JPS5970018A - Josephson monostable multivibrator - Google Patents

Abstract

PURPOSE:To obtain variable pulse width having quick rise by setting up a difference of time constants of control lines connected to a Josephson switching gate branched to output lines outputting input signal current and logical information. CONSTITUTION:A Josephson single junction element J1 is connected in parallel with an input terminal 3 as a buffer amplifier 4 of an input part 2 to supply circuit current Ig. The current Ig is selected at the size selected by the sum of input signal current Iin and the current Ig and the input signal current Iin and logical information are obtained from an output 5. The output 5 is connected to an output terminal 7 through an output line 6. A control line 8 is connected to the line 6 as a branch and the other end of the line 8 is connected to a control input 9a of a switching gate 9 selectively leading in output current from the line. A skid Q is used for the gate 9 and its negative terminal is connected to the terminal 7 through a resistor R3. The time constant of the line 6 is set up sufficiently smaller than that of the line 8, so that variable pulse width having quick rise can be outputted.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a Josephson monostable multivibrator.

A monostable multivibrator is an essential circuit when attempting to implement a subsystem with somewhat advanced processing functions, such as a memory peripheral circuit, using Josephson logic circuits.

For example, this is used when switching the current in a superconducting inductor loop using a skid (5QUID: superconducting quantum interference device), or when controlling logical operations in a latching operation. Also, Josephson is used to measure ultrafast phenomena.
A sampler also requires a one-shot pulse with a sufficiently narrow pulse width.

In such a Josephson monostable circuit, the input signal logic is embodied in the presence or absence of current, so its function is from the time when the input signal current Ijs rises to the expected pulse width, as shown in Figure 1. It can be defined as generating an output current rout.

The practical requirements associated with these functions are that the output current has a good start-up speed, that the pulse width is easy to design and has a degree of freedom, and that it is simple and easy to create as a circuit device. It is easy to use, and the circuit area is as small as possible.
etc.

However, the previously proposed Josephine monostable multivibrator could not satisfy the above requirements. In order to clarify this point, examples of two conventional types of circuits are shown and explained in FIGS. 2A and 2B, respectively.

First, in the case shown in Figure 2A, when the input signal current in is zero, the bias current In reduces the maximum Josephson current of the skids in the left branch (branch circuit), so the power supply current G flows exclusively through the skid Qs in the center line, which is in a zero voltage state.

When the input signal current 14n rises, the skid Q switches to the voltage state and the power supply current g begins to flow to the right output branch as an output current 1tnt,t. If various parameters are determined appropriately, if a Josephson connection is made during the transition period of the rise of the output current, the desired one-shot pulse can be obtained without the rise of the output current Iout.

After that, when the input signal current Iin falls, the skid Q7 switches back to the voltage state and the skid Q switches back to the zero voltage state. Constant τo=
In addition to the degree of freedom in selecting Lo/R and, ultimately, the degree of freedom in output pulse width TD being reduced, a sharp rise of the output pulse Iout cannot be expected.

On the other hand, without using the resistor R in the output branch, the second
As shown in Figure B, there is also a conventional example in which a one-shot multi is constructed using only a skid and a superconducting inductor tube.

In this conventional example, skids Q and -Q have two control lines.
, is used. However, two of them Q! Electric power uses only one control line.

In this figure, the arrows shown in the frames indicating skids Q1 to Q4 are the direction of the circuit current or power supply current of the relevant skid, and when only the control current in the direction that matches this flows, each skid transitions to the voltage state. However, due to the inductance Lo in the branch on the output side, the applied DC current σ is equal to both control lines Cq* at this point.
, Cq both flow into the left branch through the skid Q in which no magnetic field current flows, and are connected to the ground through the skid Q4 in which no magnetic field current flows in the control line Cq4, so the maximum Josephson current does not decrease. It flows out.

When an input signal current of 1 inch is applied, a current of control line C (h) of the skid Q1 flows, and the skid Q
1 transitions from a voltage state to a resistance state. Therefore, the power supply current I begins to flow to the output side branch as an output current rout through the skid Q2, where no magnetic field current is flowing through the control line Cqv yet at this point.

This output current becomes a control current flowing through the control line Cq of the skid Q4, so the skid Q4 is switched to a voltage state, and the output current or power supply current is cut off from the control line Cq, and the other control line Cq is turned off. Since a reverse current as an input signal current is applied to the line cq*', the inductance L is increased through the skids to which no magnetic field current is applied.
The current is commutated to the outer line containing ``e''.

When this commutation current becomes steady, the skid Q+ transitions to a zero voltage state due to the cancellation of both control line currents, while the skids transition to a voltage state due to the current in the control line Cqx.
The power supply current IQ is commutated from the output side branch to the left branch, and the output current 1out falls to complete the intended one-shot operation. Of course, when the input current l4tL falls, the circuit returns to its initial state.

In this method, the pulse width TD and pulse rise are determined by the inductances Le and Lo and the damping coefficient β of each built-in Josephson junction, but when the Nockles width is long, there is a degree of freedom compared to the previous example. However, the drawbacks are the large number of skids, complicated wiring patterns, and large circuit area, which are disadvantageous.

In addition, in both of the conventional circuits shown in Figures 2A and 2B, the subsequent circuit element in which part of the output line is an input line is a magnetic field coupling type circuit in which the resistance in the input line is zero, and is ultimately a skid. It also has the disadvantage that it can only drive . This is because the load condition greatly affects the one-shot operation, and the load is designed to be fixed.

Furthermore, as mentioned above, the degree of freedom in designing the pulse width TD is narrow and the rising edge cannot be expected, but with fixed circuit parameters once designed, it goes without saying that only an output pulse with a predetermined pulse width can be produced. It was not possible to make the pulse width variable.

The present invention has been made in view of the above, and it is an object of the present invention to provide a monostable multivibrator that eliminates the drawbacks of the conventional examples described above, namely, to provide a monostable multivibrator that is simple in structure, easy to manufacture, and has a small circuit area. The main purpose is to obtain an output pulse with a rise and fall of sb, and to have a large degree of freedom in changing the pulse width and amplitude. It is also intended to be able to drive the subsequent circuit of the mold, and to be able to perform variable pulse width operations without changing the constants of each element itself.

FIG. 5 shows various embodiments of the present invention. First, a comparatively basic embodiment of the monostable multivibrator shown in FIG. 5A will be explained.

First, there is an input section having an input terminal 3 for input signal current Iin. In this embodiment, in order to make the input signal current Iin, which is generally at a relatively low level, desirably at a stable and relatively large level, A buffer amplifier (hereinafter referred to as buffer amplifier) is also provided in the input section a.

Buffer amplifiers in Josephson circuit systems generally take the form of switching gates, and switching gates include current injection type, direct coupling type, and magnetic field coupling type such as skid.

Any of these may be used as the buffer amplifier in the input section of this circuit, but in the case shown in the figure, the most basic switching gate of the current injection type is a switching gate using a Josephson single junction element J1, which is convenient in construction and manufacture. is used.

That is, this Josephson single junction element J1 is arranged in parallel with the input terminal 3, and the circuit current IQ from the positive power supply Vcc is supplied to it, and the value of this circuit current IgO is expressed as
Although it does not cause the Josephson single junction element Jl to switch between the voltage state and the resistance state when it is used alone, the magnitude is selected such that the sum with the input signal current Iin causes the Josephson single junction element Jl to switch into the voltage state.

Since the Josephson single-junction element has a well-known hysteresis curve and latching mode, if the value of the circuit current Ig is determined as above, this single-junction element Jt will be turned into a voltage only when an input signal current of 1 inch is input. The level-amplified signal current I4 can be extracted from one end as an output from the input section by switching the state.

This amplified input signal current 2 differs from the actual input signal current I7% in terms of analog level and its stability or fluctuation rate; The logic information of the vibrator is the same as the input current signal Ii.

On the other hand, if the input signal current is as large as 1 inch, there is no need for the buffer amplifier 'Iff. The input signal current Iin and the logic information are exactly the same in analog terms.

In any case, it is sufficient that a signal current having at least the same logic information as the input signal current is obtained at the output of the input section.
Depending on the level of the signal currents I and i, it may be selected whether or not to provide a buffer amplifier in the input section.

Furthermore, even if the gate is of the same current injection type, if the input/output separation function, that is, the buffer function is important, and the degree of amplification is also desired to be large, it is preferable to use a known four-junction closed loop type switching gate.

The structure of this gate itself and the principle of switching operation are already well known, and are basically based on Japanese Patent Application Laid-Open No. 56-328.
Since No. 50, various improvements have been made, including Japanese Patent Application Laid-Open No. 57-99054, which are hereby referred to as reference.

Input section - output! is connected to one end 6α of the output line B, and the other end 6b of the output line is connected to the output terminal 7 of the circuit. Similarly, input section output! In , one end rα of the control line r is branchedly connected to the output line B, and the other end rb selectively draws the output current in the output line 6 and serves as a power source (therefore, in the case of , it is a negative power source) Switching gate that pulls into Vss? control input? Connected to α.

In this embodiment, the switching gater uses a skid Q which can easily be configured to operate with a negative power supply even with a positive polarity control current, and is incorporated into the circuit as follows.

The circuit current terminal of the skid Q2 is inserted in series between the ground and the lead-in power supply or the negative power supply (2), and the negative terminal and the output terminal 7 or the output line 6 are connected via the current addition resistor R8.

One feature of the present invention is that, in the above-mentioned static configuration, the time constant difference that causes a delay in each line current in the output line 6 and the control line r is actively utilized for operation.

In this embodiment, as shown in FIG. Assuming that the configuration satisfies the basic functions of
l/R.

The time constant τ2=Lx/Rt in the control line r is also selected to be sufficiently small, preferably almost zero.

In actual production, in this embodiment, the resistor R8 is
acts as a current summing resistor together with the resistor R1, so in order to reduce the time constant τi in the output line 6, it is unavoidable to avoid creating an inductance Ll intentionally. On the other hand, the inductance L is intentionally formed in the control line so that it has a large value. Hereinafter, R2) Ll=0・
Therefore, the operation will be explained assuming that the intentional time constant difference τ,-- is solely due to the value of τ.

FIG. 4A shows the correlation between the waveforms of each part, and the power supply V
After turning on cc, ■, as mentioned above, the input signal current ■
When the signal current I4, which has the same logic information as in, rises in the positive direction, this current is shunted to the output line 6 and the control line r, and the shunted portion to the output line t is transferred to the output terminal 7 without delay. As a result, a rising portion of the output current Iout is formed.

On the other hand, at this point, it's still skid Q.

is not in the zero voltage state, therefore, the output terminal 7 is shunted to the ground by the resistor R3. Therefore, if the resistance connected to the output terminal 7 is sufficiently smaller than R3, the shunting is performed.
can flow out from the output terminal 7 as an output current Iout.

Thus, the shunt branch 8 leading to the skid control type Kata via the control line rt has a transmission time constant difference τ, -τ1 . Therefore, in this case, the skid control operator is reached with a delay of a time To approximately determined by the time constant.

Time T. After a period of time, when the control current 8 is applied to the skid Q, the skid Q transitions to the voltage state. Then, the resistor R whose one end is connected to the output terminal
Since the other end of s is connected to the lead-in power supply or the negative power supply, the output line shunt branch 6 that was flowing out from the output terminal 7 as the output current 1out is drawn into the lead-in power supply , and the output current 1out becomes It will go down. Ultimately, the time TD defines the pulse width.

This explanation can be illustrated and explained that if we consider the reverse direction or negative polarity current indicated by the illustrated virtual line 17 from the negative power source (2), the result of adding both currents by both resistors + R8 is obtained as the output current Icyut. This is shown in Figure 4.

In this way, according to the present circuit I, a one-shot pulse of the desired pulse width TD can be obtained with a relatively simple configuration,
Although it is possible to obtain the effect that both the rise and fall values ν are good and the degree of freedom in selecting the pulse width and pulse amplitude is wide, the following design method can be cited as a more practical consideration.

From the viewpoint of stable operation of the circuit and design, it is desirable that both switching gates (JllQt) maintain a gap voltage, ie, a constant voltage, in the voltage state. for that purpose,
Choose a relatively large value for each resistor. Also, the output current hut
To maintain the zero level after being pulled in, the output terminal potential can be set to the ground potential by manipulating the resistance value, for example by setting Rs = Rs if other conditions are the same.
Of course it is possible.

If a magnetic field coupling type circuit is connected and driven after this circuit l, the inductance Ll on the output side will be slightly reduced.
However, as mentioned above, if the value of the resistor R1 is chosen to be large, the time constant t+=Ls/R+ can be kept sufficiently small, and therefore the rise of the output current Itmt can be kept sharp.

Further, if each parameter is determined so that the voltage at the output current terminal 7 is sufficiently smaller than the gap voltage, it is possible to drive a direct-coupling type or current injection type circuit having an input circuit whose static resistance is not zero.

Another feature of this circuit 1 is that the input signal current 1i rises in the positive direction without changing the static or physical configuration.
The advantage is that it is also possible to obtain an output current pulse that rises in the negative direction with respect to n.

In the configuration shown in FIG. 3A, contrary to the above, the inductance L in the power line is intentionally made large, while the inductance L2 in the control line is made as small as possible.

Then, as shown in FIG. 4B, as the input signal current 1in to I,a rises, the skid Q first changes to the voltage state, and the output terminal 7 is connected to the drawing current Van. A negative direction current n flows, and the output current 1out rises in the negative direction.

After that, after the delay time TD determined by the time constant τ5=Lt/Rt has elapsed, the positive direction current 6 is supplied via the output line 6, so the output current Icyut reaches the zero level as a result of the addition of both. Returning, the intended one-shot action is completed.

In either case, the circuit is reset at the fall of the power supply Vcc, (2) after the fall of the input signal current 1 inch.

As described above, the circuit 1 shown in FIG. 3A has been proven to have sufficient functionality in principle, but if we look at it in more detail, there are points that would be desirable if it were further improved. First, regarding the rising edge of the output current pulse, the rising edge of the shunt current X6 or I, which is involved in this rising edge, is caused by the large inductance L in the opposing line r or 6, or by It may become somewhat dull due to the influence of Compared to the conventional example, a practically sufficient rise can be obtained depending on the pulse width, but it is of course desirable if the speed can be further increased.

In addition, the amplitude of the output current pulse is also determined by the fact that almost the majority of the input current I,1 is consumed in the shunt to the control line r.
You may not be able to collect too much.

If only to increase the output amplitude, it is sufficient to increase the critical current value of the switching gate Jt, which is preferably provided in the input section, so that the output current becomes higher.
Since the output current output h' branch I6 and the control branch 8 are not separated, from a practical standpoint,
Designing to ensure a wide operating margin becomes somewhat complicated.

In order to solve these problems, as shown in Figure 5B, a buffer amplifier io having a current amplification ability similar to that in the input section may be inserted into the output line.

This buffer amplifier itself may be any direct coupling type or current injection type switching game (JR), and since the explanation given regarding it Jt in the input section can be used, a more detailed explanation will be omitted here.

As described above, by providing the buffer amplifier 10, it is possible to substantially separate the output current 1out and the control current Is, which eliminates the above-mentioned weaknesses, and also makes it possible to obtain a large output current amplitude commensurate with the amplification. I can do it.

Further, in this case, the Josephson critical current values of all the switching gates can be made equal, and design efficiency is improved. In the circuit shown in Figure 5B, the switching gates Js and Qt do not necessarily need to be operated in the gap voltage range. In the circuit illustrated in 3B/, other components not explained may be the same as in the circuit illustrated in 3A,
In the figures, corresponding parts are given the same reference numerals. However, the design resistance for ensuring the output zero level is R1'.

In this circuit I, when shortening the pulse width,
If you want to design a relatively large output pulse, for example, for a positive output pulse, you can increase the inductance according to the above principle, but if it becomes too large, it will be difficult to manufacture and increase the circuit area. Therefore, the value of the resistor R2 should be made small on the condition that a value greater than that is ensured.

It is clear from the above that the degree of freedom of the pulse width TD can be designed to be sufficiently large in principle in the circuit of the present invention, but even if the circuit of the present invention is manufactured using the same parameters, This includes the effect that the pulse width TD can be made variable under actual operating conditions.

That is, as shown by the virtual line Pb in FIGS. 5A and 5B, a skid q is selected as the switching gate 2, and a magnetically coupled bias current line Pb is attached to the skid q so that a magnetic field bias current can be applied to the skid q, thereby increasing the bias current. If IBt is allowed to flow, the control current I3 to the skid Q will be Is = l6j(1-g-'')-(1
), and if the current value when switching the skid Q is IBo, the delay time b' is the bias current lm
The following relationship holds true for the delay time at =00.

To'=Totn(IsJIsr-(I8(+-IB)
) (2) Therefore, the pulse width can be made variable by making the bias current In variable.

The distortion is shown schematically and qualitatively as shown in Figure 5.

The operating point P is determined by a specific control current value 180 depending on the circuit current IcO value to the skid Q, but the bias current I
When B=0, the bias current IB is set to, for example, IB =
When applied as IBI + IB = IBm,
The time TD' until the operating point P is exceeded is %' =
It can be seen that it changes as TD'+ + Tp' = TD't.

As described above, according to the present invention, it is possible to provide a Josephson monostable multivibrator that has a relatively simple and compact configuration, has a fast rise time, and has a wide degree of freedom in pulse width design. Not only does it have an extremely large effect as a single unit, but it also provides a solution to timing problems in Josephson junction circuits. In particular, considering that in the current latching operation, several timings are required to control logical operations, the present invention has the following features:
This makes it possible to easily design such a timing circuit and also to realize subsystems on the same chip.

[Brief explanation of the drawing]

Fig. 1 is an explanatory diagram of the functions required for a Josephson monostable multivibrator, Fig. 2 is a circuit diagram of a conventional Josephson monostable multivibrator, and Fig. 3 is a circuit diagram of each embodiment of the present invention. FIG. 4 is an explanatory diagram of waveforms of various parts in an example of the operation of the present invention, and FIG. 5 is an explanatory diagram of the pulse width variable operation. In the figure, / is the monostable multivibrator as a whole, ko is the input part, 3 is the input terminal, 6 is the output line, 7 is the output terminal, r is the branched control line, ? is a switching gate. @Figure 1 Figure 3A Figure 4 (A) (B) Section 6 Figure

Claims (1)

[Claims] An input section that receives an input signal current and outputs a signal current having at least the same logical information as the input signal current, one end of which is connected to the output of the input section, and the other end of which is connected to an output terminal. a Josephson switching gate that selectively pulls the output current in the output line into the power supply by a switching operation using a control input current, and a Josephson switching gate that connects one end to the output of the input section and branches to the output line. a control line, the other end of which is connected to the control input of the switching gate, the control line having a different time constant between the output line and the control line. Son monostable multivibrator.