No abstract

Three different approaches to FDM demultiplexing applying digital signal processing are investigated for L0 = 32 slot signals: The single stage method (SSM), the polyphase method (PPM), and the hierarchical multistage method (HMM). All three approaches apply bandpass sampling and are based on the processing of complex signals by linear‐phase FIR filters, where the respective signals are always oversampled by a factor of two at any stage of processing. For the three methods all relevant design results are given including the required fixed‐point signal and coefficient wordlengths. The HMM is the most efficient of the above approaches in terms of (bit) operation rate per channel, immediately followed by the PPM, while the SSM is less efficient by a factor of ten. Additional advantages claimed for the HMM are its high degree of modularity with the potential of optimization, ease of manufacturing and testing, suitability for VLSI realization, high reliability, and reduced amount of overhead circuitry. This paper is published in two parts. Part I comprises the introduction, the front end issues common to all three approaches to FDM demultiplexing and the treatment of the SSM. In Part II the work on the PPM and the HMM is reported. Furthermore, all three approaches are compared with each other and a conclusion is given.

No abstract

Three different approaches to FDM demultiplexing applying digital signal processing are investigated for L0 = 32 slot signals: The single stage method (SSM), the polyphase method (PPM), and the hierarchical multistage method (HMM). All three approaches apply bandpass sampling and are based on the processing of complex signals by linear‐phase FIR filters, where the respective signals are always oversampled by a factor of two at any stage of processing. For the three methods all relevant design results are given including the required fixed‐point signal and coefficient wordlengths. The HMM is the most efficient of the above approaches in terms of (bit) operation rate per channel, immediately followed by the PPM, while the SSM is less efficient by a factor of ten. Additional advantages claimed for the HMM are its high degree of modularity with the potential of optimization, ease of manufacturing and testing, suitability for VLSI realization, high reliability, and reduced amount of overhead circuitry. This paper is published in two parts. Part I comprises the introduction, the front end issues common to all three approaches to FDM demultiplexing and the treatment of the SSM. In Part II the work on the PPM and the HMM is reported. Furthermore, all three approaches are compared with each other and a conclusion is given.

No abstract