6 Summary

The differentiation provided by TI’s Control Law Accelerator (CLA) enables efficient execution of concurrent control loops on C2000 devices. CLA has been specially designed to boost the performance of control intensive math routines on real-time MCUs. The low-latency task-driven architecture of CLA is quite unique and reduces the sample-to-output delay, which is very critical for control applications. The direct access to key control peripherals and powerful floating-point processing capability allows CLA to offload the control tasks completely from the main CPU (C28x) thus freeing up its bandwidth to perform other system tasks. The CLA offers additional processing capabilities to C2000 devices and increases the overall device performance. The phase-shifting mechanism for scheduling CLA tasks as discussed in this report can be used to extract the maximum processing bandwidth out of the device. Another key benefit of the CLA over hardware-based control law implementations is flexibility. The CLA is fully programmable where developers can freely modify their control system without the time and high cost required to redesign a hardware-based solution. The C2000 C compiler [5] allows CLA to be programmed in C language similar to the C28x, which makes it very convenient to port existing algorithms or develop newer ones on CLA. The various software examples discussed in this application report demonstrates the key capabilities of CLA and can be used as a reference to adopt these unique features of CLA in their applications. These examples are very easy to use and do not require any special hardware platform other than the standard TI ControlCard to explore and evaluate the performance of CLA. Along with these examples, various Digital Power SDK [6] solutions also showcase the usage of CLA to reduce the overall C28x burden in various digital power solutions [7]. For further details on the software development and debugging with CLA, see [8].