Systolic IIR decimation filter
AssigneeMotorola, Inc.
InventorTim A. Williams
Life of this patent
Estimating the term…
Abstract
A systolic IIR decimation filter for use with an oversampled A/D converter. The filter includes a pole section having coefficients selected to generate complex pole pairs positioned to shape the filter output. A full multiplication of the coefficients is decomposed into an equivalent number of digital shifts and adds in accordance with a modified Booth's algorithm. The filter further includes a zero section for cancelling unwanted pole pairs produced by the decomposition and delays required for systolic operation. At least portions of the pole section are utilized in the zero section during times when redundant information is present in order to reduce equipment and provide decimation.
Claims11
A systolic inifinite impulse response (IIR) decimation filter having an input terminal for receiving an oversampled input signal which includes redundant information, said oversampled input signal having a bandwidth substantially larger than a bandwidth of an input signal, the filter comprising:
- a pole section defining at least first and second coefficients and coupled to the input terminal, the pole section multiplying the first and second coefficients by the received oversampled input signal to generate complex pole pairs positioned to shape the filter's output, at least one of the first and second coefficients and the multiplication of the received sampled input signal thereby, being decomposed into a substantially equivalent plurality of digital shifts of the sampled input signal and an addition of the plurality of shifts, said pole section also providing partially filtered output data;
- and
- a zero section formed from at least portions of said pole section and coupled to receive input signals from said pole section during the time that the redundant information is present in the pole section, said zero section providing a plurality of coefficients for multiplying with the partially filtered output data from said pole section to create zeros to compensate for unwanted poles produced by delays in said pole section caused by the decomposition of at least one of the first and second coefficients.
A method of systolicly filtering an oversampled output signal provided by an oversampling analog-to-digital (A/D) converter, comprising the steps of:
- selecting a plurality of first coefficients which, when multiplied by the oversampled output signal, will generate complex pole pairs that provide a predetermined shaping of the A/D converter output signal;
- decomposing at least one of the selected plurality of first coefficients into a substantially equivalent plurality of digital shifts of the A/D converter's output signal and adding the plurality of digital shifts;
- multiplying the A/D converter's output signal and the selected first coefficients systolicly by delaying the A/D converter's output signal and multiplied signals between shifts sufficiently to allow a continuous input of the A/D converter output signal, unwanted pole pairs being generated by the delaying of the A/D converter output signal and multiplied signals during multiplying;
- selecting a plurality of second coefficients which, when multiplied by selected samples of the A/D converter output signal and multiplied signals, will generate zeros that effectively cancel the unwanted pole pairs;
- and
- systolicly multiplying the plurality of second coefficients and the selected samples of the A/D converter output signal and multiplied signals during times that redundant information in the A/D converter output signal and the multiplied signals is present.
A method of constructing an Inifinite Impulse Response (IIR) decimation filter having an input terminal for receiving an input signal from an oversampled analog-to-digital (A/D) converter and the like which includes redundant information, comprising the steps of:
- deriving an apparatus for performing multiplies, additions and delays generally in the form ##EQU5## where:
- D=Z.sup.-1 =one delay,
- D2=two delays;
- selecting coefficients a1 and a2 to generate complex pole pairs that provide a predetermined shaping of the input signal from the converter;
- decomposing, in accordance with a predetermined algorithm, at least one of a1D and a2D2 in the derived apparatus into a substantially equivalent plurality of digital shifts and addition of the plurality of digital shifts by providing a second apparatus including the digital shifts, addition and delays required for systolic operation, the additional required delays generating unwanted pole pairs;
- and
- selecting coefficients b0, b1 and b2 to generate zeros that substantially cance the unwanted pole pairs.
Dependencies are inferred from the claim text.