Magnetic Recording Data Decoding System

This invention relates to magnetic data recording and more particularly to a system for decoding a multitrack data recording having one or more data tracks and a clock track.

In the recording of data on magnetic tape or other magnetic storage medium, a separate clock track is often employed in association with one or more data tracks. The clock information should be in predetermined phase relationship with respect to a recorded data track so that the data can be properly decoded and strobe pulses generated for appropriate clocking of the data signal in subsequent signal processing. A phase error between the data and clock tracks can cause erroneous decoding in that the clock pulses are not in correct phase relationship with the data being strobed. Such phase error is usually caused by misalignment of the recording head with respect to the magnetic tape or other recording medium and which may arise from misalignment of the respective head gaps associated with the two tracks or tilting of a multiple head structure with respect to the recorded tracks. The skew may also be of dynamic nature by reason of misalignment between the recorded tracks and the recording head as the recording medium moves past the head. Static skew error is less of a problem if the same recording apparatus is employed for both recording and reproduction; however, if a data recording is recorded on one apparatus and reproduced on another apparatus, the opportunity for greater skew error results. Especially at high bit densities of recorded data sources of skew error can exceed one bit length causing significant decoding errors.

Briefly, the present invention provides a decoding system for use with a magnetic data recording having one or more data tracks and a clock track wherein clock or strobe pulses are produced in intended phase relationship with the decoded data. A measure of the clock pulse period is continuously derived from the clock track and processed in association with information derived from the data track to produce clock pulses in intended phase relationship with the decoded data irrespective of the phase between the data and clock tracks.

The invention will be more fully understood from the following detailed description taken in conjunction with the accompanying drawings, in which:

FIG. 1 is a diagrammatic representation of a decoding system according to the invention;

FIG. 2 is a diagrammatic representation of a modification of the embodiment of FIG. 1; and

FIGS. 3a through 3i are waveforms useful in illustrating operation of the invention.

The invention is employed in the illustrated embodiment with a two-track magnetic data recording having a data track encoded in a non-return to zero (NRZ) format and an associated clock track encoded with clock pulses in properly phased relationship with the data track. This recording format is itself known in the art. A data stream is illustrated in FIG. 3a, the NRZ recorded representation of which is shown in FIG. 3b and is of the form recorded on the data track of a magnetic tape or other magnetic storage medium. The clock track illustrated in FIG. 3c is recorded on an associated clock track of the magnetic tape or other recording medium, the clock pulse being centered about the data transition times of the data track. When read, data from the two tracks of the recording medium are each differentiated to provide the respective waveforms of FIGS. 3d and 3e. The positive and negative pulses of the differentiated signal of FIG. 3d are respectively processed to provide first and second input data pulses D.sub.in (+) and D.sub.in (-) as shown in FIG. 3 f and FIG. 3g respectively. The readout of data from the data track of the recording medium and the derivation of the data pulses is accomplished in any manner well known in the art.

The data pulses of FIGS. 3f and 3g are applied to corresponding inputs of the novel decoding system shown in FIG. 1. The positive input terminal of the system is connected to the set terminal of a flip-flop 10, the reset terminal being connected to the negative input terminal. The Q output of flip-flop 10 provides the decoded data output D.sub.out of the system. An OR gate 12 has first and second inputs connected to respective positive and negative input terminals and a third input provided by a presettable down counter 14 when this counter reaches a zero or other predetermined state. The clock signal shown in FIG. 3e is processed to provide single polarity clock pulses which are applied to the clock (C.sub.in) input which is connected to the reset terminal of an up counter 16 and to the load terminal of a latch 18. The up counter 16 is clocked by a system clock 20 and provides an output count to latch 18. The latch 18 is coupled to down counter 14 and to a divide-by-2 circuit 22, the output of which is applied as one input to a comparator 24. The other input to the comparator 24 is provided by the output of down counter 14. The system clock 20 also provides a clock signal to counter 14 while a load signal for counter 14 is provided by the output from OR gate 12. The output C.sub.out (FIG. 3i) of comparator 24 is the clock output which according to the invention is in precise timed relation with the data output D.sub.out (FIG. 3h).

The clock pulses provided on the C.sub.in terminal are each operative to cause loading of the then count stored in counter 16 into latch 18 and to thereafter reset counter 16. The counter 16 is clocked by the internal system clock 20 which operates at a higher rate than the external clock input derived from the clock track of the recording medium. Thus each clock pulse derived from each transition of the clock track signal causes loading of latch 18 representing the period of the previous two clock pulses.

Upon each data pulse, the down counter 14 is loaded with the then value stored in latch 18. The down counter 14 counts down toward zero and should reach zero one bit time later, when a subsequent data pulse may or may not occur. If a data pulse does occur, the counter 14 is again loaded with the then count of latch 18 and clocked downward toward zero. If no data pulse is received, the counter 14 reaches zero causing provision of a command which is conveyed by OR gate 12 back to the counter 14 to cause reloading thereof with the value then stored in latch 18. The counter 14 is loaded at a rate equal to the frequency of the pulses derived from the clock track but in phase with the pulses derived from the data track. In the absence of data pulses, the counter 14 is recycled automatically in phase with the last received data derived from the data track. As a result, the output clock pulses C.sub.out are produced in intended phase with the decoded data even if no data transitions are occurring. In NRZ recording, a string of zeroes or ones on the data track exhibits no transitions. The count stored in latch 18 is processed by divide-by-2 circuit 22 to provide an input to comparator 24 equal to half the value stored in latch 18. When counter 14 counts down to a value equal to the half value from divider 22, comparator 24 provides a C.sub.out clock pulse. Thus, the comparator 24 provides output clock pulses C.sub.out which have the same period as those derived from the clock track of the recording medium but which are in intended phase relation with the decoded data D.sub.out derived from the data track. In practice, the division of the output count from latch 18 can be most efficiently accomplished by a "shift right" operation as is known to those versed in the art.

A clock pulse is produced by the novel decoder which occurs in intended time relationship with the decoded data irrespective of phase errors between the data and clock tracks being decoded and variations in the speed of the recording medium. The system clock 20 is preferably related to the clock rate of the recorded clock track such that counter 16 is clocked to about one-half of the expected clock period to thereby permit correction of tape speed variations and clock pulse phase variations of up to one hundred percent.

In the embodiment of FIG. 1, the counter 16 is preset in accordance with the previously measured clock period and there is derived therefrom a measure of the midpoint of the measured clock period at which time a clock pulse is regenerated. In alternative embodiment, the counter 14 counts down through a fixed reference value corresponding to the midpoint of the nominal clock period. Referring to FIG. 2, the count from counter 14 is applied to one input of comparator 24 which also receives a reference count from reference source 26 which is a fixed count equal to one-half of the nominal expected clock period. An output C.sub.out is provided by comparator 24 when the value received from counter 14 equals the fixed count from reference source 26.

The novel decoding system provides decoding of data and clock information from the respective tracks of a magnetic recording medium in a manner independent of static skew or tape speed errors. The novel system is also immune to dynamic phase jitter provided the phase error does not change at a rate greater than about fifty percent of a bit time per bit, which cannot reasonably occur in actual recording transport apparatus since the degree of error which can occur within a single bit period is limited by the physics of the transport mechanism.

For purposes of illustration in FIG. 1, circuitry is shown of four bit accuracy; in practice, a greater bit accuracy, typically six to eight bits, is employed to achieve intended performance. In an eight bit system using magnetic tape moving at 10 inches per second, such as in the widely employed Philips cassette, and a bit density of 1600 bits/inch, the clock pulses C.sub.in are typically 16 KHz in frequency with the system clock being 2.048 MHz. With these parameters, a quantizing error results of approximately one percent which is completely satisfactory in practice. The quantizing error is not cumulative even in the absence of data transitions since a measure of clock period is being continuously updated in accordance with pulses derived from the clock track.

The embodiment described above is employed with a recorded data track having a non-return to zero (NRZ) recording format in association with a separate clock track. The invention is also useful for other types of recording formats such as NRZI recording or other similar multitrack data recording formats in which the decoded data in order to generate appropriate strobing pulses for correct processing of the data.

It will be appreciated that the invention can be variously implemented to provide the novel decoding according to the invention. For example, a system clock can be employed of a first frequency for driving counter 14 and for a second frequency one-half of the first frequency for driving counter 16 in order to define the intended phase relationship between the decoded clock and data information. As a further alternative, the clock track can be decoded so that the C.sub.in pulses are only obtained from clock transitions of one polarity and the frequency of counter 16 being one-quarter the frequency of counter 14 to yield the requisite phase relationship. Accordingly, it is not intended to limit the invention by what has been particularly shown and described except as indicated in the appended claims.