorithms, logic analysis capture stations can capture data for 10x or more capture depth.ASICs and FPGAs have become massively complex, particularly for System-on-Chip (SoC) designs involving multiple cores. With this complexity comes longer and more tedious debug and validation cycles. Unfortunately, when something fails or goes wrong, gaining access to test points in highly integrated designs is next to impossible. Unless you want to spend weeks shooting in the dark at random errors while running through multiple prototypes, on-chip instrumentation is no longer optional; it’s a critical must-have. Figure 1 shows an overview of the debug process using on-chip instrumentation.

Figure 1: The instrumentation and debug cycle is critical to detect errors in FPGA prototypes.(Click graphic to zoom by 1.9x)While there are a number of ways to add instrumentation to FPGAs, a distributed approach using an instrument network is emerging as the preferred method, as it maximizes the number of potential observation points while minimizing silicon area or look-up table utilization requirements. Also critical to efficient debug is deep trace capture to see how the various parts of a system interact over time. Finally, designers must be able to observe the interactions of multiple devices and clock domains, both on- and off-chip, all fully time correlated for a true system-level perspective.[Continue reading →]

In a New Tab/Window

 Expand All