Xilinx provides a CFR solution for common wireless standards that is low cost, flexible, low power and solves your time to market needs.
Crest Factor Reduction (CFR) is one of the most fundamental building blocks in wireless communication systems today. CFR is used to reduce the dynamic range of a transmitted signal so that amplifiers being used to transmit that signal can be run with less back off. In a typical commercial wireless system, this would be used in combination with Digital Up Conversion (DUC) and Digital Pre-Distortion (DPD). There are many different algorithms for CFR, from Peak Windowing (PW) and Noise Shaping (NS) to this method of Peak Cancellation (PC), which gives rise to the product name PC-CFR. This method is much more flexible than other methods, while being higher performance resulting in greater Peak to Average Ratio (PAR) reduction and yet extremely small and cost effective in its implementation.
The PC-CFR core reduces implementation time by providing a high performance CFR solution to customers as a parameterizable core, rather than one that needs design by hand to potentially meet a multitude of different wireless standards and performance criteria.
PC-CFR supports the following air interface standards:
- MC-GSM (including frequency hopping)
Key Features and Benefits
- Support for multiple air interface standards
- Smart Peak Processing mode for supporting wide transmit bandwidth up to 400 MHz, processes incoming samples at >1.2 times instantaneous Bandwidth (iBW) reducing resource utilization
- User selectable carrier configuration agnostic Window Crest Factor Reduction (WCFR) available as a standalone or a post processing stage
- Support for power and frequency dynamics
- Support for dynamic computation of Cancellation Pulse (CP)
- Support for optional hard clipper in post processing stage
- Meets performance requirements (EVM, PAPR and ACLR) of all air interfaces
- Configurable clock-to-sample ratio of 1, 2, 3, 4 and 8
- Configurable number of Cancellation Pulse Generators (CPGs) of 1 to 12 per iteration
- Support for 1, 2, 4, 8, and 16 antennas
- Support for 1 to 8 iterations