Gated noise-predictive filter calibration
An apparatus for calibrating a noise predictive filter includes a noise-predictive filter operable to filter digital data samples to yield filtered data samples, a calibration circuit operable to calculate tap coefficients for the noise-predictive filter based at least in part on the digital data samples, and a gating circuit operable to select a portion of the digital data samples for use by the calibration circuit in calculating the tap coefficients.
1. An apparatus for calibrating a noise predictive filter, comprising:
a noise-predictive filter operable to filter digital data samples to yield filtered data samples;
a calibration circuit operable to calculate tap coefficients for the noise-predictive filter based at least in part on the digital data samples; and
a gating circuit operable to select a portion of the digital data samples for use by the calibration circuit in calculating the tap coefficients.
2. The apparatus of claim 1 , further comprising an equalizer circuit connected to an input of the noise-predictive filter.
3. The apparatus of claim 1 , further comprising a data decoder operable to decode the filtered data samples, wherein the calibration circuit is operable to calculate the tap coefficients for the noise-predictive filter when decoding fails for the filtered data samples in the data decoder.
4. The apparatus of claim 1 , further comprising an averaging circuit operable to average multiple reads of the digital data samples prior to their use by the calibration circuit in calculating the tap coefficients.
5. The apparatus of claim 1 , further comprising a data detector circuit operable to detect data values in the filtered data samples.
6. The apparatus of claim 5 , further comprising an ideal value calculator connected to an output of the data detector circuit and operable to calculate ideal values for the digital data samples, wherein the gating circuit is operable to select the portion of the digital data samples at least in part based on the ideal values.
7. The apparatus of claim 1 , wherein the gating circuit comprises a convolution circuit operable to convolve detected values of the digital data samples with a partial response target to yield ideal values for the digital data samples.
8. The apparatus of claim 1 , wherein the gating circuit comprises a subtraction circuit operable to subtract ideal values for the digital data samples from the digital data samples to yield an error signal, an absolute value circuit operable to calculate an absolute value of the error signal, and a comparator operable to compare the absolute value of the error signal with a threshold.
9. The apparatus of claim 8 , further comprising a memory operable to store a plurality of values from the output of the comparator, wherein the gating circuit is operable to perform the selection based at least in part on an output of the memory.
10. The apparatus of claim 1 , wherein the gating circuit is operable to exclude digital data samples captured at a beginning portion of a data sector from use by the calibration circuit.
11. The apparatus of claim 1 , wherein the gating circuit is operable to select parity bits and to exclude ones of the digital data samples corresponding to encoded user data bits for use by the calibration circuit in calculating the tap coefficients.
12. The apparatus of claim 1 , wherein the noise-predictive filter,
the calibration circuit and the gating circuit are implemented as part of a data storage device.
13. A method for calibrating a noise-predictive filter, comprising:
equalizing digital data samples for a data sector to yield equalized data samples;
selecting a portion of the equalized data samples as calibration samples;
calculating tap coefficients based on the calibration samples; and
filtering the equalized data samples in the noise-predictive filter using the tap coefficients, wherein the tap coefficients are calculated after decoding of the data sector fails, and wherein the equalized data samples are re-filtered in the noise-predictive filter using the tap coefficients prior to re-decoding of the data sector.
14. The method of claim 13 , wherein the tap coefficients are calculated using a covariance matrix.
15. The method of claim 13 , wherein selecting the portion of the equalized data samples as calibration samples comprises excluding data samples obtained from a beginning portion of the data sector.
16. The method of claim 13 , wherein selecting the portion of the equalized data samples as calibration samples comprises selecting parity bits from the equalized data samples as calibration samples.
17. The method of claim 13 , further comprising calculating a quality metric for each of the equalized data samples, wherein selecting the portion of the equalized data samples as calibration samples comprises selecting the equalized data samples for which the quality metric is better than a threshold quality.
18. The method of claim 17 , wherein selecting the portion of the equalized data samples as calibration samples comprises selecting the equalized data samples for which the quality metric is better than the threshold quality for a successive group of the equalized data samples.
19. A storage device, comprising:
a magnetic storage medium operable to store data in data sectors;
a head assembly disposed in relation to the storage medium to read and write the data on the storage medium;
an analog to digital converter operable to convert an analog signal from the head assembly into a stream of digital samples;
an equalizer operable to equalize the stream of digital samples to yield equalized data samples;
a noise-predictive filter operable to filter the equalized digital samples to yield filtered data samples;
a data decoder operable to decode the filtered digital samples;
a calibration circuit operable to calculate tap coefficients for the noise-predictive filter based at least in part on the digital data samples; and
a gating circuit operable to select a portion of the digital samples for a data sector for use by the calibration circuit in calculating the tap coefficients after decoding of the digital samples for the data sector fails.
20. The storage device of claim 19 , wherein the calibration circuit is operable to calculate the tap coefficients using a covariance matrix.