JPS6394412A - Reading signal reproducing circuit for magnetic recording - Google Patents

Abstract

PURPOSE:To decrease the probability that an error occurs in a reproducing signal by generating a condition signal to mask a superfluous zero cross pulse from the differential signal of a reading signal. CONSTITUTION:A differention circuit 10 analog-converts a digital signal recorded at recording media, obtains a reading signal AS and converts it to a differential signal DS. The output is sent to a positive negative discriminating circuit 20, the positive negative threshold and Vth are compared, a condition signal SS to show the positive negative condition is formed and sent to a selecting circuit 40. On the other hand, a zero pulse generating circuit 30 generates the zero cross pulse column of a differential signal and supplies a zero cross pulse ZP to the circuit 40. The circuit 40 selects the zero cross pulse applicable to the positive/negative condition shown by the signal SS out of the pulse ZP and generates a reproducing signal to show the changing-over position of a magnetic pattern from the position. Then, the changing-over point of the signal SS is selected longer than the condition shifting time to come off from the zero cross point of the signal DS and shorter than the 1/2 of a maximum period held by a reading signal.

Description

[Detailed description of the invention] [Technical field to which the invention pertains]

The present invention provides a reproducing circuit for obtaining a reproduction signal useful for reproducing digital information from an analog waveform read signal read out from digital information recorded in a magnetic pattern of a predetermined format on a magnetic storage medium such as a fixed disk device. Regarding.

[Prior art and its problems]

In the above-mentioned fixed disk device, the information magnetically written on the disk is a magnetic pattern in which the N pole and the S pole alternate sequentially, and the width of the N and 3 areas along the longitudinal direction of the trunk is in a predetermined format. 1. For example, it is written in a format defined by a system called MFM, RLL, GCR, etc. The width of this area is 0, which is the content of digital information.
, 1, and when reading the data, an electromotive force based on a change in magnetic flux from the N and 3 areas moving below the head is output as a read signal. This read signal has a peak at the boundary between the N and S page areas in the magnetic pattern, so if you accurately detect this peak position from the read signal, you can easily determine the width of the area and the written digital data from the peak interval. Can be reproduced. However, the read signal from the head has an analog waveform with many deformations and noises, and it is impossible to directly detect the peak interval from it. Therefore, a reproduction circuit, which is the object of the present invention, is provided for processing the read signal, and converts the read signal into a digital pulse or signal indicating the peak position therein. A conventional example will be described below with reference to the drawings. FIG. 3 is a block circuit diagram showing the configuration of a standard reproducing circuit. The read signal AS generated by the head 1 and amplified by the amplifier circuit 2 has an analog waveform as shown in FIG. The comparator circuit 3 has positive and negative peaks at The figure (C1) shows whether the read signal has a value on the positive side or a value on the negative side.
A status signal SS shown in is generated by this comparison circuit 3. This state signal SS is also called a gate signal. On the other hand, the read signal As is also given to the differentiating circuit 4, where it is once converted into the differentiated signal DS shown at +dl in the figure.
Further, the signal is supplied to a zero-crossing detection circuit 5, where a zero-crossing point in the differential signal DS is detected. The zero-crossing signal zS shown in FIG. 6(e) output from the zero-crossing detection circuit 5 has a waveform that switches at the zero-crossing point of the differential signal DS as shown, and the rising and falling points receive the zero-crossing signal zS. The bidirectional one-shot circuit 6 converts it into a zero-crossing pulse train ZP shown in FIG. 3(f). As can be seen from the figure, this zero-crossing pulse train ZP includes zero-crossing pulses corresponding to the positive and negative peaks of the read signal ^S, as well as the zero-crossing pulses corresponding to the positive and negative peaks of the read signal ^S.
As shown in Figure 1, it often includes an extra zero-crossing pulse corresponding to the shoulder sh that tends to occur in the read signal, and this extra zero-crossing pulse is harmful and useless for the purpose of detecting the peak of the read signal AS. In the figure, these are indicated by E surrounded by one point ilI&II. The D-type flip-flop 7 is for removing such extra zero-crossing pulses, and receives the above-mentioned state signal SS at its D input, and is triggered by the zero-crossing pulse ZP applied to its clock C, but when the peak of the read signal It does not respond to extra zero-crossing pulses until it is reset by a zero-crossing pulse corresponding to the next peak in the read signal, for example, depending on the logic value of the differential signal DS. In this way, the state signal S
The composite signal CS as the Q output of the flip-flop 7, which is a logical combination of S and zero-crossing pulse train ZP, has a waveform that switches correctly at the peak position of the read signal, as shown in the same figure (gl), and this composite signal C8 is Furthermore, it is converted into a reproduction signal sequence R5 indicating the rising and falling positions by another bidirectional one-shot circuit 8. As can be seen from the above operation, the reproduction signal indicating the peak position in the read signal As. The basic function of creating R3 is handled by the differentiating circuit 41, the zero-crossing detection circuit 5, and the one-shot circuit 6, but since there is a waveform deformation such as the aforementioned shoulder in the read signal, an extra zero-crossing pulse ZP is generated. is likely to occur, and the status signal SS has the role of masking this extra zero-crossing pulse, so to speak.In order to make the status signal SS sufficiently have this masking effect, the comparison circuit 3 must have a certain history operation width, that is, the operation described above. It is necessary to give a threshold value vth.In other words, if this threshold value is too low, the switching position of the status signal SS will be close to the shoulder where extra zero-crossing pulses are likely to occur, and the rise and fall of the However, this threshold must be kept low enough that it cuts off all peaks in the read signal, i.e. This is because if a certain peak does not intersect with the threshold value vth, the status signal SS will not be emitted, and even the zero-crossing pulse ZP indicating the normal peak will be masked and will not be output as the reproduced signal R3. As mentioned above, the width of the N and 3 areas on the track varies as shown in Figure 4 (a), and if a narrow area is sandwiched by a wide area, as shown in Figure 6 (mountain). The reason why the peak value of the read signal As corresponding to that area becomes low will be explained with reference to FIG. 5. The read signal As is originally a combination of a mountain-shaped positive peak shift p and a negative peak Vn shown by the ui line in the diagram corresponding to the boundaries of the N and 3 areas, and the height of the positive and negative angular peaks is the same as that of the trunk disk. If the radial position within the area is the same, they are all equal, but in a wide area as shown in Figure 5 (al), the positive and negative peaks Vp and Vn are isolated from each other, so the read signal As is However, on the other hand, the above-mentioned shoulder sh tends to occur between the positive and negative angle peaks Vp and νn.On the other hand, as shown in Figure To), where the area width is narrow, for example, the positive peak vp This is canceled out by the bases of the negative peak Vn on both sides, and the peak Pv becomes low. The illustrated example is the case of the above-mentioned MFM method, where the narrow region width becomes the wide region width A, but in other modulation methods, this ratio is much smaller, and the read signal AS in the narrow region width
The peak value of In Figure 6, (as shown in al, the width of the magnetic pattern area in the center of the figure is narrow and the peak value of the read signal AS is in the figure)
As shown in FIG. 4, the voltage decreases in this portion, so the threshold value vth of the comparison circuit 3 is lowered accordingly than in the case of FIG. As can be easily seen, the state signal SS output from the comparator circuit 3 in this case has a waveform shifted to the left in the figure compared to when the threshold value is high. On the other hand, in areas where the region width is wide on both sides of the narrow region width portion, an extra signal due to the shoulder portion sh is generated by the zero cross signal ZS shown in the zero cross detection circuit 5 and the zero cross signal ZS shown at +111 from the one shift circuit 6. To Pulse ZP,
In the figure, it is mixed in as shown by El-E3 surrounded by a dashed line. Of course, the generation position of such an extra zero-crossing pulse does not change much depending on the magnitude of the threshold value vth, but if the threshold value νth is small, the state signal shifts to the front in time as described above. The masking effect of the state signal Ss on extra zero-crossing pulses becomes questionable. In the error signal section El, there is a slight time difference of ΔT1 between the wave tail of the second extra zero-crossing pulse zp and the wave tail of the status signal SS, and the extra zero-crossing pulse is somehow masked. Combined signal cs of +fl in the figure from flip-flop 7 and one-shield/to circuit 8
It has been safely cut from the reproduced signal R3. However, in the error signal section E2, the time difference ΔT2 between the wave tail of the extra zero-cross pulse at the end and the wave front of the state signal is too small, so the flip-flop 7 malfunctions, and the composite signal C
S stands up in advance as shown by the arrow from the normal position shown by the chain line in the figure, that is, flip-flop 7
misidentified the extra zero-crossing pulse as a regular zero-crossing pulse and raised the composite signal, and in response to this, the reproduced signal 1? The same goes for the error signal part E3, where s also appears at an incorrect position, and the time difference ΔT3 between the wave tail of the extra zero-cross pulse and the wave tail of the status signal SSO is too small, so the composite signal C5 falls early. Shimai, E
As a result, an erroneous reproduction signal is generated as shown in . The root cause of the above-mentioned erroneous signals is that the peak value of the read signal becomes low in the narrow area of the magnetic pattern, so if this peak value is raised by applying AGC to the amplifier circuit 2, for example, That's good, but I really need AGC.
Unless the operating speed is increased, it is almost impossible to correct the peak value of the read signal corresponding to a narrow region width that is stuck within a wide region width. As is well known, the AGC used in the amplifier circuit 2 is mainly used to correct fluctuations in the average level of the read signal due to changes in the position of the trunk in the disk inner radial direction, and is used to compensate for fluctuations in the average level of the read signal due to changes in the position of the trunk in the disk inner radial direction. The performance is sufficient to even out the average level of the peaks. As described above, the state signal is useful for masking or canting extra signals that may be included in the zero-crossing pulse, but according to the prior art, the selection of the threshold for creating the state signal is delicate and difficult. If the threshold is even slightly too high, necessary pulses in the reproduced signal will be dropped, and if it is even slightly too low, the reproduced signal may become inaccurate, making it difficult to eradicate errors from the reproduced signal. Ta.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems and provide a read signal reproducing circuit that can reduce the occurrence of errors in reproduced signals compared to the conventional method.

[Key points of the invention]

The present invention has succeeded in achieving the above object by focusing on the fact that the differential signal generated from the read signal has a high peak value even in a narrow area of the magnetic pattern. A differentiation circuit that generates a differential signal, a positive/negative discrimination circuit that compares the differential signal with a positive/negative threshold value and generates a status signal indicating the positive/negative status of the differential signal, and a positive/negative discrimination circuit that detects the zero-crossing point of the differential signal and generates a short-pulse zero-crossing pulse train. A zero-crossing pulse generating circuit that emits, a state signal and one of the zero-crossing pulse trains are shifted by a fixed shift time with respect to the other, and a zero-crossing pulse that matches the positive/negative state indicated by the state signal is selected from the zero-crossing pulse train, and a selection circuit that emits a reproduction signal indicating the switching position of the magnetic pattern from the position of the selected zero-crossing pulse, and the switching point of the state signal is determined by the threshold value in the positive/negative discrimination circuit for the shift time to the zero-crossing of the differential signal. The problems in the prior art can be solved by selecting a state shift time that is longer than the state shift time that deviates from the point and shorter than 2, which is the maximum period of the read signal. As mentioned above, the analog waveform of the read signal tends to have a low peak value in the narrow area of the magnetic pattern, but its period is correspondingly short as the area width is narrow, and its rate of change over time is rather small. In other words, the frequency of the read signal waveform is higher in narrow areas, and in the MFM method, it is twice as high as in wide areas.
In the GCR method, it is tripled. Therefore, although the differential signal obtained by temporally differentiating the analog waveform of the read signal differs depending on the magnetic pattern method, it has a higher peak value in areas where the area width is narrow than in areas where it is wide. In the present invention, a status signal for masking extra zero-crossing pulses is generated from this differential signal by a positive/negative discrimination circuit 9, for example, a zero-crossing detection circuit. Therefore, for the threshold value used when creating this state signal, it is only necessary to select a value for the wide region of the magnetic pattern, and there is no need to lower the value for the narrow region. By selecting the optimum value, extra zero-crossing pulses can be effectively masked. However, as is well known, the differential signal is out of phase with the original waveform, and the state signal created from this has a phase shift from where it should be, and if left as it is, it will be difficult to mask the extra zero-crossing pulse. Therefore, it is necessary to shift the pulse train slightly in time with respect to the zero-crossing pulse train. This phase shift angle is 90 degrees as is known when the waveform of the read signal has a constant frequency, but since the N band width of the magnetic pattern varies and the frequency of the read signal also varies accordingly, the phase shift angle is not constant. However, the inventor of the present invention found that it was not necessary to change the shift time required to correct this phase shift depending on the region width, and succeeded in solving the problem using a constant shift time. The shift time suitable for this phase correction differs depending on the magnetic pattern method and actual circuit conditions, but in terms of -m, as mentioned above, the shift time is determined by the threshold value in the positive/negative discrimination circuit to determine the switching point of the status signal. It is necessary to select a state shift time that is longer than the state shift time that deviates from the zero-crossing point of the differential signal and shorter than the maximum cycle of the read signal. Note that, considering the cause of this phase shift, it is natural to shift the state signal by this shift time, but the point is that it is sufficient to shift the state signal and the zero-cross pulse train relatively. In practice, it is more advantageous to shift the zero-crossing pulse train by this shift time.

Embodiments of the Invention Examples of the present invention will be described below with reference to FIGS. 1 and 2. FIG. 1 is a circuit diagram showing a magnetic recording read signal reproducing circuit according to the present invention together with a hand 1 and an amplifier circuit 2 with AGC, and FIG. 2 is a waveform diagram of main signals in the circuit. In FIG. 1, a read signal of an analog waveform illustrated in FIG. 2(b) from the amplifier circuit 2 is first given to the differentiating circuit 10. As shown in FIG. Specifically, a well-known differential circuit IO consisting of nine combinations of capacitors and resistors as shown in the frame is sufficient, and the differential signal O3 outputted from it is typically shown in FIG. The waveform will be as shown in e). As can be seen by comparing this with Figure 3), for the read signal As, the peak value PVw at the wide area of the magnetic pattern area of the MFM method on the trunk in Figure 1 (al) is higher than the peak value PVn at the narrow area. However, in the differential signal DS, the latter is rather high.Therefore, the threshold value vth to be given to the positive/negative discrimination circuit 20 that receives the differential signal DS is as shown in tc+ in the same figure. A relatively high value can be selected.As can be seen from the diagram, the differential signal O5 is
There is a fairly large depression C at the position corresponding to the shoulder sh of the waveform of the read signal As, which can cause an extra signal to be generated, and if this depression waveform crosses the threshold value unnecessarily, even the waveform of the status signal will be distorted. Because I'm going to come here,
By setting the threshold value vth to a high value as described above, the waveform of the state signal can be made more accurate than before. The specific configuration of this positive/negative discrimination circuit 20 is a comparator circuit with an operation history using an operational amplifier equipped with a known feedback circuit, and the above-mentioned threshold value vth is set by the feedback circuit constant 1, for example, the feedback resistance value. The status signal SS output from the positive/negative discrimination circuit having such a threshold value vth is shown in FIG. 30 may be, for example, a combination of a zero-cross detection circuit and a one-shot circuit in the conventional circuit, and the zero-cross detection circuit 5 generates a zero-cross signal as shown in FIG. 2 +dl, which switches the state at the zero-cross point of the differential signal DS. After obtaining zS, the one-shot circuit 6 generates a zero-cross pulse ZP shown in to+ in the figure which is synchronized with the rise and fall of the zero-cross signal.This zero-cross pulse example Z
P includes an extra zero-crossing pulse due to the shoulder in the read signal waveform or the aforementioned dent in the differential signal waveform, and this is shown as an error portion E surrounded by a dashed line in the figure. Such extra zero-crossing pulses that occur in a practical circuit are much more complex and numerous than shown in the figure, but in the figure, for the sake of simplicity, each error portion is represented by two extra zero-crossing pulses ZP. As shown in FIG. 1, the selection circuit 40 receiving the state signal SS and the zero-crossing pulse train zP is composed of a shift circuit 41, a flip-flop 42, and a bidirectional one-shot circuit 43, of which the shift circuit 41 is A delay circuit that shifts the zero-crossing pulse train ZP for a predetermined time T.
I'll make you wait only s. Zero cross pulse Z after this shift
Ps is shown in FIG. 2(f), including of course an extra zero-crossing pulse. Flip-flop 42 is of type D as in the prior art, receives a state signal SS at its D input, is edge triggered by a shifted zero-crossing pulse ZPs received at its clock input C, and outputs as its Q output a composite signal C8. emanate. By the way, as shown in FIG. 2, the status signal SS is originally a signal that should be switched at the zero-crossing point of the differential signal corresponding to the peak position in the read signal waveform. The same figure te because it is
The rising and falling points are shifted from this normal switching position by the state shift time ΔT shown in l. The above-mentioned shift time Ts is set to be larger than this state shift time ΔT, and in this example, it is set to 2 to 3 times the state shift time. The zero-crossing pulse ZPs shifted by this amount does not overlap the switching point of the state signal and is reliably logically combined with the state signal SS after switching. Therefore, the flip-flop 42 is edge-triggered by the first regular zero-crossing pulse ZPs that arrives after the state signal SS switches, and is centered or reset depending on the state the state signal indicates at the time;
As shown in FIG. 2(h), a composite signal C3 is generated that ignores all the extra zero-crossing pulses ZPs that arrive thereafter. Since all the extra zero-crossing pulses are masked by the state signal in this way, only the reproduced signal R5 corresponding to the normal peak position in the read signal waveform is output from the one-shot circuit 43 that receives the composite signal C8 as shown in FIG. 2(i). It is emitted as shown below. As can be seen from the figure, all of the reproduced signals R8 are temporally deviated from the peak position of the read signal by the aforementioned shift time Ts, but since the amount of deviation is always constant, there is no problem at all in practice. In addition, shift time T! in this example. It is better not to set it too large. If this setting value becomes too large, the end of the extra zero-crossing pulse will overlap or interfere with the next switching of the status signal, causing an error, as shown in error part E in the figure. A signal may be generated. However, as is easily understood, when the occurrence of extra zero-crossing pulses is small, theoretically the maximum area width in the magnetic pattern, that is, the maximum period A of the read signal.
It is necessary to make it shorter than . In addition to the embodiments described above, the present invention can be implemented in various forms. The positive/negative discrimination circuit does not necessarily need to be configured as a comparison circuit that performs a history operation, and if the well-known fact that the read signal is normally output from the amplifier circuit 2 as two differential signals is used, it can be configured as a comparison circuit that performs a history operation. It consists of a comparison circuit or a comparator, each of which compares the differentiated signal of the differential signal with a threshold given to each, and then
The comparison outputs can be combined into a state signal. Further, the selection circuit can be configured by combining various known circuits, and the shift circuit may also be configured to shift the state signal instead of temporally shifting the zero-crossing pulse train.

【Effect of the invention】

As can be seen from the above description, according to the present invention, a status signal for masking the extra zero-crossing pulse that is likely to be mixed in with the zero-crossing pulse indicating the peak position of the read signal waveform is generated from the read signal itself, unlike the prior art. Since we created it from the differential signal instead of creating it, we set a relatively high value as the threshold value that is the standard when creating the state signal.
As a result, the masking effect of the state signal against extra zero-crossing pulses can be made more reliable than in the past, and the probability that errors will occur in the reproduced signal can be reduced. The effects of the circuit of the present invention are hardly affected by the magnetic pattern method used in magnetic recording, and in practice, the same circuit can support various modulation methods such as MFM, RLL, or OCR.

[Brief explanation of the drawing]

Figures 1 and 2 are for explaining the present invention, Figure 1 of the inner tube is a circuit diagram showing an embodiment of the magnetic recording read signal reproducing circuit according to the present invention together with related circuits, and Figure 2 is a circuit diagram showing the inside of the circuit. FIG. 2 is a waveform diagram of main signals of FIG. Figure 3 and subsequent figures are for explanation of the prior art. Figure 3 is a block circuit diagram showing an overview of a magnetic recording read signal reproducing circuit according to the prior art, and Figure 4 is a waveform of main signals in the circuit showing its normal operation. 5 is a waveform diagram of the read signal showing the correlation between the area width of the magnetic pattern and the read signal,
FIG. 6 is a waveform diagram of the main signals in the circuit showing the pattern of error occurrence in the reproduced signal in a conventional circuit.
4: Differential circuit, 5: Zero cross detection circuit, 6: One shift circuit, 7: Flip-flop, 8: One shift circuit, 10+ differentiator circuit, 20: Positive/negative discrimination circuit, 30: Zero cross pulse generation circuit, 40: Selection circuit, 41: Shift circuit, 42: Flip-flop, 43: One-four-bit circuit, AS=read signal, C: dent in differential signal waveform, C1
Synthetic signal, DS: Differential signal, E, El-mB3: Part of error in signal, Pv: Peak value of waveform, PVn: Peak value of waveform corresponding to narrow part of magnetic pattern region width, PV
w: peak value of the waveform corresponding to the wide part of the magnetic pattern area, R3: reproduced signal, Sh: shoulder in the read signal waveform,
SS: Status signal, Tj) rack, ↑3: Shift time, Δ
T: Tj shift time, ΔT1 to ΔT1 time difference, Vp, V
+: positive and negative peak voltage, Vth: L threshold, ZP: zero-crossing pulse or zero-crossing pulse train, ZPs: shifted zero-crossing pulse, ZS: zero-crossing signal,
It is. Mt. Kuu Figure 2 Figure 4

Claims (1)

[Claims] A differentiation circuit that temporally differentiates an analog waveform read signal read out from digital information recorded in a magnetic pattern in a predetermined format on a magnetic storage medium and converts it into a differential signal of the analog waveform, and a differential signal that converts the differential signal into a differential signal of the analog waveform. a positive/negative discrimination circuit that emits a status signal indicating the positive/negative status of the differential signal in comparison with the differential signal;
A zero-crossing pulse generation circuit detects the zero-crossing point of a differential signal and generates a short-pulse zero-crossing pulse train, and a zero-crossing pulse generation circuit generates a short-pulse zero-crossing pulse train by shifting one of the state signal and the zero-crossing pulse train with respect to the other by a fixed shift time and then generating a zero-crossing pulse train from within the zero-crossing pulse train. a selection circuit that selects a zero-crossing pulse that matches the positive/negative state indicated by the state signal and generates a reproduction signal indicating the switching position of the magnetic pattern from the position of the selected zero-crossing pulse; A magnetic recording read signal reproduction characterized in that the threshold value is selected to be longer than the state shift time in which the switching point of the state signal deviates from the zero crossing point of the differential signal and shorter than 1/2 of the maximum cycle of the read signal. circuit.