DE1217440B - Method and circuit arrangement for the transcoding reading of a message encoded in the two-frequency code or in the Williams code - Google Patents

Description

FEDERAL REPUBLIC OF GERMANY

GERMAN

PATENT OFFICE

EDITORIAL

Int. σ .:

H 03 k

German KL: 21 al - 36/20

Number: 1217440

File number: S 81315 VIII a / 21 al

Filing date: September 6, 1962

Opening day: May 26, 1966

In data processing technology, but also in other areas of communications technology often record or reproduce large amounts of information quickly. For this are multiple Magnetic layer storage, for example magnetic tape, magnetic drum or magnetic disk storage, used with very high recording density. The recording density depends on the particular recording method away. A particularly high recording density can thus be achieved when using the be ίο knew binary non-return-to-zero font, in particular the Williams font or the two-frequency font.

With the usual return-to-zerc font, the binary signal one through a pulse-shaped positive magnetization, the binary signal zero by a pulse-shaped negative magnetization on the one provided for recording the signal Position of the recording medium shown. Because of the very short current impulse with every informa- ao tion element, a very steep write current edge is necessary. Hence the procedure for special high recording frequencies inexpedient.

In the advantageous and therefore mostly preferred non-return-to-zero font, the binary signal as One is represented by positive magnetization, the binary signal zero by negative magnetization. An uninterrupted sequence of binary ones results in a positive sequence magnetized places. A change in magnetization only occurs in connection with a change in information on. - There is also a variant of the non-return-to-zero font, in contrast to the previous one described so-called mark font the so-called change font, in which for each to be recorded binary information element one exactly one change, d. H. a reversal of magnetization, however, there is no change for binary zeros. So there is a certain binary information not always assigned the same magnetization state.

For example, in the case of a sequence of binary ones, the magnetization changes continuously.

With the Williams script, a binary pulse phase modulation, one or two magnetization changes are possible in each recording cycle interval Γ, i.e. for each binary information element. A binary one is represented by a negative magnetization during the first half and a positive magnetization during the second half of the clock interval T , a binary zero by a positive magnetization during the method and circuit arrangement for
transcoding reading one encoded in the two-frequency code or in the Williams code
news

Applicant:

Siemens & Halske Aktiengesellschaft,

Berlin and Munich,

Munich 2, Wittelsbacherplatz 2

Named as inventor:

Dipl.-Ing. Attila Nowak, Munich

first and negative magnetization during the second half of the clock interval. When reading the Recording, the signal edge occurring in the middle of the clock interval is evaluated. She is at a binary one a positive edge, d. H. a transition from negative to positive magnetization, a negative edge in the case of a binary zero.

With the two-frequency font, a binary frequency modulation, one or two magnetization changes are possible in each recording cycle interval T. A binary zero is represented by a single change in magnetization within a clock interval, a binary one by two changes in magnetization. When reading the recording, a check is made as to whether there is a signal edge in the middle of the clock interval. In this case a binary one is decoded, otherwise a binary zero. Polarities do not need to be taken into account.

According to a known method for reading recordings that are coded in the two-frequency script, a sampling clock pulse (gate pulse) with a duty cycle of 1: 1 is first formed from the read & signal pulses, the pulses of which lie in the middle of each recording clock interval T , namely from 0.25 · T to 0.75 · T. If a signal edge coincides with one of these gate pulses of width 0.5 · T , it can only be an edge that identifies binary information one. In principle, those from the maximal and

609 570/487

ί 217 440

The signal pulses derived from the read voltage are clearly assigned to the pulse-shaped voltages of the writing process, but experience shows that their flanks show very strong temporal shifts, which become more pronounced the higher the storage density is selected.
The permissible tolerances for the position of the evaluating pulse edge, based on the position of the> r pulse, are only ± 0.25 · T, because the Tori pulse is only triggered by the preceding read signal pulse, i.e. it cannot begin earlier than 0.25 · T before the target position. On the other hand, the gate pulse must not begin later than 0.25 · T i of the nominal position, because otherwise a read signal edge (binary signal 1) to be evaluated could not be determined. A broadening of the pulse and thus an apparent improvement in the tolerance field is ruled out because of the risk that a subsequent read-and-write pulse coming too early will still coincide with the gate pulse and then no longer trigger the next gate pulse. This would cause the next information item to be incorrectly decoded. The tolerances m ± 0.25 · T are too small, especially for recordings with a very high recording density.
The object of the invention is to increase the security of the coded reading of messages that are binary-encrypted in the two-frequency code [he Williams code. According to the invention, this is sufficient if the evaluation interval is always larger than the recording clock interval and the number of changes in the input variable detected during each evaluation interval is evaluated as information.
In a particularly expedient embodiment of the invention, each edge of the input variable to be identified is determined (that is, not just the signal edge near the middle of the recording clock interval T ) and defines the beginning of an input value interval. Due to the two-sequence or Williams code, one or more edges occur in each off-Mrteintervall. de starts an evaluation interval. In order to reduce the number of clock pulses i abzuleitenn from the evaluation intervals on the number of Informationsemiten, the intervals have the same total when only two intervals catch one at a time; in the case of more overlapping intervals, s is ended as the second interval started at the same time as the end of the Inter-Us started first. An output clock pulse is derived from each evaluation interval, preferably at the end, and the output information is obtained from a flow signal which is generated during the overlap time of the evaluation interval.

In a preferred embodiment of the sound arrangement According to the invention, at least 3i, preferably similar, switching elements are to be provided, which are activated cyclically during the evaluation period. For example, can monostable multivibrators are provided, of which one cyclically each at the beginning of an interval is set in the working position and ch an adjustable period of time, which is equal to the desired duration of the evaluation interval, automatically returns to the rest position or at the same time as Reset at the regular end of the previous evaluation interval, if more than two flip-flops at the same time in the working position are. Because of the recovery time of the monostable, which is necessary for an exact adherence to the time span But flip-flops are the use of five

monostable multivibrators are particularly advantageous.

The evaluation interval is expediently about

Made 1.35 times as long as the recording clock interval. The reason for this results from the

ίο further description.

The main advantage of the reading arrangement according to the invention is that which it enables large tolerance with regard to the position of the signal edge occurring in the middle of the clock interval, which identifies binary information one. It is precisely this flank that is strong in its position Subject to fluctuations, because this is the point in time when the highest transmission frequency occurs. The transmission system, consisting of writing

ao amplifier, write head, magnetic layer, read head and read amplifier have similar physical properties such as one consisting of a write amplifier, low-pass filter, attenuator and read amplifier Arrangement. To the greatest possible storage density

To achieve this, the writing frequency is chosen to be very high, ie almost equal to the cut-off frequency of the transmission system. With such a high transmission frequency, there is a strong phase shift and time shift (similar to a disruptive time modulation) and thus a considerable uncertainty in the position of the middle edge. In the solution according to the invention, according to which several successive signal edges are evaluated in context, so to speak averaged over time, the position of the middle signal edge may in principle be -0.5 T or +0.85 T from the underlying numerical values Nominal position differ, i.e. at least twice as much as in the previously known solution, according to which only a single signal edge is evaluated on its own. The Williams and the two-frequency font do not show any fundamental differences as pulse-shaped voltage curves.

Only the assignment of whether a certain information element is a binary one or a binary zero represents is different. Because the application of the two-frequency font explained for the example Reading method in a similar way for the Williams script is not possible in the individual received.

Further details and advantageous developments of the invention are detailed below Description of an exemplary embodiment according to the invention can be found on the basis of the drawings. It shows

F i g. 1 shows the basic mode of operation of the previously known circuit arrangement for reading in Binary coded information recorded in two-frequency writing,

F i g. 2 shows the principle of operation of a circuit arrangement according to the invention for reading from binary coded information recorded in two-frequency script and

3a and 3b show the block diagram of a special

len embodiment of a circuit arrangement according to the invention.

In FIG. 1 is the basic mode of operation of the previously known circuit arrangement for reading

of information (line 2) recorded in two-frequency writing (line 1). line 1 can be understood as a representation of the pulse-shaped voltages during the writing process. At the Reading is performed by peak scanning or similar methods at each maximum or minimum of the Read voltage generates a pulse edge. This creates information impulses, which are impulsive The voltages of the writing process. Therefore, line 1 can also be called differentiated and limited output voltage can be perceived when reading. However, then usually occur Displacements of the pulse edges with respect to their nominal positions, which are larger, the higher the storage density is.

The first signal edge that occurs triggers a delay chain or a sampling clock pulse (line 3). The difficulty lies in the fact that special circuit means must be used to prevent a middle edge, which identifies binary information one, does not also trigger the sampling clock pulse, otherwise there would be too many sampling clock pulses.

Clock pulses and information pulses are emitted at the output of the reading arrangement, and one clock pulse per recording clock interval (line 4) and one information pulse per read binary information element one (line 5).

In FIG. 2 is the principle of operation of a circuit arrangement according to the invention for Reading the information recorded in two-frequency writing (line 1) (line 2) is shown. In principle, a clock pulse is generated by each signal edge (line 3) as well as in a cyclical manner Follow the next of a series of, for example, five monostable multivibrators in the working position is set (lines 4 to 8). Each monostable multivibrator is either after a period of time equal to 1.35 times the recording clock interval reset to the rest position or already with the regular resetting of the monostable multivibrator that was operated just before, if three monostable multivibrators are in the working position at the same time (lines 4 to 8). By this measure it is achieved that a reading clock pulse (line 9) from the lapping back of the can derive monostable multivibrators in the rest position. Otherwise it wouldn't be an even Clock pulse sequence arise, but at those places where binary information One is recorded, one clock pulse too many. In the following, let us assume that from the tilting back the monostable In contrast to the clock pulse derived from flip-flops, referred to as a secondary reading clock pulse the primary reading clock pulse derived directly from each signal edge. The primary reading pulse also contains the information, the secondary reading pulse only indicates the clock. A bistable Toggle switch (line 10) is activated by each clock pulse of the primary reading clock pulse and by each clock pulse of the secondary Reading pulse is set to the rest position. The read Information can be determined. If another bistable trigger circuit (line 11) is dependent of the aforementioned output signal at those times when the clock pulses of the primary reading cycle occur, set in the working position and generally at those times in which the clock pulses of the second reading clock pulse occur, reset to the rest position, so sets the output signal of this bistable multivibrator represents the read information in return-to-zero form. Another bistable multivibrator (line 12) depending on the last-mentioned output signal set or reset by the second reading pulse, the output signal represents this

ίο bistable toggle switch the read information in non-return-to-zero form.

In order to explain the considerable increase in the permissible tolerance range for the position of the pulse edges, which is achieved with the arrangement according to the invention, all combinations of two consecutive characters Z ₁ and Z ₂ are discussed below. However, three empirical facts are given in advance.
First, the first part of a binary one always seems too long after a sequence of binary zeros. Second, the second part of a binary one always appears too long before a sequence of binary zeros. Third, the pulse width of a single binary zero enclosed in a series of binary ones is reduced.

Now let us first consider the case ₁ = I, Z ₂ = I. The position of the two pulses within the query time (1.35 · Γ) is completely irrelevant. In order to prevent the first pulse of the character Z _{2 from being} counted, the evaluation circuit is switched back to its basic position _{after the second pulse of the character Z 1.}

Second, consider the case Z ₁ = I, Z ₂ = O. The extension of the interrogation time to 1.35 · T has a favorable effect here, since, as a result of the second empirical fact mentioned above, the second pulse of the character Z ₁ can arrive very late. Thirdly, consider the case Z ₁ = O, Z ₂ = O. With this character string, only the lowest recording frequency occurs. Therefore, in this case, the available tolerance fields are hardly used. The case poses no difficulties.

Fourth, consider the case Z ₁ = O, Z ₂ = I. This is the most difficult reading case. Here the first edge of the character Z ₂ would be expected after the time 1.5 · T. In order not to evaluate it incorrectly in connection with the first character, the query time must be limited to a value that is sufficiently smaller than 1.5 · T. Hence the choice of the maximum query time of 1.35 · T. The remaining security initially appears to be very low. However, as a result of the first empirical fact cited above, the tolerance conditions are significantly eased in this particular case, namely insofar as the first flank of the character Z ₂ appears too late and therefore by no means before the end of the previous evaluation interval. Practical investigations have shown that this type of character recoding according to the invention works very reliably. A pre-equalization of the characters for the purpose of eliminating information-dependent errors therefore turned out to be unnecessary.
In Fig. FIGS. 3 a and 3 b show a circuit arrangement according to the invention for reading information recorded in two-frequency writing on a magnetic tape. The differentiated and limited, i.e. square-wave read signal voltage

is converted in a clock shaping stage 1 into information clock pulses T ₁ , which are often required in the remaining parts of the circuit arrangement. You are, for example, on five bistable multivibrators 2 to 6 of a control chain. This control chain creates five monostable multivibrators? up to 11 in a cyclical sequence in the working position. They are caused by control circuits 12 to 16 to be reset to the rest position after an adjustable time, preferably 1.35 · T. A bistable multivibrator 17 is used to generate a binary signal one circulating in the control chain once by means of an auxiliary signal Έ when a read command E and an information clock T _{1 are} present. The trailing edges of the pulses emitted by the monostable multivibrators 7 to 11 are combined in a mixer 18 and, amplified by a clock amplifier 19, result in the reading clock T _L.

Each information clock pulse T ₁ resets a flip-flop 20 to the rest position (FIG. 3b). If, however, two clock pulses T ₁ occur between two clock pulses T _L , a bistable flip-flop circuit 21 is additionally set to the working position and reset to the rest position by the next following clock pulse T _L. Therefore, a pulse occurs at the output of the bistable multivibrator 21 only if the information read was one. The output signal of the bistable multivibrator 21 is therefore recoded in return-to-zero writing, with impulses for each binary information one. With the aid of a 1-bit generator 22, an output voltage in non-return-to-zero representation is generated from the output signals of the flip-flops 20 and 21, the read command E and a read clock pulse Tj.

The read cycle for decoding the two-frequency writing is started by pressing a key 23. As a result, a bistable trigger circuit 24 is set in the working position, which in turn sets a bistable trigger circuit 25 in the working position. The bistable multivibrator 25 puts the belt motors into operation (which are not shown in the drawing). With the aid of the auxiliary signals Z ₁ derived from the bistable multivibrators 24 and 25, a monostable multivibrator 26 is set into the working position. When this monostable multivibrator 26 in is tilted back, the magnetic tape has meanwhile reached its set speed. Then the output signal of the monostable flip-flop 26, together with *, an information pulse T ₁ , sets a flip-flop 27, the read flip-flop, whose output signal is the read command L. Then the actual, already described decoding process begins (reference numerals 1 to 22).

In order to put the bistable flip-flops 24 and 25 into operation before the magnetic tape is set in motion, a clock pulse Γ is required. This is with an oscillator 28 and a HiMe Clock shaping stage 29 is generated.

The evaluation intervals can be determined by means of the monostable multivibrators 7 to 11 also be done by counters. In special cases it can be advantageous to change the time intervals to measure and save between neighboring pulses analog or digital and then through a data processing system in connection with one another, in particular taking into account systematic error to be evaluated.

The invention has been explained on the basis of specific exemplary embodiments. Of course there are many Modifications, further developments, combinations with known and the like without further ado conceivable.

So it can be z. B. with appropriate modification particularly advantageous also with submarine cable or Apply local cable telegraphy.

Claims (7)

Patent claims: 1. Method for transcoding a reading in the two-frequency code or in the Williams code binary encrypted message, preferably for the recovery of a message using two-frequency writing or by Williams script in a high density magnetic layer memory stored information, characterized in that the evaluation interval each larger than the recording cycle interval and the number of during changes in the input variable occurring at each evaluation interval are determined and provided as information is evaluated. 2. The method according to claim 1, characterized in that each edge of the to be evaluated Input variable is determined and defines the beginning of an evaluation interval that in at least one further edge occurs in each evaluation interval and one further evaluation interval lets begin that the intervals have the same length if only two intervals correspond in time overlap, but if there are more intervals, the second interval started at the same time as the regular end of the interval started first is ended that at the end of each evaluation interval an output clock pulse is generated and that the output information is derived from an auxiliary signal coming about during the overlapping times of the evaluation intervals will. 3. The method according to any one of the preceding claims, characterized in that the Regular evaluation interval is 1.35 times as long as the recording interval. 4. Circuit arrangement for performing the method according to one of claims 1 to 3, characterized in that at least three preferably similar switching elements are provided that are activated cyclically for the duration of the evaluation intervals. 5. Circuit arrangement according to claim 4, characterized in that as switching elements at least three monostable flip-flops (7, 8, 9, 10, 11) are provided, one of which is each The beginning of an evaluation interval is set in the working position and after an adjustable one Period of time equal to the desired duration of the evaluation interval automatically returns to the rest position or reset at the same time as the regular end of the previous evaluation interval if more than two flip-flops are in the working position at the same time (Fig. 3a). 6. Circuit arrangement for performing the method according to one of claims 1 to 3, characterized in that the evaluation intervals are determined by counters. 7. Circuit arrangement for performing the method according to one of claims 1 to 3, characterized in that the distances between adjacent read voltage pulses are measured with the aid of counters be determined and that from them with a data processing system under consideration the inner connections, especially 10 special under elimination of systematic errors, the binary information is determined will. Considered publications: British Patent No. 851520;
"IRE International Convention Record", 8, Part 9, 1960, pp. 179 to 185. For this purpose 2 sheets of drawings