Adjustable digital filter for high speed digital transmission
AssigneeRockwell International Corporation
InventorMichael A. Caloyannides
Life of this patent
Estimating the term…
Abstract
An adjustable digital filter combines the functions of a conventional digital filter and a conventional equalizer and permits elimination of the additional delay stages conventionally required for the equalizer. Reduced complexity and expense are attained, along with improved performance, since the elimination of "delay-and-add" stages required by prior art equalizers permits detection and correction of error with less delay than the prior art. Hence, improved performance is realized along with reduced cost.
Claims30
An adjustable digital filter for use in a digital data transmission and receiving system for digitally filtering and equalizing a received data signal supplied to the filter as digital data samples x.sub.j at twice the baud rate of data transmission and including successive, alternating integer-baud and half-baud samples, comprising:
- a plurality of successive delay stages for storing successive said digital data samples,
- a corresponding plurality of adjustable tap gain means respectively associated with said plurality of delay stages, each receiving the samples currently stored in the respectively associated delay stages and a corresponding tap gain control as inputs thereto, and producing as an output thereof the product of the inputs thereto,
- means for summing the outputs of said tap gain means and producing a filtered and equalized output sample y.sub.j in each baud interval for the corresponding integer baud sample x.sub.j,
- means responsive to each sample y.sub. j for generating a corresponding estimated value l.sub.o D.sub.j representing an estimate of the value of a correctly equalized and filtered output sample y.sub.j,
- means responsive to said summing means and also responsive to said estimated value generating means for generating an error signal Y.sub.j = y.sub.j - l.sub.o D.sub.j,
- first means for adjusting the tap gain control inputs to each of the adjustable tap gain means associated with said delay stages storing integer baud samples of said succession of samples, as a function of said error signal Y.sub.j and said digital data samples x.sub.j, and
- second means for adjusting the tap gain control inputs to each adjustable tap gain means associated with a delay stage storing half-integer baud samples of said succession of samples as a function of the adjustment by said first adjusting means to the tap gain control input to the adjustable tap gain means associated with at least one integer baud sample storing delay stage respectively corresponding thereto.
An adjustable digital filter for use in a digital data transmission and receiving system for digitally filtering and equalizing a received data signal supplied to the filter as digital data samples x.sub.j at twice the baud rate of data transmission and including successive, alternating integer-baud and half-baud samples, comprising:
- means including a plurality of (2N- 1) successive delay stages for storing (2N- 1) successive ones of said digital data samples x.sub.j,
- a corresponding plurality of adjustable tap gain means respectively associated with said plurality of delay stages, each receiving the sample currently stored in the respectively associated delay stage and a corresponding tap gain control as inputs thereto, and producing as an output thereof the product of the inputs thereto,
- means for summing the outputs of said tap gain means and producing a filtered and equalized output sample y.sub.j in each baud interval for the corresponding integer-baud sample x.sub.j,
- means responsive to each sample y.sub.j for generating a corresponding, estimated value l.sub.o D.sub.j,
- means responsive to said summing means and also responsive to said estimated value generating means for generating an error signal Y.sub.j = y.sub.j - l.sub.o D.sub.j where l.sub.o D.sub.j represents an estimate of the value of the correctly equalized and filtered output sample y.sub.j,
- means for generating tap gain control inputs for adjusting the said adjustable tap gain means for each said sample x.sub.j in accordance with
- 0 .congruent. E.sub.j {Y(j) x(j+ i)}
- where i= 0, 1, 2, . . . (2N- 1), and identifies the tap gain means (i) of the plurality (2N- 1) thereof. y(j) - y(j).]
Dependencies are inferred from the claim text.