FR2642588A1 - Device for shaping frequency analog signals, with high dynamic range - Google Patents

Abstract

The present invention relates to a device for shaping a frequency analog signal, characterised in that it comprises means S defining a variable reference threshold V.ANA, the changes in which, governed by the level of the input signal, follow a non-linear sequence, and a comparator 10 which receives from a first input 12 an input signal to be shaped and which receives on a second input 14 the reference threshold V.ANA.

Description

The present invention relates to a setting device
form of analog frequency signals.

In the context of the present invention, the shaping
of analog frequency signals is to transform the signals
analog to square logic signals, toggling between two levels, with the same frequency as analog signals.

The purpose of shaping is therefore to eliminate all signals
low amplitude noise possibly superimposed on the useful signal.

According to the present invention, the shaping of signals is carried out using a device which comprises - means defining a variable reference threshold, the
changes driven by the level of the input signal follow a sequence
non-linear, - a comparator which receives on an initial input, an input signal
to be shaped and which receives on a second input the threshold of
reference.

 Other characteristics, objects and advantages of the present invention will appear on reading the detailed description which follows and with reference to the appended drawings, given by way of nonlimiting examples and in which - Figure I represents a general view schematic, in the form of functional blocks, of a device according to a first embodiment of the present invention, - FIG. 2 represents a truth table of a FORSAX signal which controls the door circuit, - FIG. 3 represents a truth table of POM and NOC signals which respectively control the adder-subtractor means and the register means, - Figure 4 schematically illustrates the main events occurring during each sampling period, - Figure 5 shows an example of realization of the logic means, - in FIG. 6 represents an example of an embodiment of the additive-sub-tractor means.

- Figure 7 shows an exemplary embodiment of the network, - Figure S represents a timing diagram of signals obtained using the device according to the first embodiment, - Figure 9 shows a general schematic view, in the form of functional blocks , of an improved device according to a second embodiment of the present invention, capable of processing two input signals, - Figure 10 shows an exemplary embodiment of means defining the reference threshold for the device illustrated in the figure 9, - FIG. 11 represents a timing diagram of the signals obtained with the device according to the second embodiment, - FIG. 12 diagrammatically represents a module making it possible to reset the reference threshold to zero, - FIG. 13 diagrammatically represents a module device output. and FIG. 14 schematically represents means defining the different clock signals from an internal clock signal.

FIRST EMBODIMENT: FORMATION OF A SINGLE
ENTRY SIGNAL
The device shown in FIG. 1 comprises a comparator 10 and means S defining a reference threshold.

 Comparator 10 receives on a first non-inverting input 12 an input signal to be shaped.

 The reference threshold from the means S is applied to the second inverting input 14 of the comparator 10.

 Comparator 10 delivers at its output 16 a high level signal if the amplitude of the input signal is greater than the reference threshold. Conversely, the comparator 10 delivers at its output 16 a low level signal if the amplitude of the input signal is less than the reference threshold.

 The signal obtained at the output of comparator 10 is sampled by a signal D4. This signal D4 is generated by clock means shown schematically in FIG. 14 from an Internal clock signal: H. Internal.

 As indicated above, the means S defining the reference threshold comprise a network 100, a gate circuit 200, register means 30u, adder / subtractor means 400 and logic means 500. These logic means 500 are sensitive to the level detected on the output 16 of the comparator 10 during three consecutive sampling periods.

 The sampled signal available at output 16 of comparator 10 at a given time is referenced KN.

 The sampled signals available at the output 16 of comparator 10 during the two immediately preceding sampling periods - are referenced KN - I and KN - 2. These last two signals are stored respectively in registers 600, 65G, at the rate of the signal sampling D4.

 The function of network 100 is to define at its output the appropriate reference threshold. Its output 102 is connected to the reference input 14 of the comparator IO. More precisely, the network 100 has the function of transforming a TODAC digital signal of X bits which it receives on its inputs 101, into an analog reference signal V. ANA, available on its output 102. The network 100 can thus define 2X predetermined reference thresholds.

 Preferably, the network 100 is formed of a resistive network which receives at its input 101 a digital word A of control of X bits (corresponding to the signal TODAC) and delivers at its output 102 an analog reference threshold V. ANA equal to B x CA The term B determines the minimum analog threshold, obtained when A equals 0. For example B equals 114mV. Preferably C equals 1.4. In this case, the 2X reference thresholds correspond to a sequence of geometric progression 1.4.

 FIG. 7 shows an exemplary embodiment of a network 100.

 According to the embodiment shown in this FIG. 7, the network 100 includes 9 resistors referenced R 110 to R 118, connected in series between two supply terminals CRN and VRF, the origin of which will be explained below.

The output 152 of the network 155 delivers the analog signal
V.ANA as previously indicated. This output 1G2 of the network 100 is connected to the intermediate points of the resistive divider bridge R110 to R118 by a network of solid-state switches S12G to SI 33.

 These switches 5120 to 5133 are controlled by the X-bit TODAC signal from the gate circuit 2û0.

 According to the illustrated embodiment, the word TODAC has three bits referenced respectively TODAC O, TODAC 1 and TODAC 2.

Note that some of the solid state switches
S 125 to S 133 are controlled by a bit complemented by the TODAC signal.

This complement is obtained at the output of inverters 14G, 141, 142 respectively.

 The arrangement of the network of solid-state switches shown in FIG. 7 corresponds to a particular preferential but non-limiting arrangement. Many other configurations are possible. For this reason, the interconnections of the network of solid-state switches will not be described in more detail below.

Those skilled in the art will readily understand that the network 100 shown in FIG. 7 makes it possible to transform the digital signal
X-bit TODAC into a V.ANA analog reference signal.

 Preferably, the resistance values R 110 to R 118 and the structure of the network of transistorized switches are chosen such that the amplitude of the analog reference signal V.ANA evolves according to a non-linear law as a function of the signal TODAC, for example a geometric progression of constant ratio.

 According to a particular non-limiting embodiment, the resistors R Il 0 to R 118 can respectively take the following values: 18000 ohms, 3430 ohms, 245û ohms, 1750 ohms, 1250 ohms, 892 ohms, 638 ohms, 455 ohms and 1140 ohms.

 In addition, according to the present invention, to allow input signals of positive or negative amplitude to be processed, the network 150 is preferably adapted to allow the direction of evolution of the analog signal V to be chosen accordingly. ANA, c that is, to authorize either a growth or a decrease of the analog reference signal V.ANA when the digital input signal TODAC increases.

The direction of evolution of the analog reference signal
V.ANA is chosen by the aforementioned CRN signal applied to one of the ends of the resistive bridge R110 to R118, and which evolves between two high and low logic states. The high logic level preferably corresponds to a potential of 5.5 volts1 while the low logic level of the CRN signal corresponds to a ground potential. In addition, the potential VRF applied to the second end of the resistive bridge R 111 to R 118 is defined by a second resistive bridge R 150, R 151, R 152, two transis Sized switches S 153, S 154, an inverter 155 and a follower amplifier 156.

 The resistors R 150 to R 152 are connected in series between a positive supply potential VDD (+ 5 volts preferably) and the ground. The non-inverting input of the amplifier 156 is connected to the common point of the resistors R 150 ′ and R 151 by means of the controlled switch S 153. The latter receives on its control input the signal CRN supplemented by l '' inverter 155.

 The non-inverting input of the amplifier 156 is also connected to the common point of the resistors R 151, R 152 via the transistorized switch S 154. This is controlled directly by the signal CRN. The output of amplifier 156 is looped back to its inverting input to form a follower stage. The signal VRF applied to the second end of the resistive divider bridge R llG to R 118 is available at the output of amplifier 156.

By way of nonlimiting example, the resistors R 15O,
R 151 and R 152 can have the following values respectively 2000 ohms, 1505 ohms and 2005 ohms.

So, taking the above values VDD = 15 volts and
CRN evolving between +5 and 0 volts, the resistive bridge R 110 to R 118 is supplied with * 3 volts (CRN = +5 and VRF = +2) when CRN is at high logic level, and conversely is supplied with -3 volts (CRN = 0 and
VRF = 3) when the CRN signal is at low logic level.

 The register means 300 have the function of defining an Ao word of X bits. By way of example X equals 3. In this case, the register means can be formed from 3 1-bit registers.

 The register means 300 are controlled by the clock signal D4. They can be reset by a RESET signal. The content of the register means 30G can be incremented or decremented by the adding / subtracting means 4û0. However, the variation of a command input, referenced NOC in FIG. 1, connected to the logic means 555 prohibits any modification of the content of the register means 355. According to an arbitrary convention, it will be assumed subsequently that any modification of the content of the register means 300 is prohibited when the NOC signal is at the high logic level.

 The gate circuit 200 is interposed between the output 3û2 of the register means 300 and the input 101 of the network 100.

 The door circuit 200 is designed to selectively apply the command word Ao contained in the register means 300 to the network 100. For this, the door circuit 200 is controlled by a FORSAX signal generated by the logic means 500.

According to an arbitrary convention, when the FORSAX signal is at the low logic level, the gate circuit 200 blocks the signal Ao Issu from the register means 300 and applies to the network 100 a control word Al corresponding to the low logic level. Consequently, the network 100 applies a minimum reference threshold to the comparator 10
B.

On the other hand, when the FORSAX signal is at the high logic level, the gate circuit 200 applies the control word Ao coming from the register means 300 on the network 110. Consequently, the network 100
Ao applies a reference threshold B x C to comparator 10
The gate circuit 200 can be simply formed of X gates with two inputs which each receive the FORSAX signal on a first of their inputs and which respectively receive on their second input one of the bits from the register means 300.

 The adding / subtracting means 400 have their input 0l and their output 402 (each with X bits) respectively connected to the output 3Q2 and to the input 301 of the register means 300. They are controlled by a POM signal generated by the logic means 500.

 According to an arbitrary convention, when the POM signal is at a low logic level, the means 400 increment the content of the register means 300 by D units, for example one unit. On the other hand, when the POM signal is at the high logic level, the means 400 decrement the content of the register means 300 by E units. Very advantageously, E corresponds to two units. Thus, if C equals 1.4, the decrementing of the content of the register means 300 by two units amounts to dividing the reference threshold by 2.

 FIG. 6 shows an example of a non-limiting embodiment of the adding / subtracting means 400.

 According to the embodiment represented in FIG. 6, the means 400 include 9 NAND doors with two inputs 401 to 409, three NOR doors 410 to 412, three inverters 413 to 415 and three doors 416 417, 418 which combine two AND functions. two inputs whose outputs are combined according to the NAND logic function.

 The inputs 401 of the means 4G5 are formed of a three-wire bus referenced respectively threshold û, threshold 1 and threshold 2 in the direction of the increasing weights of the bits originating from the register means 300.

 The outputs 402 of the means 455 are also formed of a three-wire bus referenced respectively DO, D1 and D2 in the direction of increasing weight of the bits.

 The embodiment shown in Figure 6 corresponds to a particular embodiment without limitation. Other configurations are possible. For this reason, the particular embodiment of FIG. 6 will not be described in more detail below.

 It will be noted that the means 4G5 are adapted to define an upper stop equal to 7, if this state is already reached and that the POM signal nevertheless commands an incrementation, and conversely are adapted to define a lower stop equal to O, if this state is already reached but that despite everything the POM signal commands a decrementation.

 FIG. 5 shows an exemplary embodiment of the logic means 500. These include two inverters 51ru, 512, a NOR gate with two inputs 513, a NAND gate 514 with two inputs and an exclusive OR gate 515 with two inputs .

 The signal KN - 2 from register 655 is applied to the input of the inverter 51t.

 The inputs of gate 513 respectively receive the signal KN-1 and the signal from the inverter 510.

 The FORSAX signal is available at the exit of gate 513.

 Gate 514 receives the FORSAX signal and a D3 signal which corresponds to the inverse of KN. Its output is connected to the input of the inverter 512 ..

 The POM signal is available at the output of this Inverter 512.

 Gate 515 receives the signals D3 and FORSAX as input.

The NOC signal is available at the output of gate 515.

 FIG. 2 gives the truth table of the FORSAX signal as a function of the signals KN - 1, KN - 2, contained respectively in the registers 6O0 and 650.

 It will be noted that the FORSAX signal is at the high logic level when KN - 1 is at the low level and KN - 2 is at the high level. In all other cases the FORSAX signal is at low logic level.

 Figure 3 gives the truth table of the NOC and POM signals according to the KN, KN-I, KN-2 signals.

 Note that the POM signal is at the high logic level when KN and KN-1 are at the low logic level while KN-2 is at the high logic level. In all other cases the POM signal is at low logic level.

The NOC signal is at the high logic level in 4 cases: either when KN, KN-1, KS-2 are at low logic level, or when KN and
KN-2 are at the low logic level while KN-1 is at the high logic level, i.e. when KN and KN-2 are at the high logic level while
KN-1 is at the basic logic level either when KN-1 and KN-2 are at the high logic level while KN is at the low logic level. In the other cases, the NOC signal is at low logic level.

 Figure 4 illustrates the main events occurring successively during each sampling period. In FIG. 4, 4 successive sampling periods N-2, N-1, N and N + 1 have been arbitrarily represented.

Sampling periods are defined by a signal
D4 whose period corresponds to the sampling period. The rising edge of the clock signal D4 coincides with the start of each sampling period.

 FIG. 4 also shows an HL signal.

This has the same period as the D4 signal, but a duty cycle
Less than 1. The rising edge of the HL signal occurs at three-quarters of each sampling period. The falling edge of the HL signal coincides with the rising edge of the next D4 signal.

 The main events occurring during each sampling period, as shown in Figure 4 for the sampling period N, are as follows.

 After a slight delay consecutive to the rising edge of the sampling signal D4, the FORSAX signal defined on the basis of the signals KN-1 and KN-2 contained in the registers 600 and 65û is applied to the circuit with doors 20ru. Thus, a TODAC signal is presented at the input of the network 100 and an appropriate analog reference threshold V.ANA is applied to the input 14 of the comparator 10. If the FORSAX signal is at logic level 1 the analog threshold applied to the comparator 10 is minimal and equal to B. If on the other hand the FORSAX signal is at the low logic level, the threshold applied to comparator 10 is defined by the word Ao contained in the register means 30ru, word Ao which itself depends on the POM and NOC signals Imposed at the end of the previous N-1 sampling period.

 Thereafter, the output of comparator 10 stabilizes before the rising edge of the signal HL appears.

 On the rising edge of this signal, the output level of comparator 10 is stored in memory. More precisely, according to the embodiment shown in the appended figures, on the rising edge of the signal HL, a signal D3 is stored which corresponds to the complement of the output of the comparator 16.

While the HL signal is high the signals
NOC and POM obtained at the output of doors 512 and 515 stabilize. Therefore, the adder / subtractor means 4û0 present at the input of the register means 300 a word Ao adapted to the next sampling period N
On the rising edge of the sampling signal D4 of the sampling period N + 1, the circuit ensures on the one hand the transfer of the content KN-1 from the state register 600 to the state register 650 and
KN in the status register 600, on the other hand, the evolution of the means 300 on the basis of the NOC and POM signals defined previously. Thereafter, but with a slight delay after the rising edge of the sampling signal D4 , a new FORSAX signal is presented on the gate circuit 200 and a new resulting analog threshold value V.ANA is presented to the comparator 10.

 The device output signal is formed using a logical combination of the KN - 1 and KN - 2 signals. Different solutions can be used for this.

According to a first solution, as illustrated in the figure
I, these signals KN - 1 and KN - 2 are applied to the inputs of a door
Exclusive OR 700 which attacks a flip-flop 702 clocked by the clock signal D4. The output signal is available on flip-flop 702.

 According to a second solution, as shown in FIG. 5, in dashed lines, the exclusive OR gate 700 can be replaced by a NAND gate 520 receiving as input the signals KN-1 and KN-2 inverted by inverters 510, 522.

 Such a NAND 520 generates at its output a ZZBAR signal identical to the signal from the above-mentioned gate 700 and therefore capable of being applied to the input of the flip-flop 702.

 FIG. 8 shows various signals explaining the operation of the device previously described and illustrated in FIGS. 1 to 7.

 On the top of Figure 8 we see an input signal shown in solid continuous line, the 8 reference thresholds likely to be generated by the network 110, represented in solid continuous lines (we will note in Figure S the evolution non-linear of the different thresholds according to a ratio of l, L), and the reference threshold actually generated by the ICO network, represented in broken dashed lines.

 The underlying lines in FIG. 8 respectively represent the internal H signals D4, HL, the output of the comparator and the signals D3, FORSAX, POM, NOC and output of the device obtained on the flip-flop 7ru2.

 During the sampling periods TO and T1 a spurious pulse appears on the input signal. This parasitic pulse remains below the V.ANA threshold. The comparator output remains at the low logic level. As a result, the D3 and NOC signals remain high while the FORSAX, POM and output signals remain low.

 The input signal crosses the threshold V. ANA during the sampling period T2. Consequently, the signals D3 and NOC go low on the rising edge of the corresponding HL signal.

In addition, the output signal goes high on the rising edge of the clock signal D4 initiating the next sampling period D3.

 Simultaneously, the signal NOC having passed to the low level, the threshold V. ANA is incremented. The signal remaining above the threshold level when the rising edge of the signal HL appears during the sampling period T3, the signals D3, FORSAX, POM, NOC and output remain identical at the end of the period T3.

 Thus, the threshold V. ANA applied to the comparator 1G is again increased on the rising edge of the signal D4 initiating the sampling period T4. During this the input signal is again greater than the threshold when the rising edge of the signal HL appears. The D3, FORSAX, POM, NOC and output signals remain; identical to their previous state at the end of period T4. The threshold V.NA applied to the comparator 10 and therefore again incremented on the appearance of the rising edge of the sampling signal D4 initiating the period T5.

When the rising edge of the HL signal appears during this period T5, the input signal is, on the other hand, below the threshold.
V. ANA. Thus, on the rising edge of the signal HL, the signals D3 and NOC return to the high level. As a result, the FORSAX signal also goes high after the rising edge of the signal appears.
D4 which initiates the period T6.

Therefore, during this period T6 the gate circuit 200 blocks the signal from the register means 300. A minimum threshold is therefore applied to the comparator 10. When the rising edge of the signal HL appears during this period T6, the input signal is greater than the minimum threshold applied. The signal D3 therefore descends to the low level on the rising edge of the signal HL. Likewise, the FORSAX signal goes down to the low level on the rising edge of the signal.
D4 which initiates the following period T7. The signal NOC having remained at the high level at the end of the periods T5 and T6, during the period T7 the device applies to the input 14 of the comparator IG a threshold identical to that used during the period T5.

On the rising edge of the signal HL of period T7 the input signal is again below the threshold. As a result, the signal D3 returns to the high level on the rising edge of the signal HL. Similarly, the FORSAX signal returns to the high level on the rising edge of the signal D4 which initiates the period T8;
On the rising edge of the signal HL during the period T8, the input signal is detected below the threshold for the second consecutive time. For this reason, the POM signal goes from low level to high level while the NOC signal goes from high level to low level on the rising edge of the HL signal. The FORSAX signal and the output signal descend to the low level on the rising edge of the signal D4 which initiates the period T9. At the start of this period T9, the POM signal having gone high, while the NOC signal has gone to the level low, the content of the register means 300 is decremented.

 A new threshold is then defined which preferably corresponds to half of the maximum threshold retained during the processing, according to the arrangements described above.

The P051 signal drops back to the low level while the signal
NOC rises to the high level on the rising edge of the signal HL of the period T9 consecutive.

 The device is then ready to detect the appearance of a new pulse of the input signal exceeding the analog threshold V.ANA defined at the start of the period T9.

SECOND EMBODIMENT: FORMATION OF TWO SIGNALS OF ENTER
There is shown in Figure 9, in the form of schematic functional blocks, a second embodiment of the device according to the present invention adapted to simultaneously process two different input signals referenced respectively ENR and ENV.

 It suffices to multiplex the input signals in order to apply them asternatlvement on the comparator 10.

As shown in Figure 9, the input signals
ENR and ENV are applied to respective amplifiers 20, 30, the outputs of which are connected to the same input 12 of the comparator 10 by gates 23, 33, controlled alternately in the state passing by multiplexing signals of opposite phases MUXI and MUX2.

 The 3G amplifier receiving the ENV input signal is mounted on the follower stage. For this, the signal ENV is applied to the non-inverting input of the amplifier 30, while the output of the amplifier 35, directed towards the controlled gate 33 is also looped back to its inverting input.

 The amplifier 20 which receives the input signal ENR is arranged in a controlled gain stage. For this, the signal ENR is applied to the non-inverting input of the amplifier 20. The output of the same amplifier 20 is looped back to its inverting input by a resistor R 21. The inverting input of the amplifier 20 is connected to the mounting mass by a resistor R22. In a manner known per se, the gain of the amplifier 20 is thus equal to 1 + (R21 / R22).

 The structure of the means S defining the reference threshold applied to the input 14 of the comparator 10 remains essentially identical to the arrangements previously described with reference to FIGS. 1 to 8.

 In fact, in FIG. 9, there is a network 100, a gate circuit 255, register means 300, add-on / subtractor means 400 and logic means 500.

 It is however necessary to provide a pair of status registers 600, 650, and X data registers 300 respectively for each of the input signals ENR and ENV. In addition, it is necessary to multiplex these registers in coincidence with the signals MUXI and .N X2.

The signals MUXI and MUX2 of opposite phases are respectively formed by an address signal D5 and its complement, with a duty cycle equal to I and whose period is twice the period of the sampling signal D4. The signal D5 and its complement are also formed by the means shown schematically in Figure 14 from the internal signal H.

 In FIG. 9, the status registers assigned to the signal ENR are referenced 600R and 650R. Likewise, the status registers assigned to the input signal ENV are referenced 600V and 650V.

 The X data registers allocated to the signal ENR are referenced 300R and the X data registers allocated to the input signal ENV are referenced 300V.

 The output signal of the comparator 10 is applied alternately to the state registers 60cR and 600V by a multiplexing assembly g '' controlled by the signal D5. Similarly, the signals obtained at the output of the state registers 600R and 650R or 600V and 650V are applied to the logic means 500 by means of a multiplexing assembly 810 controlled by the signal D5.

 In a comparable manner, the signal NOC and the output of the adder / subtractor means 400 are applied alternately to the register means 300R and 300V respectively by means of a set this multiplexing 820 controlled by the address signal D5, and the outputs of the register means 300R and 300V respectively are applied to the gate circuit 200 by means of a multiplexing assembly 830 also controlled by the address signal D5.

 According to an arbitrary convention, when the signal D5 is at the high level. door 23 is busy. The ENR input signal is processed.

Therefore. The assembly 800 connects the output 16 of the comparator to the register 65rR. The assembly 810 connects the outputs of the registers 6OOR and 650R to the logic means 500. The assembly 820 applies the signal NOC and the output of the additive / subtractor means 400 to the register means 355R. The assembly 830 connects the output of the register means 300R to the gate circuit 200.

Conversely when the signal D5 is at low logic level, the gate 33 is on. The ENV signal is processed. Therefore,
The assembly 855 connects the output 16 of the comparator 10 to the register 6G5V.

The assembly 815 connects the outputs of the 600V and 650V registers to the logic means 500. The assembly 820 applies the NOC signal and the output of the add-on / subtractor means 400 to the register means 300V and the assembly 830 connects the output of the means to 300V register at door circuit 255.

 FIG. 10 shows a particular example of embodiment of the means S illustrated diagrammatically in FIG. 9.

 We see in Figure 10 a first two-line input bus which carries signals BERBUS 3 and BERBUS 4 from a module shown schematically in Figure 12 and intended to reset respectively the register means 300R or the means with register 300V if the input signal ENR or ENV respectively has not crossed the associated threshold during a determined time delay.

 We also see in FIG. 10 another 9-line bus which carries signals DO to D7 respectively, as well as the complement of signal D5.

 The signals Dc, Dl and D2 come from the adder / subtractor means 400 as indicated previously with reference to FIG. 6.

 The signals D3, D4, D5 and its complement have been previously described. The signal D6 corresponds to the NOC signal coming from the gate 515. The signal D7 corresponds to a general reset signal of the device.

 The gate circuit 200 is formed by 3 NOR gates with two inputs 2pu2, 2G4, 206. The gates 252, 204 and 206 receive on one of their inputs the signals From the assembly 830. They receive on their second input the FORSAX signal. The gates 202, 204 and 206 deliver at their output the signals TODAC O, TODAC 1 and TODAC 2 directed to the network 100. The register means 300R are formed by three flip-flops D, 302R, 304R and 306R. Similarly, the register means 300V are formed by three flip-flops D, 302V, 304V and 306V. The assembly 820 is formed by two NOR doors 822, 824 associated respectively with the register means 300R and with the register means 300V. The gate 822 receives on its inputs the signal NOC (D6) and the complement of the signal D5.

The gate 824 associated with the register means 300V receives on its inputs the signal NOC (D6) and the signal D5.

 The flip-flops 3G2R, 354R and 356R are associated respectively with gates 353R, 355R and 3G7R. These gates have the function of applying to the input of flip-flops 3G2R, 3G4R and 306K either the data present at the output of the same flip-flop, or the value available respectively on the line DO, DI or D2 of the bus.

 Each of the gates 353R, 305R and 307R combines two AND functions with two inputs, the outputs of which are combined according to the NAND logic function. A first AND function of each of these gates receives the output of the multiplexing gate 822 and the DO signals respectively. Dl and D2. The second AND function receives the complement of the output of the multiplexing gate 822 obtained by an inverter 823 and the complemented output of the flip-flops 302R, 304R and 356R respectively.

 The output of gates 303R, 305K and 3G7R is looped back to the input of the respective flip-flops.

 The flip-flops 352R, 354R and 306R are clocked by the clock signal D4. The reset input of flip-flops 302R, 3C4R and 306R is connected to the output of an ET 826 gate which receives signals D7 and BERBUS 3 on its inputs. Thus, flip-flops 302R, 304R, 306K can be reset to zero either when a general reset of the device is required (validation of signal D7), or when the input signal ENR has not crossed the associated threshold for a predetermined period (validation of signal BER BUS 3).

 The outputs of flip-flops 302R, 304R, 306R, correspond to the lines CUBUSO, CUBUS I and CUBUS2.

 Similarly, the flip-flops 302V, 304V, 360V, are associated respectively with gates 303V, 355V, 307V. The doors 303V, 305V, 307V have the function of applying to the input of flip-flops 302V, 3G4V, 306V, either the data available at the output of these flip-flops respectively, or the data available on the lines DO, D1 and D2 respectively. Each of the gates 303V, 355V, 307V, combines two AND functions with two inputs whose outputs are combined according to the NAND logic function.

 A first AND function receives the signal from the multiplexing gate 824 and the signals DO, D1 and D2 respectively. A second AND function receives the complement of the output of the multiplexing gate 824, obtained thanks to an inverter 825 and respectively the complemented output of the flip-flops 302V, 304V and 306V.

The output of doors 303V, 305V, 307V is looped back to the input of flip-flops 352V, 304V and 306V. These flip-flops are clocked by the signal D4. Their reset input is connected to the output of an AND 827 gate which receives the D7 signal and the signal as an input.
BERBUS4. Thus, the flip-flops 3G2V, 304V, 306V are reset to zero either when a general reset of the device is required (validation of the signal D7), or when the signal ENV has not crossed the associated threshold for a predetermined period ( validation of signal BERBUS4).

 The outputs of the 3û2V, 304V and 306V flip-flops are available on the CUBUS5, CUBUS6 and CUBUS 7 lines respectively.

 The multiplexing assembly 830 includes 3 doors 832, 834 and 836. Each of these doors combines two AND functions with two inputs, the outputs of which are combined according to the NAND logic function.

 A first AND function receives the signal D5 and respectively the signals CUBUSû, CUBUS 1 and CUBUS 2. The second AND function of each gate 832, 834 and 836 receives the complement of the signal D5 and respectively the signals CUBUS5, CUBUS6 and CUBUS7.

The output of doors 832, 834, 836 is connected to the input of the circuit with doors 200 (doors NOR 202, 204 and 206).

 It will be noted that the output of the gates 832, 834 and 836 is connected to the input of the adder / subtractor means 400 by means of inverters 833, 835 and 837 at the output of which the signals SEUILO, SEUIL1 and SEUIL2 are available respectively.

 The assembly 800 includes 4 doors 802, 804, 806 and 808 associated respectively with 4 flip-flops D forming the registers 600R, 650R, 600V and 655V.

Each of doors 8G2 to 808 combines two AND functions with two inputs whose outputs are combined according to the logic function
NAND.

 A first AND function of the gate 8O2 associated with the flip-flop 600R receives the signals D3 and D5. The second AND function of the gate 802 receives the complement of the output of the flip-flop 600R and the complement of D5. The output of gate 802 is looped back to the input of flip-flop 600R.

 A first AND function of gate 804 associated with flip-flop 655 receives D5 and the complement of the output of flip-flop 600K.

The second AND function of gate 854 receives the complement of the output of the flip-flop 655R and the complement of D5. The output of door 854 is connected to the input of flip-flop 650R.

 A first AND function of gate 806 associated with flip-flop 65.V receives the signal D3 and the complement of D5. The second AND function of gate 806 receives the signal D5 and the complement of the output of the flip-flop 600V. The output of gate 806 is connected to the input of the flip-flop 655V.

 A first AND function of the gate 808 associated with the flip-flop 650V receives the complement of D5 and the complement of the output of the flip-flop 600V. The second AND function of gate 808 receives the complement of the output of the flip-flop 650V and the signal D5. The output of gate 858 is connected to the input of the 650V scale.

 The flip-flops 600R, 650R, 600V and 650V are clocked by the signal D4. The reset input for flip-flops 600R, 650R, 600V, 65OV is connected to line D7.

The outputs of flip-flops 605R, 650R, 600V, 650V, are connected respectively to lines CUBUS4, CUBUS3, CUBUS9 and
CUBUS8.

 The assembly 815 is formed by two gates 812, 814 which combine two AND functions with two inputs whose outputs are complemented according to the NAND logic function.

 A first AND function of gate 812 receives the signals D5 and CUBUS 3. The second AND function of gate 812 receives the signal CUBUS 8 and the complement of D5.

A first AND function of gate 814 receives the signal
CUBUS4 and the D5 signal. The second AND function of gate 814 receives the signal CUBUS9 and the complement of D5.

 The KN-I and KN-2 complements are available respectively at the output of doors 814 and 812. The outputs of doors 814 and 812 are therefore connected to the logic means 500 described above with reference to FIG. 5.

 There is shown in Figure 11 attached various signals obtained on the circuit shown in Figure 10 and described above.

 We see at the top of Figure 11 in solid lines, an ENR input signal (we assume according to the representation in Figure 11 that the signal ENV remains zero) and in broken dashed lines, the reference threshold actually generated by the network 110.

We also see in figure II the signals H. INTERNAL, D4, D5, HL, D3, at the output of the comparator, FORSAX, POM,
NOC and exit R.

 The sampling periods allocated respectively to the input ENR and to the input ENV are referenced R and V respectively in FIG. 11. It will be noted that the periods R correspond to the high level of the signal D5 and conversely the periods V correspond to the level bottom of signal D5.

 As indicated above, with reference to FIG. 7, the device according to the present invention is preferably designed to allow processing of signals having useful pulses, either positive or negative.

 For this, the configuration of the network 155 is determined by the level of the signal CRN. In addition, as shown diagrammatically in FIG. 14, the output of the comparator 10 attacks, after passing through a 4G inverter, an input of a door 42 ensuring the completed logic or exclusive function. Gate 42 receives on its second input the aforementioned coding signal CKN. The output of gate 42 is connected to the input of a flip-flop D44. The flip-flop D4 is clocked by the signal HL. Its reset is ensured by signal D7. The signal D3 is available at the output of flip-flop 44.

 FIG. 14 also shows an exemplary embodiment of means making it possible to generate the signals D4, D5 and its complement and D7 from the internal clock signal. The means represented for this purpose in FIG. 14 are susceptible of numerous variants of réallsatron and will therefore not be described in more detail below.

 FIG. 9 shows an OR gate 755R receiving the signals KN-1 and KN-2 contained in the flip-flops 6ccK as an input. 655R, and the output of which is connected to a flip-flop 702R clocked by the signal D4. The flip-flop 752R therefore generates at its output a reshaped signal of the same frequency as the signal ENR.

 Similarly, FIG. 9 shows a 755V OR gate which receives KN-1 and KN-2 signals contained in the 655V and 650V flip-flops at the input, and the output of which is connected to a 752V flip-flop clocked by the signal. D4. The flip-flop 702V generates at its output a reshaped signal of the same frequency as the signal ENV.

 FIG. 13 shows another alternative embodiment of means making it possible to generate the reshaped signals of the same frequency as the input signals ENR and ENV.

 According to the representation of FIG. 13, these means comprise two doors 710, 714 and two flip-flops D712, D716, respectively associated.

 Each of the doors 710, 714 combines two AND functions with two inputs, the outputs of which are combined according to the NAND logic function.

 A first AND function of gate 710 receives the complement of D5 and the output signal completed by the flip-flop 712. The second AND function of gate 710 receives the signal D5 and the complement of signal Z, BAR. This complement is obtained at the output of an inverter 718. It will be recalled that the signal ZZBAR is available at the output of a NAND gate 525 of the logic means 500, as shown in FIG. 5.

 The output of door 710 is connected to the input of flip-flop 712.

 Similarly, a first AND function of gate 714 receives the signal D5 and the signal obtained on the complemented output of flip-flop 716. The second AND function of gate 714 receives the complement of DS- and the complement of ZZBAR. The output of door 714 is connected to the input of flip-flop 716.

 The flip-flops 712 and 716 are clocked by the signal D4 and reset to zero by the signal D7.

 The outputs R and V are available at the output of flip-flops 712 and 716 respectively.

 Furthermore, the device according to the present invention is preferably provided with means making it possible to reset to zero the threshold applied on the comparator 10 for processing a given input signal, if this input signal does not cross the threshold imposed by the network 100 for a predetermined period.

 Such means prevent the device from being blocked with a very high threshold greater than the useful input signal after the appearance of a parasitic pulse of large amplitude.

 Many arrangements can be made for this purpose.

 There is shown diagrammatically in FIG. 12 means fulfilling this function.

 Essentially, the means represented in FIG. 12 are designed to initialize a counter when a change in level occurs at the output and to reset the corresponding threshold to zero if a new level development of the same signal is not not intervened before the counter reaches a predetermined number.

 More specifically, the means shown in Figure 12 are designed to alternately monitor the output R and the output V.

 4 for this purpose, the means shown in FIG. 12 include a selector 955, a timer 91ru, an inverter 920 and two doors N 'D 935 and 940.

 The selector 955 receives on its inputs 952 and 954 the signals output R and output V respectively.

 It delivers on its output 9G6 a logic signal which switches between a high level and a low level depending on whether the output R or V is monitored. The switching of the signal at the output of the selector 950 are synchronized with the rising edges of the output R or V output signals as appropriate.

 More specifically, the operation of the selector is as follows.

 It will be assumed that the signal on output 906 is at the high level when the signal output R is monitored and conversely the signal on output 906 is at low level when the signal output V and is monitored.

 When the signal on output 906 is high, the selector monitors the appearance of a rising edge on the signal output R.

The signal on output 956 returns to low level when such a rising edge appears. On the other hand, changes in the output signal V have no effect on the selector 900 as long as its output 906 is at the high level.

 Conversely, when the signal on output 906 is at the low level, the selector 900 monitors the changes in the output signal V and returns to the high level when a rising edge appears on the output signal V.

 On the other hand, changes in the output signal R have no effect on the selector 900 as long as the signal on the output 906 is at low level.

 Each edge of the signal on output 906, whether rising or falling, ensures the reset of timer 910.

 This timer 910 is preferably formed by a counter receiving on its counting input a clock signal of fixed frequency. The output 912 of timer 910 is validated when a predetermined account is reached.

The validation of the output 912 of the timer 910 itself ensures the resetting of this timer 910 and the switching of the selector 900. The validation of the output 912 of the timer 91G indeed indicates that the monitored input signal, R or
V did not go above the associated threshold during the predetermined period.

 The output 912 of the timer 91û is also connected to one of the inputs of each of the doors 930 and 940. The gate 930 receives on its second input the signal available on the output 906. The gate 940 receives on its second input the complement of the output 906 obtained at the output of an inverter 920. The gates 930 and 940 generate on their output signals BER BUS 3 and BERBUS 4 respectively, which as previously indicated are applied on AND gates 906 and 927 to ensure the resetting of data flip-flops 302R, 304R, 306R and 302 V, 304V, 306 V respectively.

 Of course the present invention is not limited to the particular embodiments which have just been described but extends to all variants in accordance with its spirit.

 In the foregoing description, it has been indicated to simplify the description that the network 100 delivers at its output 1G2, an analog reiterative threshold V.ANA equal to B X CA. In reallocation, the network 15G represented in FIG. 7 delivers at its output 102, an analog reference threshold V.ANA equal to VRF (B X CA). However, it is preferable to use, at the input of the shaping device, a high-pass filter which references the signal with respect to VRF.

Claims (17)

R E V E N D I C A T I O N S  1. Device for shaping a frequency analog signal, characterized in that it comprises - means (S) defining a variable reference threshold (V.ANA), including  the changes controlled by the level of the input signal follow a  non-linear sequence, and - a comparator (10) which receives on a first input (12), a signal  input to be shaped and which receives on a second input (14) the  reference threshold (V.ANA).  2. Device according to claim 1, characterized by the  fact that the means defining the reference threshold include  a network (100) capable of transforming a digital signal (TODAC)  that it receives as input in a reference analog signal (V.ANA)  available on its output, connected to the reference input (14) of the  comparator (10),  digital register means (300) which contain a number of  X bits representing the reference threshold,  a gate circuit (20G) placed between the network input (100) and the  output of the register means (300), for selectively applying the  signal available in the register means (300) on the input of the  network (100),  adding / subtracting means (400) capable of incrementing /  decrement the content of the register means (30), and  logic means (500) sensitive to the level detected at the output of the  comparator (10) which control the register means (300), the  door circuit (200) and the adding / subtracting means  (400).  3. Device according to claim 2, characterized in that it comprises sampling means (600, 654, D4) which sample the signal available at the output of the comparator, and by the fat that the logic means (555) are sensitive to the level detected at the output of the comparator (15) during several consecutive sampling periods.  4. Device according to one of claims 2 or 3, characterized in that the logic means (500) control the door circuit (205) using a FORSAX signal, that the door circuit (200) applies the signal available in the register means (300) to the input (IcI) of the network (155) when the FORSAX signal is at a first level (preferably low) while the gate circuit (20O) applies to the 'input (1SI) of the network (15û) a signal imposing a minimum threshold, when the FORSAX signal is at a second level (preferably high), by the fact that the logic means (505) generate a FORSAX signal at the second level when the KN-I signal, corresponding to the comparator output (ire) during the last completed sampling period, is at low level. while the signal KN-2, corresponding to the output of the comparator (10) during the penultimate completed sampling period is at the high level. by the fact that the logic means (50ru) control the register means (30O) using an NOC signal, that the NOC signal is determined as a function of the level of the comparator (10) during three periods of consecutive sampling (N, N-1, N-2), that the NOC signal imposes the state of the register means (300) for the next sampling period (N + 1), and that the NOC signal authorizes the evolution of the content of the register means (300) if it is at a first level (preferably low) while it prohibits the evolution of the content of the register means (300) if it is at a second level (high preferably), by the fact that the logic means (500) generate a second level NOC signal if -a) the signals KN, KN-1 and KN-2 sampled at the output of the  comparator are at low level, or -b) signals KN, KN-2 are at low level while (KN-1) is at  high level, or -c) the signals KN, KN-2 are at the high level while the signal (Kn-l)  is at the low level, or -d) the signals KN-1, KN-2 are at the high level while the signal (KN)  is at a low level, by the fact that the logic means (5 or) control the add / subtractor means (400) using a POM signal, that the signal POM is determined as a function of the level of the comparator (10) during three consecutive sampling periods (N, Nl and N-2), that the POM signal imposes the state of the adding / subtracting means (400) for the period next (N + 1), and that the POM signal controls the increment of the register means (3ru3) if it is in a first level (preferably low) and that it controls on the other hand the decrementation of the register means ( 30G) if it is in a second level (preferably high), and by the fact that the logic means (500) generate a second level signal (POM) if the signals KN and KN-1 obtained at the output of the comparator (10) are at the low level while the KN-2 signal is at the high level.  5. Device according to one of claims 2 to 4, characterized in that the network (105) is a resistive network.  6. Device according to one of claims 2 to 5, characterized in that the network (100) generates a reference threshold having a geometric progression.  7. Device according to claim 6, characterized in that the network (100) generates a reference threshold having a geometric progression of 1.4.  8. Device according to one of claims 2 to 7, characterized in that the adding / subtracting means generate increment steps of a unit for the register means (300). and generate two unit decrement steps for the register means (300).  9. Device according to one of claims 2 to 8 characterized by the fact that the network (ion) comprises a resistive bridge (R Il 0 to R 118) whose intermediate points are connected to the output (102) by a network d 'switches (S120 to SI 32) controlled by the digital input signal (TODAC) -out of the door circuit (255).  10. Device according to claim 9, characterized in that the resistive bridge (R 110 to R 118) is connected between two supply points (CRN, VRF) one of which corresponds to a coding signal (CRN) evolving between two levels, high, low, depending on whether the input signal has positive or negative pulses.  11. Device according to claim 10, characterized in that the second supply point of the resistive bridge changes between two values according to the level of the coding signal (CKN).  12. Device according to one of claims 10 or 11, characterized in that the second supply point (VRF) of the resistive bridge (R111 to R118) is generated by a follower stage (156) which receives as input either of two potentials defined by a resistive bridge (R! 5r ,, R 151, Rl52), via switches (5153, S 154) controlled by the coding signal (CRN) and its complement .  13. Device according to one of claims 10 to 12, characterized in that the comparator output signal (10) is applied to the input of an exclusive OR gate (22) whose second input receives the value coding signal logic (CRN).  14. Device according to one of claims 3 to 13, characterized in that the output signal is formed by logical combination of the signals KN-1 and KN-2 obtained at the output of the comparator during the last and the front -last full sampling period.  15. Device according to one of claims 1 to 14, characterized in that it comprises switches (23, 33) controlled by multiplexing to alternately apply different signals to the first input (12) of the comparator (10).  16. Device according to claim 15, characterized in that it comprises a pair of registers (6for, 650R; 650V, 650V) associated with each input signal (ENR; ENV) to store the states (KN-I and KN-2) at the output of the comparator (10) during the two previous sampling periods associated respectively with each input signal (ENR, ENV), by the fact that a register (600R, 630V) of each pair is connected at the output of the comparator (1G) by a multiplexing assembly (800) which routes the output signal of the comparator (1O) to the appropriate register (600R, 655V) according to the processed input signal (ENR) or (ENV) , by the fact that the registers (600R, 650R 600V, 655V) are connected to the logic means (500) by a demultiplexing assembly (810) which applies the content of one or the other pair of registers to the logic means ( 500), depending on the input signal processed (ERN) or (ENV>, by the fact that it includes a set of X registers (300R, 30 5V), respectively associated with each input signal ENR, ENV to store each a number of X bits associated with each input signal (ENR, ENV), by the fact that the sets of X registers (300R, 300V) are connected to the output of the adding / subtracting means (400) by a multiplexing assembly (82G) which directs the output of the additive / subtracting means (400) to one or the other of the assemblies according to the input signal processed ( ENR) or (ENV), by the fact that the sets of X registers (300R, 300V), of each set are connected to the gate circuit (zoo) via a demultiplexing set (830) which applies the content of one or other of the sets of registers to the door circuit (200) according to the processed input signal (ENR) or (ENV), and by the fact that the various multiplexing or demuliplexing assemblies (800, 810, 820, 830), are controlled by the same signal address (D5) which switches between two states according to the processed input signal (ENR) or (ENV).  17. Device according to one of claims 1 to 16, characterized in that it further comprises means (910, 910, 920, 930, 940) adapted to ensure a reset of the reference threshold if a signal has not crossed the associated threshold for a specified period.