Sampled amplitude read channel employing an adaptive non-linear correction circuit for correcting non-linear distortions in a read signal

Abstract

A sampled amplitude read channel is disclosed for magnetic disk storage systems comprising an adaptive non-linear correction circuit for correcting non-linear distortions in the read signal, such as asymmetry caused by the non-linear response of a magneto-resistive (MR) read head. The analog read signal is sampled and the discrete time sample values equalized into a desired partial response prior to sequence detection. The non-linear correction circuit is inserted into the read path prior to the sequence detector and adaptively tuned by a least-mean-square (LMS) adaptation circuit. In one embodiment, the non-linear correction circuit is a discrete-time Volterra filter comprising a linear response for implementing an equalizing filter, and a non-linear response for attenuating non-linear distortions in the read signal. The filter coefficients of both the linear and non-linear sections of the Volterra filter are adaptively adjusted by the LMS adaptation circuit. In an alternative embodiment, the non-linear correction circuit operates in the analog domain, prior to the sampling device, where the cost and complexity can be minimized. The analog correction circuit implements an inverse response to that of the non-linearity in the read signal, and the response is adaptively tuned using an LMS update value computed in discrete-time for a Volterra filter, without actually implementing a Volterra filter. Further, the LMS update value for the analog correction circuit can be implemented using a simple squaring circuit.

Description

FIELD OF INVENTION

The present invention relates to the recording and reproduction of binary data in magnetic disk storage systems for digital computers, particularly to a sampled amplitude read channel employing an adaptive non-linear correction circuit for correcting non-linear distortions in the read signal, such as asymmetries caused by the non-linear response of a magneto-resistive (MR) read head.

BACKGROUND OF THE INVENTION

In magnetic disk drive storage devices, sampled amplitude read channels employing partial response (PR) signaling with maximum likelihood (ML) sequence detection have provided a substantial increase in storage capacity by enabling significantly higher linear bit densities. Partial response signaling refers to a particular method for transmitting symbols represented as analog pulses through a communication medium. The benefit is that at the signaling instances (baud rate) there is no intersymbol interference (ISI) from other pulses except for a controlled amount from immediately adjacent, overlapping pulses. Allowing the pulses to overlap in a controlled manner leads to an increase in the symbol rate (linear recording density) without sacrificing performance in terms of signal-to-noise ratio (SNR).

Partial response channels are characterized by the polynomials(1−D)(1+D)ⁿ where D represents a delay of one symbol period and n is an integer. For n=1,2,3, the partial response channels are referred to as PR4, EPR4 and EEPR4, with their respective frequency responses shown in FIG. 1A. The channel's dipulse response, the response to an isolated symbol, characterizes the transfer function of the system (the output for a given input). With a binary “1” bit modulating a positive dipulse response and a binary “0” bit modulating a negative dipulse response, the output of the channel is a linear combination of time shifted dipulse responses. The dipulse response for a PR4 channel (1−D²) is shown as a solid line in FIG. 1B. Notice that at the symbol instances (baud rate), the dipulse response is zero except at times t=0 and t=2. Thus, the linear combination of time shifted PR4 dipulse responses will result in zero ISI at the symbol instances except where immediately adjacent pulses overlap.

It should be apparent that the linear combination of time shifted PR4 dipulse responses will result in a channel output of +2, 0, or −2 at the symbol instances (with the dipulse samples normalized to +1, 0, −1) depending on the binary input sequence. The output of the channel can therefore be characterized as a state machine driven by the binary input sequence, and conversely, the input sequence can be estimated or demodulated by running the signal samples at the output of the channel through an “inverse” state machine. Because noise will obfuscate the signal samples, the inverse state machine is actually implemented as a trellis sequence detector which computes a most likely input sequence associated with the signal samples.

The performance of the trellis sequence detector in terms of bit error rate depends on the amount of noise in the system, including noise due to the spectrum of the read signal diverging from the ideal partial response. Linear distortions in the read signal can generally be suppressed using a linear equalizer which may operate on the continous-time analog read signal or the discrete-time samples of the read signal. Typical read channels employ both an analog equalizer, such as a biquad analog filter, followed by a nth order finite-impulse response (FIR) discrete-time filter. Linear equalizers, however, are not effective in attenuating non-linear distortions in the read signal, such as asymmetries caused by the non-linear response of a magneto-resistive (MR) read head.

An MR read head comprises an MR sensor element with a resistance which is proportional to the strength of the magnetic flux; the read signal is generated by applying a current to the MR element and measuring the voltage across it as it passes over the magnetic transitions recorded on the disk. FIG. 2A is a plot of the head's resistance versus the magnetic flux which illustrates that the response can be very non-linear. The asymmetry effect of this non-linearity on the read signal is illustrated in FIG. 2B which shows the head's response to an isolated positive and negative magnetic transition. In this example, the magnitude of the pulse induced by the positive magnetic transition is greater than the magnitude of the pulse induced by the negative magnetic transition (note that the asymmetry in the pulses may be reversed, and other asymmetries may also be present in the read signal). Ultimately, the non-linear distortions prevent the read signal from attaining the desired partial response target, introducing noise into the sample values which degrades the performance of the trellis sequence detector. This undesirable effect is generally ameliorated by applying a magnetic biasing field across the MR element so that it operates in a linear region of the response while still providing sufficient sensitivity and stability. However, biasing the MR element does not remove the non-linear aberrations from the read signal entirely, particularly when errors in the biasing field move the MR element away from the optimal linear region of operation.

There is, therefore, a need for an improved sampled amplitude read channel for use in disk storage systems that provides a non-linear correction circuit for correcting non-linear distortions in the read signal, such as asymmetries caused by the non-linear response of an MR read head.

SUMMARY OF THE INVENTION

A sampled amplitude read channel is disclosed for magnetic disk storage systems comprising an adaptive non-linear correction circuit for correcting non-linear distortions in the read signal, such as asymmetry caused by the non-linear response of a magneto-resistive (MR) read head. The analog read signal is sampled and the discrete time sample values equalized into a desired partial response prior to sequence detection. The non-linear correction circuit is inserted into the read path prior to the sequence detector and adaptively tuned by a least-mean-square (LMS) adaptation circuit. In one embodiment, the non-linear correction circuit is a discrete-time Volterra filter comprising a linear response for implementing an equalizing filter, and a non-linear response for attenuating non-linear distortions in the read signal. The filter coefficients of both the linear and non-linear sections of the Volterra filter are adaptively adjusted by the LMS adaptation circuit. In an alternative embodiment, the non-linear correction circuit operates in the analog domain, prior to the sampling device, where the cost and complexity can be minimized. The analog correction circuit implements an inverse response to that of the non-linearity in the read signal, and the response is adaptively tuned using an LMS update value computed in discrete-time for a Volterra filter, without actually implementing a Volterra filter. Further, the LMS update value for the analog correction circuit can be implemented using a simple squaring circuit.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and other aspects and advantages of the present invention will be better understood by reading the following detailed description of the invention in conjunction with the drawings, wherein:

FIG. 1A shows the frequency response for a PR4, EPR4 and EEPR4 read channel.

FIG. 1B shows the dipulse responses for the PR4, EPR4 and EEPR4 read channels of FIG. 1A.

FIG. 2A is a plot of the resistance of an MR read head versus magnetic flux.

FIG. 2B illustrates the asymmetry introduced into the read signal from the non-linear response of the MR read head.

FIG. 3 is a block diagram of the sampled amplitude read channel of the present invention comprising a non-linear correction circuit implemented in discrete-time as an adaptive Volterra filter updated according to an LMS algorithm.

FIG. 4 shows details of the LMS adaptation circuit for the Volterra filter of FIG. 3.

FIG. 5 is a block diagram of the sampled amplitude read channel of the present invention comprising a non-linear correction circuit implemented in the analog domain and updated using a stochastic gradient estimate generated from the LMS adaptation circuit for updating a discrete-time Volterra filter.

FIG. 6 shows further details of the LMS adaptation circuit for updating the linear equalizer and the analog correction circuit, including an orthogonal projection operation for constraining the phase delay and gain of the linear equalizer.

FIG. 7A shows details of a two-sided non-linear correction circuit implemented in the analog domain.

FIG. 7B shows details of a one-sided non-linear correction circuit implemented in the analog domain.

FIG. 8 shows the gain response of the linear equalizer of FIG. 6 constrained to a constant g at the normalized frequency of 1/4T.

FIG. 9 shows the phase response of the linear equalizer of FIG. 6 constrained to kπ at the normalized frequency of 1/4T.

FIG. 10 illustrates operation of the orthogonal projection operation for constraining the gain and phase response of the linear equalizer.

FIG. 11 illustrates the effect of the non-linear correction circuit of the present invention on a PR4 read signal.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

Overview

The sampled amplitude read channel of the present invention is intended to operate within a magnetic disk storage device as illustrated in FIG. 3. A magnetic disk 2 comprising a plurality of concentric, radially spaced data tracks is rotated under a read head (e.g., an MR read head) which transduces the magnetic transitions recorded on the disk into an analog read signal 4 comprising polarity alternating pulses. A non-linear distortion 6 is introduced into the read signal 4, such as asymmetry distortion in the pulses caused by the non-linear response of an MR read head. The read channel includes circuitry for equalizing the read signal into a desired partial response, including a variable gain amplifier (VGA) 8, an analog filter 10, and an adaptive linear discrete-time equalizer 12. Since the linear equalizer 12 cannot compensate for non-linear distortions in the read signal, the read channel further comprises an adaptive non-linear correction circuit. The non-linear correction circuit is adapted to compensate for variations in the non-linear distortions that occur over time due to variations in fly-height, radial location of the head, temperature fluctuations, and other parameter variations in the system. After equalizing the read signal according to the desired partial response and attenuating the non-linear distortions, the sample values of the analog read signal are demodulated by a trellis sequence detector 16 which computes an estimated binary sequence most likely to have generated the read signal samples.

A least-mean-square (LMS) algorithm implements the adaptation strategy for the non-linear correction circuit. In a first embodiment, the non-linear correction circuit is implemented as a nth order discrete-time Volterra filter with coefficients updated according to an estimated stochastic gradient. In an alternative embodiment, the non-linear correction circuit is implemented in the analog domain wherein the cost and complexity of the design can be significantly reduced. The analog correction circuit comprises a simple second order response of the formx+αx² which estimates the inverse response of the second order non-linearity caused by an MR read head. The LMS update for the coefficient α is estimated using the stochastic gradient computed for the center coefficient in the quadratic component of a Volterra filter, without actually implementing a Volterra filter.Discrete-Time Volterra Filter

In the embodiment of the present invention shown in FIG. 3, the non-linear correction circuit is implemented as a discrete-time, adaptive Volterra filter which operates on the sample values of the read signal at the output of the sampling device 14. The Volterra filter comprises a linear component H₁(z) 12 for performing the equalizing function in the read channel, and at least one higher order component (a second order component H₂(z₁,z₂) 18 shown in FIG. 3) for attenuating a non-linear distortion in the read signal, such as asymmetry caused by the non-linear response of an MR read head. The output of the linear component H₁(z) 12 is combined with the output of the non-linear component H₂(z₁,z₂) 18 at adder 34 to generate the equalized, linearized signal samples 32 processed by the discrete-time sequence detector 16.

The linear component H₁(z) 12 of the Volterra filter is implemented as a conventional finite-impulse-response (FIR) filter comprising a vector of coefficients C36, and the non-linear second order component H₂(z₁,z₂) 18 is implemented as an nxn matrix of coefficients C_nxm 38. The output y_k 32 of the Volterra filter can be written asy_k=C^TX_k+X_k^TC_nxnX_k where X_k is a vector of the input sample values 24. The coefficients for both the linear component H₁(z) 12 and the non-linear component H₂(z₁,z₂) 18 are updated by a least-mean-square (LMS) adaptation circuit 20 which adjusts the coefficients C36 and C_nxn 38 in a manner that minimizes the squared error e_k 22 computed as the difference 28 between the sample values y_k 32 output by the Volterra filter and estimated ideal sample values ^S_k 26 corresponding to the desired partial response. A simple slicer circuit 30 generates the estimated ideal sample values ^S_k 26 by comparing the output samples of the Volterra filter 32 to thresholds which represent the decision boundaries for the ideal partial response signal samples.

The LMS adaptation circuit 20 approximates a minimum mean-square-error (MMSE) algorithm which minimizes an expected squared error functionV(H₁,H₂)=E[(y_k−^S_k)²]. The gradient of V with respect to C_n (the nth coefficient in the linear filter H₁(z) 12) is $\frac{\partial V}{\partial C_n}=E\left[2\left(y_k-{\hat{S}}_k\right)\frac{\partial {\hat{y}}_k}{\partial C_n}\right]=2E\left[e_k{X}_{k-n}\right]$where e_k is the error value y_k−^S_k. The LMS algorithm, otherwise known as the stochastic gradient algorithm, updates the coefficient C_n by removing the expectation operator E from the above gradient equation and following the residual gradient estimate to the minimum of VC_n^(m+1)=C_n^(m)−μ2e_kX_k−n. A similar computation leads to the LMS update algorithm for the non-linear second order component H₂(z₁,z₂) 18 of the Volterra filter. The MMSE stochastic gradient equation for H₂(z₁,z₂) 18 is $\frac{\partial V}{\partial H_2\left(i,j\right)}=E\left[2e_k\frac{\partial {\hat{y}}_k}{\partial H_2\left(i,j\right)}\right]=E\left[2e_k{X}_{k-i}{X}_{k-j}\right]$which leads to an LMS stochastic gradient equation ofH₂^(m+1)(i,j)=H₂^(m)(i,j)−μ2e_kX_k−iX_k−j.

The circuitry for implementing the LMS adaptation circuit 20 for the Volterra filter of FIG. 3 is shown in FIG. 4. The discrete-time sample values X_k 24 of the read signal and the error signal e_k 22 are input into a first stochastic gradient circuit 40 for implementing the above LMS equation for the linear component H₁(z) 12 (the additional r index shown in the figure accounts for the delay introduced by H₁(z)), and a second stochastic gradient circuit 42 for implementing the above LMS equation for the non-linear second order component H₂(z₁,z₂) 18. At every sample interval, the previous coefficients C and C_nxn are loaded into the LMS update circuits 40 and 42, updated coefficients are computed, and the updated coefficients restored to the filters for use in generating the next output sample y_k 32. Alternatively, the LMS update operation can be decimated to reduce the number of multiplications and additions required at the cost of reducing the convergence speed. For details concerning a decimated implementation for an LMS adaptation strategy, see the above referenced U.S. patent entitled “GAIN AND PHASE CONSTRAINED ADAPTIVE EQUALIZING FILTER IN A SAMPLED AMPLITUDE READ CHANNEL FOR MAGNETIC RECORDING.”

The non-linear second order component H₂(z₁,z₂) 18 of the Volterra filter may be overly complex and expensive to implement depending on the design constraints for a particular storage system. For example, a second order component H₂ comprising a 5×5 matrix of coefficients requires 20 multiplies and 14 accumulates for every update. Therefore, the present invention provides an alternative embodiment which significantly reduces the cost and complexity by implementing the non-linear correction circuit in the analog domain, while still performing the LMS update procedure in discrete-time using the LMS update for a Volterra filter, without actually implementing the more complex Volterra filter. This embodiment of the present invention is discussed in detail in the following section.

Analog Non-Linear Correction Circuit

Details of the sampled amplitude read channel of the present invention employing an analog non-linear correction circuit are shown in FIG. 5. The elements shown are essentially the same as in FIG. 3 described above, except for the addition of an analog non-linear correction circuit 44 and the deletion of the second order component H₂(z₁,z₂) 18 of the Volterra filter. The analog non-linear correction circuit 44 is adaptively tuned by the LMS adaptation circuit 20 using the stochastic gradient update computed for the second order component H₂(z₁,z₂) 18 of the Volterra filter.

As with the linear component H₁(z) 12 of the Volterra equalizer, the MMSE for the coefficients of the second order component H₂(z₁,z₂) 18 is a function of the analog filter 10, the recording density, and noise spectrum and distribution, in addition to the non-linear distortion caused by MR asymmetry. However, the “center” coefficient of the nxn second order component H₂(z₁,z₂) 18, defined as the center coefficient H₂(d,d) on the diagonal of the nxn matrix, is a consistent indicator of the second order asymmetry distortion caused by the MR read head. If the center coefficient H₂(d,d) is positive, it means that the Volterra equalizer needs to add a portion of X²_k−d to ^S_k when the MR asymmetry is under-corrected. Conversely, if the center coefficient H₂(d,d) is negative, it means that the Volterra equalizer needs to subtract a portion of X²_k−d from ^S_k when the MR asymmetry is over-corrected. Therefore, the stochastic gradient update for the center coefficient H₂(d,d) can be integrated and the integral representing the MR asymmetry error can be used to update the non-linear correction circuit 44. From the above LMS equation, the stochastic gradient update 46 shown in FIG. 5 for the center coefficient H₂(d,d) is computed as $\frac{\partial V}{\partial H_2\left(d,d\right)}=\mu 2e_k{X}_{k-d}^2$The stochastic gradient update 46 is then integrated by a discrete-time accumulator 48, and the integral 50 used to adaptively tune the analog non-linear correction circuit 44.

Details of the LMS adaptation circuit 20 of FIG. 5 are shown in FIG. 6. The discrete-time sample values X_k 24 of the read signal and the error signal e_k 22 are input into a constrained stochastic gradient circuit 52 for implementing the above LMS equation for the linear component H₁(z) 12 (i.e., the linear equalizer). As explained in greater detail below, the phase response of the linear equalizer H₁(z) 12 is constrained using an orthogonal projection operation Pv₁v₂^⊥ so that the stochastic gradient update for the center coefficient H₂(d,d) can be accurately computed relative to the equalizer's group delay. The discrete-time signal samples X_k 24 are passed through a delay element 54 to generate the sample value X_k−d. The delayed sample value X_k−d is then squared 56, and the squared sample value X²_k−d multiplied 58 by the error signal e_k 22 and a gain value μ to implement the above equation $\frac{\partial V}{\partial H_2\left(d,d\right)}=\mu 2e_k{X}_{k-d}^2$

The analog correction circuit 44 is designed to approximate the inverse response of the non-linearity in the read signal, thereby cancelling the non-linear distortion. Since the non-linear response of an MR read head is dominated by a second order component, the inverse response f⁻¹(x) can be approximated as a second order polynomial of the formf⁻¹(x)=x+αx². The above inverse response is considered “two-sided” since it will add a second order compensation factor into both the positive and negative pulses in the read signal. However, the non-linearity may be biased toward distorting only the positive or negative pulses, in which case it may be better to employ a “one-sided” inverse response of the form: $f^{-1}\left(x\right)=\{\begin{array}{cc}x& x\ge 0\\ x+\alpha x^2& x<0\end{array}\mathrm{or}{f}^{-1}\left(x\right)=\{\begin{array}{cc}x+\alpha x^2& x\ge 0\\ x& x<0\end{array}$depending on whether the distortion affects the negative or positive pulses, respectively. The coefficient α is adaptively tuned using the integrated stochastic gradient 50 for the center coefficient in the second order factor of the Volterra filter, as described above, in order to minimize the deviation (the error e_k 22) of the read signal's response from the desired partial response.

The circuitry for implementing the one-sided inverse response is shown in FIG. 7A. The analog read signal 61 output by the VGA 8 is squared by analog squarer 60, and the squared signal is then scaled by the coefficient α 62 which is adaptively tuned by the integrated stochastic gradient 50 generated by the LMS adaptation circuit 20 of FIG. 6. The scaled, squared read signal is then added to itself at adder 64 to implement the above two-sided inverse response. The circuitry for implementing the one-sided inverse response is shown in FIG. 7B, which is essentially the same as that of FIG. 7A except for the addition of a comparator for comparing the analog read signal to zero at comparator 66. The result of the comparison is used as the control signal to multiplexer 68 for selecting either the unmodified analog read signal, or the analog read signal after second order compensation, as the output of the non-linear correction circuit.

The stochastic gradient estimate for the center coefficient H₂(d,d) of the Volterra filter will be an accurate estimate of the stochastic gradient for the analog non-linear correction circuit 44 only if the linear equalizer H₁(z) 12 has a known, fixed (preferably integer) delay and is close to linear phase, giving it a nearly constant group delay across frequencies. Otherwise, there will be an unkown, fractional time delay between the read signal sample X_k 24 and the error signal e_k 22 making it more difficult or impossible to synchronize the error signal e_k 22 to the appropriate signal sample X_k 24. Thus, an aspect of the present invention is to implement the linear equalizer H₁(z) 12 to have a fixed, integer group delay.

In the present invention the LMS update algorithm is modified to including an orthogonal projection operator Pv₁v₂^⊥, as seen in the LMS adaptation circuit of FIG. 6, which constrains the phase response of the linear equalizer at the preamble frequency of 1/4T, thereby fixing the phasy delay at 1/4T to an integer delay. Since the group delay of the linear filter is characteristically constant across a significant portion of the frequency spectrum as shown in FIG. 9, the orthogonal projection operator Pv₁v₂^⊥ effectively constrains the linear filter such that it exhibits a known, integer group delay with little variation across frequencies. Consequently, the equalized error signal e_k 22 can be easily synchronized to the read signal samples X_k 24 for computing the stochastic gradient estimate for the center coefficient H₂(d,d). The gain of the linear equalizer H₁(z) 12 is also constrained at the preamble frequency in order to attenuate interference from the timing recovery and gain control feedback loops which may prevent the LMS adaptation circuit from converging.

Operation of the orthogonal projection operator Pv₁v₂^⊥ will now be described in relation to the gain and phase response of the linear equalizer H₁(z) 12. Referring to FIG. 8, the gain (magnitude) response of the linear equalizer H₁(z) 12 has been constrained to a predetermined value (denoted by g) at the normalized frequency of 0.5 (1/4T). Similarly, the phase response of the linear equalizer H₁(z) 12 has been constrained to kπ at the normalized frequency of 0.5 as shown in FIG. 9. In effect, these constraints allow the gain and phase to vary (adapt) at all frequencies except at the normalized frequency of 0.5, thereby constraining the filter's frequency response in a manner that attenuates interference from the gain and timing loops. In addition, constraining the phase reasponse to kπ at the normalized frequency of 0.5 results in an integer group delay across frequencies so that the read signal samples X_k 24 can be appropriately synchronized to the error signal e_k 22 as described above. The gain constraint g is relatively arbitrary except that it is selected to optimize the dynamic range of the filter's coefficients. However, constraining the filter's response at the normalized frequency of 0.5 and selecting a phase constraint of kπ reduces the complexity of the orthogonal projection operator Pv₁v₂^⊥.

Turning now to the implementation details of the orthogonal projection operator Pv₁v₂^⊥, the equalizer's frequency response is $C\left(e^{\mathrm{j2\pi }f}\right)=\sum _k{C}_k{e}^{-j\mathrm{k2\pi j}\mathrm{fT}}$where C_k are the coefficients of the equalizer's impulse response. At the preamble frequency (1/4T), the equalizer's frequency response is $C\left(e^{j\frac{\pi }{2}}\right)=\sum _k{C}_k{e}^{-jk\frac{\pi }{2}}$where the sampling period has been normalized to T=1. In matrix form, the equalizer's frequency response at the preamble frequency is, $C\left(e^{j\frac{\pi }{2}}\right)={\underset{\_}{C}}^T\left[\begin{array}{c}{\left(e^{j\frac{\pi }{2}}\right)}^0\\ {\left(e^{j\frac{\pi }{2}}\right)}^{-1}\\ ⋮\\ {\left(e^{j\frac{\pi }{2}}\right)}^{-\left(N-1\right)}\end{array}\right]={\underset{\_}{C}}^T\left[\begin{array}{c}{\left(j\right)}^0\\ {\left(j\right)}^{-1}\\ ⋮\\ {\left(j\right)}^{-\left(N-1\right)}\end{array}\right]={\underset{\_}{C}}^T\left[\begin{array}{c}1\\ -j\\ -1\\ j\\ ⋮\end{array}\right]$Those skilled in the art will recognize that shifting the time base will lead to four different, but functionally equivalent, frequency responses at the preamble frequency (i.e., [1,−j,−1,j, . . . ]C, [−j,−1,j,1, . . . ]C, [−1,j,1,−j, . . . ]C and [j,1,−j,−1, . . . ]C). Constraining the phase response of the equalizer to an integer multiple of π at the preamble frequency (1/4T) implies that the imaginary component of its frequency response is zero, ${\underset{\_}{C}}^T\left[\begin{array}{c}0\\ -1\\ 0\\ 1\\ ⋮\end{array}\right]={\underset{\_}{C}}^T·V_1=0$If the imaginary component of the frequency response is constrained to zero, as described above, then constraining the magnitude of the equalizer to g at the preamble frequency (1/4T) implies that the real component of the frequency response equals g, ${\underset{\_}{C}}^T\left[\begin{array}{c}1\\ 0\\ -1\\ 0\\ ⋮\end{array}\right]={\underset{\_}{C}}^T·V_2=g$Therefore, the equalizer's coefficients C_k must be constrained to satisfy the following two conditions:C^T·V₁=0 and C^T·V₂=g. The above constraints are achieved by multiplying the computed gradient X_k·e_k by an orthogonal projection operation Pv₁v₂^⊥ as part of the modified LMS algorithm.

To understand the operation of the orthogonal projection operation, consider an equalizer that comprises only two coefficients: C₀ and C₁ as shown in FIG. 10. The phase constraint condition C^T·V₁=0 implies that the filter coefficient vector C^T must be orthogonal to V₁. When using an unmodified LMS algorithm to update the filter coefficients, the orthogonal constraint is not always satisfied as shown in FIG. 10. The present invention, however, constrains the filter coefficients to a subspace <C> which is orthogonal to V₁ by multiplying the gradient values X_k·e_k by a projection operation Pv₁^⊥, where the null space of the projection operation Pv₁^⊥ is orthogonal to <C>. The updated coefficients correspond to a point on the orthogonal subspace <C> closest to the coefficients derived from the unmodified LMS algorithm as shown in FIG. 10.

Similar to the phase constraint projection operation Pv₁^⊥, a second orthogonal projection operation Pv₂^⊥ constrains the filter coefficients such that the coefficient vector C^T satisfies the above gain constraint: C^T·V₂=g. The combined orthogonal projection operator Pv₁v₂^⊥ eliminates two degrees of freedom in an N-dimensional subspace where N is the number of filter coefficients (i.e., the orthogonal projection operator Pv₁v₂^⊥ has a rank of N−2).

An orthogonal projection operator for V₁ and V₂ can be computed according toPv_x^⊥=I−Pv_x=I−V_x(V_x^TV_x)⁻¹V_x^T where Pv₁v₂^⊥=Pv₁^⊥·Pv₂^⊥ since V₁ is orthogonal to V₂. The orthogonal projection operator Pv₁v₂^⊥ computed using the above equation for an equalizer comprising ten filter coefficients is a matrix ${\mathrm{Pv}}_1{v}_2^{\perp }=\begin{array}{cccccccccc}4& 0& 1& 0& -1& 0& 1& 0& -1& 0\\ 0& 4& 0& 1& 0& -1& 0& 1& 0& -1\\ 1& 0& 4& 0& 1& 0& -1& 0& 1& 0\\ 0& 1& 0& 4& 0& 1& 0& -1& 0& 1\\ -1& 0& 1& 0& 4& 0& 1& 0& -1& 0\\ 0& -1& 0& 1& 0& 4& 0& 1& 0& -1\\ 1& 0& -1& 0& 1& 0& 4& 0& 1& 0\\ 0& 1& 0& -1& 0& 1& 0& 4& 0& 1\\ -1& 0& 1& 0& -1& 0& 1& 0& 4& 0\\ 0& -1& 0& 1& 0& -1& 0& 1& 0& 4\end{array}$

The above matrix Pv₁v₂^⊥ is an orthogonal projection matrix scaled by 5 (multiplied by 5) so that it contains integer valued elements which simplifies multiplying by X_k·e_k in the above LMS update equation. The scaling factor is taken into account in the selection of the gain value μ. Constraining the gain to g and the phase to kπ at the normalized frequency of 0.5 simplifies implementing the orthogonal projection matrix Pv₁v₂^⊥: half of the elements are zero and the other half are either +1, −1, or +4. Thus, multiplying the projection matrix Pv₁v₂^⊥ by the gradient values X_k·e_k requires only shift registers and adders. In the preferred embodiment, the orthogonal projection operator Pv₁v₂^⊥ is actually decimated in order to reduce the cost and complexity of the implementation. For details concerning the decimated embodiment of the orthogonal projection operator Pv₁v₂^⊥, see the above referenced co-pending U.S. patent application entitled, “GAIN AND PHASE CONSTRAINED ADAPTIVE EQUALIZING FILTER IN A SAMPLED AMPLITUDE READ CHANNEL FOR MAGNETIC RECORDING.”

The objects of the invention have been fully realized through the embodiments disclosed herein. Those skilled in the art will appreciate that the various aspects of the invention can be achieved through different embodiments without departing from the essential function. For example, the aspects of the present invention could be applied to attenuate non-linear distortions in the read signal other than those caused by the non-linear response of an MR read head. The particular embodiments disclosed are illustrative and not meant to limit the scope of the invention as appropriately construed from the following claims.

Claims

1. A sampled amplitude read channel for reading data recorded on a magnetic disk storage medium by detecting an estimated binary sequence from a sequence of discrete-time sample values generated by sampling an analog read signal emanating from a magneto-resistive (MR) read head positioned over the magnetic disk storage medium, the sampled amplitude read channel comprising: (a) an adaptive non-linear correction circuit for correcting a non-linearity in the analog read signal; (b) an LMS adaptation circuit, responsive to the discrete-time sample values, for adjusting a parameter of the adaptive non-linear correction circuit; and (c) a sequence detector for detecting an estimated binary sequence from the sequence of discrete-time sample values.

2. The sampled amplitude read channel as recited in claim 1, wherein the adaptive non-linear correction circuit comprises an n-th order component of a discrete-time Volterra filter where n is greater than one.

3. The sampled amplitude read channel as recited in claim 2, wherein: (a) the component of the Volterra filter is second order (n=2): yk=xkTH2xk

4. The sampled amplitude read channel as recited in claim 1, wherein the adaptive non-linear correction circuit is an analog correction circuit.

5. The sampled amplitude read channel as recited in claim 4, wherein the LMS adaptation circuit comprises: (a) a squarer for squaring one of the discrete-time sample values to generated a squared sample value; (b) a sample value estimator, responsive to the discrete-time sample values, for generating estimated sample values corresponding to the partial response; (c) an error generator for generating an error value between the discrete-time sample values and the estimated sample values; and (d) a multiplier for multiplying the squared sample value by the error value to generate an update value for use in adjusting the parameter of the adaptive non-linear correction circuit.

6. The sampled amplitude read channel as recited in claim 5, wherein the update value is a LMS update value for a coefficient in a second order component of a Volterra filter.

7. The sampled amplitude read channel as recited in claim 5, wherein the adaptive non-linear correction circuit comprises: (a) an analog squaring circuit for squaring the analog read signal to generate a squared analog signal; and (b) a multiplier for multiplying the squared analog signal by a coefficient, wherein the coefficient is adjusted in response to the update value.

8. The sampled amplitude read channel as recited in claim 7, wherein the adaptive non-linear correction circuit further comprises: (a) an adder for adding the analog read signal to the squared analog read signal to generate a corrected analog read signal; (b) a comparator for comparing the analog read signal to a threshold; and (c) a multiplexer for selecting between the analog read signal and the corrected analog read signal in response to the comparator.

9. The sampled amplitude read channel as recited in claim 5, further comprising: (a) a linear discrete-time equalizer for equalizing the discrete-time sample values according to a desired partial response; and (b) a constraint circuit for constraining a phase response of the linear discrete-time equalizer at a predetermined frequency.

10. The sampled amplitude read channel as recited in claim 9, wherein the constraint circuit implements an orthogonal projection operation.

11. The sampled amplitude read channel as recited in claim 9, wherein the predetermined frequency is a frequency near the maximum amplitude of the partial response spectrum.

12. The sampled amplitude read channel as recited in claim 9, wherein the predetermined frequency is substantially 1/4T where T is a baud rate period of the data recorded on the magnetic disk storage medium.

13. A sampled amplitude read channel for reading data recorded on a magnetic disk storage medium by detecting an estimated binary sequence from a sequence of discrete-time sample values generated by sampling an analog read signal emanating from a read head positioned over the magnetic disk storage medium, the sampled amplitude read channel comprising: (a) a sampling device for sampling the analog read signal to generate the sequence of discrete-time sample values; (b) a discrete-time Volterra equalizer comprising: (i) a discrete-time linear filter for equalizing the discrete-time sample values according to a desired partial response; and (ii) a discrete-time non-linear filter for attenuating a non-linearity in the discrete-time sample values; (c) an LMS adaptation circuit, responsive to the discrete-time sample values, for adjusting coefficients of the linear and non-linear filters; and (d) a sequence detector for detecting an estimated binary sequence from the sequence of discrete-time sample values output by the discrete-time Volterra equalizer.

14. The sampled amplitude read channel as recited in claim 13, wherein: (a) the non-linear filter outputs sample values yk according to: yk=xkTH2xk

15. A sampled amplitude read channel for reading data recorded on a magnetic disk storage medium by detecting an estimated binary sequence from a sequence of discrete-time sample values generated by sampling an analog read signal emanating from a magneto-resistive (MR) read head positioned over the magnetic disk storage medium, the sampled amplitude read channel comprising: (a) an adaptive non-linear analog correction circuit for correcting a non-linearity in the analog read signal; (b) a discrete-time linear equalizer for equalizing the discrete-time sample values according to a desired partial response; (c) an LMS adaptation circuit, responsive to the discrete-time sample values input into the discrete-time linear equalizer, for adjusting a parameter of the adaptive non-linear correction circuit; and (d) a sequence detector for detecting an estimated binary sequence from the sequence of discrete-time sample values.

16. The sampled amplitude read channel as recited in claim 15, wherein the LMS adaptation circuit comprises: (a) a squarer for squaring one of the discrete-time sample values to generated a squared sample value; (b) a sample value estimator, responsive to the discrete-time sample values output by the linear equalizer, for generating estimated sample values corresponding to the partial response; (c) an error generator for generating an error value between the discrete-time sample values and the estimated sample values; and (d) a multiplier for multiplying the squared sample value by the error value to generate an update value for use in adjusting the parameter of the adaptive non-linear analog correction circuit.

17. The sampled amplitude read channel as recited in claim 16, wherein the update value is a LMS update value for a coefficient in a second order component of a Volterra filter.

18. The sampled amplitude read channel as recited in claim 16, wherein the adaptive non-linear analog correction circuit comprises: (a) an analog squaring circuit for squaring the analog read signal to generate a squared analog signal; and (b) a multiplier for multiplying the squared analog signal by a coefficient, wherein the coefficient is adjusted in response to the update value.

19. The sampled amplitude read channel as recited in claim 18, wherein the adaptive non-linear analog correction circuit further comprises: (a) an adder for adding the analog read signal to the squared analog read signal to generate a corrected analog read signal; (b) a comparator for comparing the analog read signal to a threshold; and (c) a multiplexer for selecting between the analog read signal and the corrected analog read signal in response to the comparator.

20. The sampled amplitude read channel as recited in claim 16, further comprising a constraint circuit for constraining a phase response of the linear discrete-time equalizer at a predetermined frequency.

21. The sampled amplitude read channel as recited in claim 20, wherein the constraint circuit implements an orthogonal projection operation.

22. The sampled amplitude read channel as recited in claim 20, wherein the predetermined frequency is a frequency near the maximum amplitude of the partial response spectrum.

23. The sampled amplitude read channel as recited in claim 20, wherein the predetermined frequency is substantially 1/4T where T is a baud rate period of the data recorded on the magnetic disk storage medium.

24. A sampled amplitude read channel for reading data recorded on a magnetic disk storage medium by detecting an estimated binary sequence from a sequence of discrete-time sample values generated by sampling an analog read signal emanating from a read head positioned over the magnetic disk storage medium, the sampled amplitude read channel comprising: (a) an adaptive non-linear correction circuit for correcting a non-linearity in the analog read signal, comprising: (i) a n-th order polynomial circuit, responsive to the analog read signal, for generating a corrected analog read signal; (ii) a comparator for comparing the analog read signal to a threshold; and (iii)a multiplexer for selecting between a first analog signal and the corrected analog read signal in response to the comparator; (b) an adaptation circuit, responsive to the discrete-time sample values, for adjusting a coefficient of the n-th order polynomial circuit; and (c) a sequence detector for detecting the estimated binary sequence from the sequence of discrete-time sample values.

25. The sampled amplitude read channel as recited in claim 24, wherein n=2 in the n-th order polynomial circuit.

26. The sampled amplitude read channel as recited in claim 24, wherein the threshold is zero.