JP2000200462A - Signal processing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent performing needless training after convergence upto an optimum state by confirming that an adaptive type transversal filter is converged to an optimum state. SOLUTION: This device is provided with an adaptive type transversal filter 5 performing adaptive operation for compensating a tap coefficient so that a calculated error between a target equalization level and an output is made smaller, a quality evaluating means 10a performing evaluation of a quality value of the output by comparing an output of the adaptive type transversal filter 5 and an ideal sample value, and a control means 9a having a function setting a tap coefficient of the adaptive type transversal filter 5 and selecting execution/non-execution of adaptive operation. And the control means 9a sets a tap coefficient to the adaptive type transversal filter 5 to a quality value given a good evaluation and disables adaptive operation of the adaptive type transversal filter 5.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a signal processing apparatus for optimally setting an FIR (transversal filter) and a method of adjusting the signal processing apparatus.

[0002]

2. Description of the Related Art FIG. 6 is an explanatory diagram of a conventional example. In FIG. 6, a conventional signal processing device includes a head 1, a variable gain amplifier (VGA) 2, a low-pass filter (LPF) 3,
A / D converter 4, an adaptive FIR 5, a PLL (Phase Lock Loop) 6, an automatic gain controller (AGC) 7, and a detector 8 are provided.

The head 1 amplifies a signal read from the magnetic disk 1a by a head IC or the like and outputs the amplified signal. The variable gain amplifier 2 amplifies the signal from the head 1 under the control of the automatic gain control unit 7. The low-pass filter 3 is for cutting high-frequency noise of the signal from the variable gain amplifier 2. The A / D converter 4 converts an analog signal from the low-pass filter 3 into a digital signal. The adaptive FIR 5 performs an adaptive operation so that the set value of the FIR becomes an optimum value. P
LL6 includes a phase comparator, a low-pass filter, an error amplifier, a voltage controlled oscillator, and the like, and generates an arbitrary frequency according to an input signal. The automatic gain control unit 7 controls the variable gain amplifier 2 according to an input signal. The detector 8 detects a signal from the adaptive FIR 5.

In the prior art shown in FIG. 6, an adaptive FIR
The number of taps was reduced so that 5 could be easily adapted. In the case of this small number of taps, the adaptive FIR5 easily converged to the optimum state with a small amount of training, but the equalization error was large due to the small number of taps. However, when the number of taps was increased in order to reduce the equalization error, the convergence of the adaptive FIR5 became more and more slow.

As a conventional example, Japanese Patent Application Laid-Open No. 7-1410
As disclosed in Japanese Patent Application Laid-Open No. 3 (1999) -1995, there is a method of using FIR to set a tap coefficient of FIR. However, this is a fixed and non-adaptive tap coefficient, and does not evaluate a quality value.

[0006]

SUMMARY OF THE INVENTION
There were the following issues. : FIR with a large number of taps slows down convergence, so if the training time is short, it may not converge to an optimal state.

[0007] Training must be performed for a sufficiently long time to ensure convergence to an optimal state. : The convergence time of the FIR also changed depending on the SN ratio (ratio of the signal level to the noise level) of the signal.

The present invention solves such a conventional problem, confirms that the FIR has converged to an optimum state, and prevents unnecessary training after converging to the optimum state. The purpose is to:

[0009]

FIG. 1 is a diagram illustrating the principle of the present invention. In FIG. 1, 3 is a low-pass filter (LPF),
5 is an adaptive transversal filter (FIR), 8a
Is a data detection means, 9a is a control means, 10a is a quality evaluation means, 11a is a setting storage means, and 12a is a quality storage means.

The present invention is configured as follows to solve the above-mentioned conventional problems. (1): An adaptive transversal filter 5 that calculates an error between a target equalization level and an output and corrects a tap coefficient so as to reduce the error, and an output of the adaptive transversal filter 5 Quality evaluation means 10a for comparing the output with an ideal sample value to evaluate the quality value of the output, and selecting the setting of the tap coefficient of the adaptive transversal filter 5 and the operation / non-operation of the adaptive operation. Control means 9a having a function of performing the adaptive transversal filter 5 by setting tap coefficients of the adaptive transversal filter 5 to those having a good quality value evaluation. Operation is deactivated.

(2) In the signal processing device of the above (1), a low-pass filter 3 is provided in a stage preceding the adaptive transversal filter 5, and the control means 9a can set filter characteristics of the low-pass filter 3. Has functions.

(3) In the signal processing device of (1) or (2), the quality evaluation means 10a includes data detection means 8a for detecting data from the output of the adaptive transversal filter 5, Data detection means 8a
Is compared with the ideal pattern, and the number of times the detection result of the data detection means 8a is incorrect is output as a quality value.

(4) In the signal processing apparatus of (1) to (3), if the tap coefficients of the adaptive transversal filter 5 and the variable gain amplifier and the low-pass filter are included, the respective set values are stored. A setting storage unit 11a is provided.

(5) In the signal processing device of (1) to (4), a quality storage means 12a for storing the quality value of the quality evaluation means 10a is provided.

(6): An adaptive transversal filter 5 that calculates an error between a target equalization level and an output and performs an adaptive operation of correcting a tap coefficient so as to reduce the error.
And a quality evaluation for comparing the output of the adaptive transversal filter 5 with an ideal sample value by comparing the output of the low-pass filter 3 provided before the adaptive transversal filter 5 with an ideal sample value. Means 10a; and control means 9a for controlling the low-pass filter 3 and the adaptive transversal filter 5. The control means 9a initializes the low-pass filter 3 and the adaptive transversal filter 5, Next, the filter characteristics of the low-pass filter 3 are changed, and the filter characteristics are set based on the output of the quality evaluation unit 10a at the time when the evaluation is good. Activating the adaptive operation, the tap coefficient is evaluated from the output of the quality evaluation means 10a at that time to evaluate the quality value. Set to those of was good, then,
A method for adjusting a signal processing device, wherein the adaptive operation of the adaptive transversal filter is deactivated.

(Operation) The operation based on the above configuration will be described. The adaptive transversal filter 5 calculates an error between the target equalization level and the output and performs an adaptive operation to correct the tap coefficient so as to reduce the error.
0a, the output of the adaptive transversal filter 5 is compared with an ideal sample value to evaluate the quality value of the output, and the control means 9a sets the tap coefficients of the adaptive transversal filter 5 and The adaptive transversal filter 5 has a function of selecting operation / non-operation of the adaptive operation. The control means 9a sets the tap coefficient of the adaptive transversal filter 5 to a value having a good evaluation of the quality value, and Is deactivated. For this reason, it is possible to confirm that the adaptive transversal filter 5 has converged to the optimum state, and to prevent unnecessary training after converging to the optimum state.

The control means 9a has a function of setting the filter characteristics of the low-pass filter 3 provided before the adaptive transversal filter 5. Therefore, the filter characteristics such as the cutoff frequency and the boost amount can be adjusted in accordance with the range in which the adaptive transversal filter 5 can perform the adaptive operation.

Further, the quality evaluation means 10a compares the output of the data detection means 8a with an ideal pattern, and outputs the number of times the detection result of the data detection means 8a is incorrect as a quality value. For this reason, the wrong number can be easily calculated by a counter or the like, and the quality value can be easily compared by dividing the wrong number by the total number of data of the ideal pattern in which the comparison is made.

Further, a setting storage means 11a for storing the set values of the tap coefficients of the adaptive transversal filter 5, the variable gain amplifier, and the low-pass filter if they are included. Therefore, the setting storage unit 11a
Tap coefficient and variable gain amplifier converged to the optimal state of
If a low-pass filter or the like is included, each set value can be set, and unnecessary training can be prevented from being performed later.

Further, there is provided a quality storage means 12a for storing the quality value of the quality evaluation means 10a. For this reason, by storing the quality of the optimal setting in the quality storage means 12a, the quality can be periodically evaluated and readjustment can be performed if the quality is extremely poor.

The control means 9a initializes the low-pass filter 3 and the adaptive transversal filter 5,
Next, the filter characteristics of the low-pass filter 3 are changed, and the filter characteristics are set based on the output of the quality evaluation unit 10a at the time when the evaluation is good. A signal that activates the adaptive operation and sets the tap coefficient to a value having a good quality value evaluation from the output of the quality evaluation means 10a at that time, and then deactivates the adaptive operation of the adaptive transversal filter. It is a method of adjusting the processing apparatus. For this reason, it is necessary to confirm that the optimum setting of the filter characteristics of the low-pass filter 3 and the adaptive transversal filter 5 have converged to an optimum state, and to prevent unnecessary training after converging to the optimum state. it can.

[0022]

2 to 5 are views showing an embodiment of the present invention. Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(1): Description of Signal Processing Device FIG. 2 is an explanatory diagram of the signal processing device. The signal processing device of the present invention is a PR4M which is a signal detection method of a magnetic disk drive.
L (partial response + maximum likelihood detector).

In FIG. 2, the signal processing device includes a head 1, a variable gain amplifier (VGA) 2, a low-pass filter (LPF) 3, an A / D converter 4, an adaptive FIR 5,
PLL (phase lock loop) 6, automatic gain control unit 7, PR4 detector 8, signal processing system control unit 9, signal processing system quality evaluation unit 10, signal processing system setting storage unit 11, signal processing system quality storage unit 12, A training pattern generator 13 is provided. Also, signal processing system quality evaluation means 1
For 0, a PR4 conversion 14, a subtracter 15, a squarer 16, and an adder 17 are provided.

The head 1 is for amplifying a signal read from the magnetic disk 1a by a head IC or the like and outputting the amplified signal. The variable gain amplifier 2 amplifies the signal from the head 1 under the control of the automatic gain control unit 7. The low-pass filter 3 cuts the high-frequency noise of the signal from the variable gain amplifier 2, and the cutoff frequency and the boost amount are set by the signal processing system control unit 9.

The A / D converter 4 converts an analog signal from the low-pass filter 3 into a digital signal. The adaptive FIR 5 performs an adaptive operation so that the set value of the FIR becomes an optimum value (“1” in the case of PR4,
"0" and "-1"). PLL6
Is composed of a phase comparator, a low-pass filter, an error amplifier, a voltage controlled oscillator and the like, and generates an arbitrary frequency according to an input signal. The automatic gain control unit 7 controls the variable gain amplifier 2 according to an input signal. PR4
The detector 8 detects a data signal from the adaptive FIR 5.

The signal processing system control means 9 sets the low-pass filter 3, sets the tap coefficient of the adaptive FIR 5, selects the operation / non-operation of the adaptive operation of the adaptive FIR 5, and selects the adaptive FIR 5.
Acquisition of tap coefficients after adaptive operation of IR5, adaptive FIR
5 to check whether the FIR tap coefficients are not saturated during the adaptive operation, and the obtained low-pass filter 3 and adaptive F
An operation of storing the set value and the quality value of the IR5 in the storage unit is performed.

The signal processing system quality evaluation means 10 evaluates the quality of the signal processing system. The signal processing system setting storage unit 11 stores setting values of the low-pass filter 3 and the adaptive FIR 5. Signal processing system quality storage means 1
Numeral 2 stores the quality value at the time of the optimal setting value of the adaptive FIR5. The training pattern generator 13
For example, a training pattern such as a random pattern is generated.

The PR4 converter 14 converts a signal from the training pattern generator 13. Subtractor 1
Reference numeral 5 denotes a difference between the output from the PR4 converter 14 and the output from the adaptive FIR 5. The squarer 16 is the subtractor 1
5 is squared. The adder 17 calculates the sum of the output values of the squarer 16.

(Explanation of Operation of Signal Processing Apparatus) A head IC output signal read from the magnetic disk 1a by the head and amplified by the head IC is applied to the adaptive FIR by the variable gain amplifier 2.
5 is amplified until it has the desired magnitude. The gain of the variable gain amplifier 2 is automatically controlled by the automatic gain control unit 7. Thereafter, the high-pass noise component is cut by the low-pass filter 3, and the sampled value of the PR4 equalized signal is output from the PR4 detector 8 through the A / D converter 4 and the adaptive FIR 5. At this time, the low-pass filter 3
And tap coefficient setting of adaptive FIR5, adaptive FIR5
Of the tap coefficient after the adaptive operation of the adaptive FIR 5, check whether the FIR tap coefficient is saturated during the adaptive operation of the adaptive FIR 5, the obtained low-pass filter 3 and the adaptive FIR 5
Is stored in the storage means.

(2) Description of Method for Optimizing LPF and FIR First, data generated by the training pattern generator 13 is written to a medium (magnetic disk 1a). Thereafter, adjustment is performed as follows.

1. The signal processing system control means 9 initializes the low-pass filter 3 and the adaptive FIR 5. As the initial setting value, for example, a value obtained by a theoretical calculation or an average value of some units is set.

2. If the automatic gain control unit 7 for automatically controlling the gain of the variable gain amplifier 2 is not provided, the gain of the variable gain amplifier 2 is adjusted. This adjustment is performed while changing the setting of the gain of the variable gain amplifier 2.
The output of the signal processing system quality evaluation means 10 is monitored, and the variable gain amplifier 2 is set with the setting with the best evaluation.

3. The low-pass filter 3 is adjusted. This adjustment is performed by the signal processing system control means 9 by the low-pass filter 3.
The output of the signal processing system quality evaluation means 10 is monitored while gradually changing the cutoff frequency or the boost amount (or both the cutoff frequency and the boost amount) which is the filter characteristic of The filter 3 is set. Note that this low-pass filter 3
If the adaptive operation of the adaptive FIR 5 is possible for a wide range of settings, this adjustment can be omitted.

4. Adjust the adaptive FIR5. This adjustment procedure is as follows. a) The signal processing system control means 9 activates the adaptive means of the adaptive FIR 5 (the adaptive operation is not performed in the above-mentioned 1 to 3).

B) Read a certain number of data (for example, the number of data in which the variation of the quality value is within the regulation). Thereafter, the signal processing system control means 9 reads the set value and the quality value of the adaptive FIR 5 and stores them in the signal processing system setting storage means 11 and the signal processing system quality storage means 12. Hereinafter, this is referred to as an adaptive operation.

C) When the adaptive FIR 5 detects that the FI
If the function to output a saturation warning signal when the R tap coefficient reaches the upper limit or lower limit of the settable range,
When a saturation warning signal is output during the adaptive operation, the signal processing system control means 9 determines that the FIR tap coefficients have been saturated.

When the adaptive FIR 5 does not have a function of outputting a saturation warning signal, the signal processing system control means 9 sets the FIR set value read after the adaptive operation very close to the upper or lower limit of the settable range. In this case, it is determined that the FIR tap coefficient is saturated.

When it is determined that the FIR tap coefficient is saturated, the setting of the adaptive FIR5 is set such that the setting value read after the immediately preceding adaptive operation is reduced at a fixed rate as a whole (for example, , All tap coefficients multiplied by 0.9).

D) If the quality value read after the adaptation operation is significantly different from the best value among the quality values of all the adaptation operations up to that point, the signal processing system control means 9 causes the adaptation means to be affected by noise or the like. Judge that it did not work well,
The FIR is set again with the setting value of the one having the best quality value among all the adaptive operations up to that time.

E) The above steps b) to d) are repeated. However,
The repetition ends when the following conditions are satisfied.

When the quality does not substantially change near the best value in several adaptive operations. In this case, the FI
It is determined that R has been adapted to an optimum state and is stable, and any one of the FIR set values considered to be stable or an average value is set as the optimum setting.

When the number of repetitions of b) to d) reaches a predetermined maximum number of repetitions. In this case, the optimum value is set as the optimum value of the quality value of all the adaptive operations.

As described above, after one of the above-mentioned steps 1 to 4 has been completed, the above-mentioned steps 2 to 4 may be repeated several times with the adjusted low-pass filter 3 and adaptive FIR 5 set as initial settings.

(Explanation of Operation of Signal Processing System Quality Evaluation Means) As a signal processing system quality value, a training pattern
The difference between the output of the 4-conversion 14 and the output of the adaptive FIR 5 is squared by the squarer 16, and the sum thereof is calculated by the adder 17. Note that means for taking an absolute value may be provided instead of the squarer 16.

The obtained low-pass filter 3, adaptive FI
The optimum value of R5 and the quality value at that time are stored in the signal processing system setting storage unit 11 and the signal processing system quality storage unit 12, respectively. Once adjusted, the settings of the low-pass filter 3 and the adaptive FIR 5 are achieved simply by setting the values stored in the storage means.

The signal processing system control means 9 periodically evaluates the quality of the signal processing system, and if the quality is significantly inferior to the value stored in the signal processing system quality storage means 12, the low-pass filter 3 , Adaptive FIR5 is again adjusted optimally.

Although a training pattern obtained by PR4 conversion 14 is used as a means for calculating an ideal equalized waveform in the above description, the target equalization level calculated by the adaptive means of the adaptive FIR 5 (the determination circuit 46). Output) can also be used.

(3) Description of the case where a counter is provided in the signal processing system quality evaluation means FIG. 3 is an explanatory diagram in the case where a counter is provided in the signal processing system quality evaluation means. This is applied to PR4ML (partial response + maximum likelihood detector), which is a signal detection method of the above.

In FIG. 3, the signal processing device includes a head 1, a variable gain amplifier (VGA) 2, a low-pass filter (LPF) 3, an A / D converter 4, an adaptive FIR 5,
PLL (phase lock loop) 6, automatic gain controller 7, PR4 detector 8, signal processing system control means 9, signal processing system quality evaluation means 10, training pattern generator 13
Is provided. The signal processing system quality evaluation means 10 includes a comparator 21, an error counter 22, an error rate calculator 23, and a read bit number counter 24.

The head 1 amplifies a signal read from the magnetic disk 1a by a head IC or the like and outputs the amplified signal. The variable gain amplifier 2 amplifies the signal from the head 1 under the control of the automatic gain control unit 7. The low-pass filter 3 cuts the high-frequency noise of the signal from the variable gain amplifier 2, and the cutoff frequency and the boost amount are set by the signal processing system control unit 9.

The A / D converter 4 converts an analog signal from the low-pass filter 3 into a digital signal. The adaptive FIR 5 performs an adaptive operation so that the set value of the FIR becomes an optimum value (“1” in the case of PR4,
"0" and "-1"). PLL6
Is composed of a phase comparator, a low-pass filter, an error amplifier, a voltage controlled oscillator and the like, and generates an arbitrary frequency according to an input signal. The automatic gain control unit 7 controls the variable gain amplifier 2 according to an input signal. PR4
The detector 8 detects a data signal from the adaptive FIR 5.

The signal processing system control means 9 sets the low-pass filter 3, sets the tap coefficient of the adaptive FIR 5, selects the operation / non-operation of the adaptive operation of the adaptive FIR 5, and sets the adaptive FIR 5.
Acquisition of tap coefficients after adaptive operation of IR5, adaptive FIR
During the adaptive operation of No. 5, an operation such as checking whether or not the FIR tap coefficient is saturated is performed.

The signal processing system quality evaluation means 10 evaluates the quality of the signal processing system. The training pattern generator 13 is for generating a training pattern such as a random pattern.

The comparator 21 compares the output of the training pattern generator 13 with the output of the PR4 detector 8.
The error counter 22 counts the number of errors from the comparator 21. The error rate calculator 23 calculates
The error rate is obtained by dividing the counted number of errors by the counted number of read bits. The read bit number counter 24 counts the number of read bits based on the clock of the PLL 6.

(Explanation of the Operation of the Signal Processing Apparatus) The head IC output signal obtained by amplifying the signal from the magnetic disk 1a by the head IC is amplified by the variable gain amplifier 2 until the output signal of the adaptive FIR 5 becomes a desired magnitude. You. The gain of the variable gain amplifier 2 is automatically controlled by the automatic gain control unit 7. Then, low-pass filter 3
, The high-frequency noise component is cut off, and the signal passes through the A / D converter 4 and the adaptive FIR 5, and is output from the PR4 detector 8 to the PR4.
A sample value of the equalized signal is output. At this time, the signal processing system control means 9 sets the low-pass filter 3,
It performs operations such as setting tap coefficients of the adaptive FIR5, acquiring tap coefficients after the adaptive FIR5 adaptive operation, and checking whether the FIR tap coefficients are not saturated during the adaptive FIR5 adaptive operation. .

(Explanation of Operation of Signal Processing System Quality Evaluating Means 10) As a signal processing system quality value, a comparator 21 compares a training pattern from training pattern generator 13 with an output signal from PR4 detector 8 as a data detector. And the error counter 22 counts the number of errors. On the other hand, the read bit number counter 24 counts the number of read bits based on the clock of the PLL 6. The error rate calculation unit 23 obtains an error rate by dividing the number of errors counted by the error counter 22 by the number of read bits counted by the read bit number counter 24, and outputs this to the signal processing system control unit 9 as a quality value.

The read bit number counter 24 may be provided separately, and the signal processing system control means 9 may read the count value of the read bit number and determine the error rate (quality value).

As described above, the advantage of providing the counter is that it is not necessary to keep the number of read bits from the magnetic disk constant during quality evaluation. If there is no counter, it is necessary to always keep the number of read bits read at the time of signal evaluation constant. Otherwise, even at the same error rate, the error count value increases as the number of read bits increases, resulting in an inappropriate value as the quality value.

(4) Description of Adaptive FIR FIG. 4 is an explanatory diagram of the adaptive FIR. In FIG. 4, the adaptive FIR, which is an adaptive equalizer, includes delay circuits 31 and 32, multipliers 33 and 34, and a coefficient correction circuit (1,
2) 35, 36 and an adder 37 are provided.

The delay circuits 31 and 32 delay the input signal (one sample time delay D) and output it.
The multipliers 33 and 34 are for multiplying two input signals. The coefficient correction circuits 35 and 36 correct tap coefficients. Note that these coefficient correction circuits 35, 3
6 is connected to the signal processing system control means 9. Adder 3
Numeral 7 is for adding the input signals.

The input signal x _{i + 1} is delayed by one sample time in the delay circuit 31, and the signal x _i is output. This signal x
_i is input to the adder 37, and further delayed by one sample time in the delay circuit 32, and the delayed signal x
_i−1 is input to the multiplier 33 and the coefficient correction circuit 35. The coefficient correction circuit 35, an adder by 37 which is the output equalizer output y _i and the signal x _i of outputting the correction coefficient c _1, the multiplier 33 outputs c ₁ of the coefficient correction circuit 35 and the signal x _i-1 And input to the adder 37. On the other hand, the coefficient correction circuit 36
Is a correction coefficient c based on the equalizer output y _i and the input signal x _{i + 1.
-1} is output, and the output c _-1 of the coefficient correction circuit 36 is output from the multiplier 34.
And the input signal x _{i + 1} are input to the adder 37.

(5) Description of the coefficient correction circuit FIG. 5 is an explanatory diagram of the coefficient correction circuit. 5, the coefficient correction circuit includes multipliers 41 and 42, subtractors 43 and 4,
4, a delay circuit 45, a determination unit 46, and a setting selection unit 47 are provided.

The multipliers 41 and 42 calculate a multiplied value. The subtracters 43 and 44 calculate a subtraction value. The delay circuit 45 delays an input signal and outputs the delayed signal. The determination unit 46 outputs the calculated target equalization level. When the input signal y _i is greater than 0.5 (y _i > 0.5), the determination unit 46 outputs “1”, and outputs the input signal y _i Is larger than −0.5 and smaller than 0.5 (−0.
It outputs "0" when 5 <y _i <0.5) and outputs "-1" when the input signal y _i is smaller than -0.5 (y _i <-0.5).

The operation of the coefficient correction circuit is as follows.
Then, the output signal y _i of the equalizer is determined, and the determined signal is input to the subtractor 43. The subtractor 43 subtracts the signal determined by the determination unit 46 from the signal y _i . On the other hand, multiplies the input signal x _i and the output signal of the subtracter 43 by the multiplier 41, and inputs a signal multiplication to the subtracter 44. Subtractor 4
4, a delay circuit 4 for delaying the output signal of the subtractor 44
5 is subtracted from the output signal of the multiplier 41, and the subtracted output is input to the multiplier 42. The multiplier 42 multiplies the output signal of the subtractor 44 by a constant α to obtain a correction coefficient c.
And the correction coefficient c is output from the setting selector 47.

It should be noted that, for a certain constant α, the larger this value is, the better the followability of the adaptive FIR is, but the greater the erroneous adaptation due to noise or the like is. Conversely, if the value of this constant α is small,
The adaptability of the adaptive FIR becomes poor, and erroneous adaptation due to noise or the like is reduced.

The setting selecting section 47 is connected to the signal processing system control means 9 to set tap coefficients, select the operation / non-operation of the adaptive operation, and obtain the tap coefficient (correction coefficient c) after the adaptive operation. During the adaptive operation, whether the FIR tap coefficients are saturated or not is checked.

As described above, it is possible to make adjustments while checking the quality of the signal processing system. As a result, it is possible to confirm that the adaptive FIR has converged to the optimal state, and to prevent unnecessary training after the adaptive FIR has converged to the optimal state. Further, when the setting of the signal processing system deviates from the optimum value due to the aging of the signal, this can be confirmed and a new optimum adjustment can be performed.

[0069]

As described above, the present invention has the following effects. (1): An adaptive transversal filter calculates an error between a target equalization level and an output, performs an adaptive operation of correcting a tap coefficient so as to reduce the error, and performs quality adaptation by the adaptive transversal filter. Is compared with the ideal sample value to evaluate the quality value of the output, and the control means selects the setting of the tap coefficient of the adaptive transversal filter and the activation / deactivation of the adaptive operation. The adaptive transversal filter has a function, and the control means sets the tap coefficient of the adaptive transversal filter to a value having a good evaluation of the quality value to deactivate the adaptive operation of the adaptive transversal filter. Check that the versal filter has converged to the optimal state and prevent unnecessary training after converging to the optimal state Door can be.

(2) Since the control means has a function of setting the filter characteristics of the low-pass filter provided in the preceding stage of the adaptive transversal filter, the control means is adjusted to a range where the adaptive operation of the adaptive transversal filter is possible. Filter characteristics such as cutoff frequency and boost amount can be adjusted.

(3) The quality evaluation means compares the output of the data detection means with the ideal pattern and outputs the number of times the detection result of the data detection means is incorrect as a quality value. It is easy to calculate the number of times, and the quality value can be easily compared by dividing the incorrect number by the total number of data of the ideal pattern.

(4): If a tap coefficient of the adaptive transversal filter, a variable gain amplifier and a low-pass filter are included, setting storage means for storing respective setting values is provided. If a coefficient, a variable gain amplifier, a low-pass filter, and the like are included, respective setting values can be set, and unnecessary training can be prevented later.

(5) Since the quality storage means for storing the quality value of the quality evaluation means is provided, the quality of the optimum setting is stored in the quality storage means, so that the quality evaluation is periodically performed, and the quality is evaluated. If it is significantly worse, readjustment can be performed.

(6): The control means initializes the low-pass filter and the adaptive transversal filter.
By changing the filter characteristic of the low-pass filter, setting the filter characteristic with a setting having good evaluation from the output of the quality evaluation means at that time, and then operating the adaptive operation of the adaptive transversal filter. The tap coefficient is set to a value having a good quality value evaluation from the output of the quality evaluation means at that time, and thereafter, the adaptive processing of the adaptive transversal filter is deactivated so as to be a method of adjusting the signal processing device. Therefore, confirm that the optimal setting of the filter characteristics of the low-pass filter and that the adaptive transversal filter has converged to the optimal state,
Unnecessary training after the convergence to the optimal state can be prevented.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is an explanatory diagram of a signal processing device according to an embodiment.

FIG. 3 is an explanatory diagram in a case where a counter is provided in a signal processing system quality evaluation unit in the embodiment.

FIG. 4 is an explanatory diagram of an adaptive FIR according to the embodiment.

FIG. 5 is an explanatory diagram of a coefficient correction circuit according to the embodiment.

FIG. 6 is an explanatory diagram of a conventional example.

[Explanation of symbols]

 Reference Signs List 3 Low-pass filter (LPF) 5 Adaptive transversal filter (FIR) 8a Data detection means 9a Control means 10a Quality evaluation means 11a Setting storage means 12a Quality storage means

Claims (6)

[Claims] 1. An adaptive transversal filter for calculating an error between a target equalization level and an output and performing an adaptive operation for correcting a tap coefficient so as to reduce the error, and an output of the adaptive transversal filter. Quality evaluation means for comparing the quality value of the output with an ideal sample value; and a function of setting the tap coefficient of the adaptive transversal filter and selecting an operation / non-operation of the adaptive operation. The adaptive transversal filter sets the tap coefficient of the adaptive transversal filter to a value having a good evaluation of the quality value to deactivate the adaptive operation of the adaptive transversal filter. A signal processing device characterized by the following. 2. A signal processing apparatus according to claim 1, wherein a low-pass filter is provided at a stage preceding said adaptive transversal filter, and said control means has a function of setting filter characteristics of said low-pass filter. . 3. The quality evaluation means includes data detection means for detecting data from an output of the adaptive transversal filter, and compares the output of the data detection means with an ideal pattern to detect the data. 3. The signal processing apparatus according to claim 1, wherein the number of times the detection result of the means is incorrect is output as a quality value. 4. The apparatus according to claim 1, further comprising setting storage means for storing tap values of said adaptive transversal filter, a variable gain amplifier, and setting values of said low pass filter if said low pass filter is included. A signal processing device according to any one of claims 1 to 3. 5. A signal processing apparatus according to claim 1, further comprising a quality storage means for storing a quality value of said quality evaluation means. 6. An adaptive transversal filter for calculating an error between a target equalization level and an output and performing an adaptive operation for correcting a tap coefficient so as to reduce the error, and a stage preceding the adaptive transversal filter. A low-pass filter provided; quality evaluation means for comparing an output of the adaptive transversal filter with an ideal sample value to evaluate a quality value of the output; and a low-pass filter and an adaptive transversal filter. Control means for performing control, wherein the control means initializes the low-pass filter and the adaptive transversal filter, and then changes the filter characteristics of the low-pass filter to thereby control the quality evaluation means. From the output, set the filter characteristics with a setting that has a good evaluation. Operating the adaptive operation of the adaptive filter to set the tap coefficients to those having a good evaluation of the quality value from the output of the quality evaluation means at that time, and then deactivate the adaptive operation of the adaptive transversal filter. A method for adjusting a signal processing device, comprising: