IPhO 2016 · Theory problem 2

Nonlinear Dynamics in Electric Circuits

Please read the general instructions in the separate envelope before you start this problem.

Introduction

Bistable non-linear semiconducting elements (e.g. thyristors) are widely used in electronics as switches and generators of electromagnetic oscillations. The primary field of applications of thyristors is controlling alternating currents in power electronics, for instance rectification of AC current to DC at the megawatt scale. Bistable elements may also serve as model systems for self-organization phenomena in physics (this topic is covered in part B of the problem), biology (see part C) and other fields of modern nonlinear science.

Goals

To study instabilities and nontrivial dynamics of circuits including elements with non-linear \(I-V\) characteristics. To discover possible applications of such circuits in engineering and in modeling of biological systems.

Part A. Stationary states and instabilities (3 points)

Fig. 1 shows the so-called S-shaped \(I-V\) characteristics of a non-linear element \(X\). In the voltage range between \(U_{\mathrm h}=4.00\ \mathrm V\) (the holding voltage) and \(U_{\mathrm{th}}=10.0\ \mathrm V\) (the threshold voltage) this \(I-V\) characteristics is multivalued. For simplicity, the graph on Fig. 1 is chosen to be piece-wise linear (each branch is a segment of a straight line). In particular, the line in the upper branch touches the origin if it is extended. This approximation gives a good description of real thyristors.

Figure 1

Figure 1: \(I-V\) characteristics of the non-linear element \(X\).

A.1 — 0.4pt

Using the graph, determine the resistance \(R_{\mathrm{on}}\) of the element \(X\) on the upper branch of the \(I-V\) characteristics, and \(R_{\mathrm{off}}\) on the lower branch, respectively. The middle branch is described by the equation

\[ I=I_0-\frac{U}{R_{\mathrm{int}}}. \tag{1} \]

Find the values of the parameters \(I_0\) and \(R_{\mathrm{int}}\).

The element \(X\) is connected in series (see Fig.2) with a resistor \(R\), an inductor \(L\) and an ideal voltage source \(\mathcal E\). One says that the circuit is in a stationary state if the current is constant in time, \(I(t)=\mathrm{const}\).

Figure 2

Figure 2: Circuit with element \(X\), resistor \(R\), inductor \(L\) and voltage source \(\mathcal E\).

A.2 — 1pt

What are the possible numbers of stationary states that the circuit of Fig. 2 may have for a fixed value of \(\mathcal E\) and for \(R=3.00\ \Omega\)? How does the answer change for \(R=1.00\ \Omega\)?

A.3 — 0.6pt

Let \(R=3.00\ \Omega\), \(L=1.00\ \mu\mathrm H\) and \(\mathcal E=15.0\ \mathrm V\) in the circuit shown in Fig. 2. Determine the values of the current \(I_{\mathrm{stationary}}\) and the voltage \(V_{\mathrm{stationary}}\) on the non-linear element \(X\) in the stationary state.

The circuit in Fig. 2 is in the stationary state with \(I(t)=I_{\mathrm{stationary}}\). This stationary state is said to be stable if after a small displacement (increase or decrease in the current), the current returns towards the stationary state. And if the system keeps moving away from the stationary state, it is said to be unstable.

A.4 — 1pt

Use numerical values of the question A.3 and study the stability of the stationary state with \(I(t)=I_{\mathrm{stationary}}\). Is it stable or unstable?

Part B. Bistable non-linear elements in physics: radio transmitter (5 points)

We now investigate a new circuit configuration (see Fig. 3). This time, the non-linear element \(X\) is connected in parallel to a capacitor of capacitance \(C=1.00\ \mu\mathrm F\). This block is then connected in series to a resistor of resistance \(R=3.00\ \Omega\) and an ideal constant voltage source of voltage \(\mathcal E=15.0\ \mathrm V\). It turns out that this circuit undergoes oscillations with the non-linear element \(X\) jumping from one branch of the \(I-V\) characteristics to another over the course of one cycle.

Figure 3

Figure 3: Circuit with element \(X\), capacitor \(C\), resistor \(R\) and voltage source \(\mathcal E\).

B.1 — 1.8pt

Draw the oscillation cycle on the \(I-V\) graph, including its direction (clockwise or anticlockwise). Justify your answer with equations and sketches.

B.2 — 1.9pt

Find expressions for the times \(t_1\) and \(t_2\) that the system spends on each branch of the \(I-V\) graph during the oscillation cycle. Determine their numerical values. Find the numerical value of the oscillation period \(T\) assuming that the time needed for jumps between the branches of the \(I-V\) graph is negligible.

B.3 — 0.7pt

Estimate the average power \(P\) dissipated by the non-linear element over the course of one oscillation. An order of magnitude is sufficient.

The circuit in Fig. 3 is used to build a radio transmitter. For this purpose, the element \(X\) is attached to one end of a linear antenna (a long straight wire) of length \(s\). The other end of the wire is free. In the antenna, an electromagnetic standing wave is formed. The speed of electromagnetic waves along the antenna is the same as in vacuum. The transmitter is using the main harmonic of the system, which has period \(T\) of question B.2.

B.4 — 0.6pt

What is the optimal value of \(s\) assuming that it cannot exceed \(1\ \mathrm{km}\)?

Part C. Bistable non-linear elements in biology: neuristor (2 points)

In this part of the problem, we consider an application of bistable non-linear elements to modeling of biological processes. A neuron in a human brain has the following property: when excited by an external signal, it makes one single oscillation and then returns to its initial state. This feature is called excitability. Due to this property, pulses can propagate in the network of coupled neurons constituting the nerve systems. A semiconductor chip designed to mimic excitability and pulse propagation is called a neuristor (from neuron and transistor).

We attempt to model a simple neuristor using a circuit that includes the non-linear element \(X\) that we investigated previously. To this end, the voltage \(\mathcal E\) in the circuit of Fig. 3 is decreased to the value \(\mathcal E'=12.0\ \mathrm V\). The oscillations stop, and the system reaches its stationary state. Then, the voltage is rapidly increased back to the value \(\mathcal E=15.0\ \mathrm V\), and after a period of time \(\tau\) (with \(\tau<T\)) is set again to the value \(\mathcal E'\) (see Fig. 4). It turns out that there is a certain critical value \(\tau_{\mathrm{crit}}\), and the system shows qualitatively different behavior for \(\tau<\tau_{\mathrm{crit}}\) and for \(\tau>\tau_{\mathrm{crit}}\).

Figure 4

Figure 4: Voltage of the voltage source as a function of time.

C.1 — 1.2pt

Sketch the graphs of the time dependence of the current \(I_X(t)\) on the non-linear element \(X\) for \(\tau<\tau_{\mathrm{crit}}\) and for \(\tau>\tau_{\mathrm{crit}}\).

C.2 — 0.6pt

Find the expression and the numerical value of the critical time \(\tau_{\mathrm{crit}}\) for which the scenario switches.

C.3 — 0.2pt

Is the circuit with \(\tau=1.00\times10^{-6}\ \mathrm s\) a neuristor?

General Data Sheet

Speed of light in vacuum \(c\) \(=\) \(299\,792\,458\ \mathrm{m\cdot s^{-1}}\)
Vacuum permeability (magnetic constant) \(\mu_0\) \(=\) \(4\pi\times10^{-7}\ \mathrm{kg\cdot m\cdot A^{-2}\cdot s^{-2}}\)
Vacuum permittivity (electrical constant) \(\varepsilon_0\) \(=\) \(8.854\,187\,817\times10^{-12}\ \mathrm{A^2\cdot s^4\cdot kg^{-1}\cdot m^{-3}}\)
Elementary charge \(e\) \(=\) \(1.602\,176\,620\,8(98)\times10^{-19}\ \mathrm{A\cdot s}\)
Mass of the electron \(m_{\mathrm e}\) \(=\) \(9.109\,383\,56(11)\times10^{-31}\ \mathrm{kg}\)
\(=\) \(0.510\,998\,946\,1(31)\ \dfrac{\mathrm{MeV}}{c^2}\)
Mass of the proton \(m_{\mathrm p}\) \(=\) \(1.672\,621\,898(21)\times10^{-27}\ \mathrm{kg}\)
\(=\) \(938.272\,081\,3(58)\ \dfrac{\mathrm{MeV}}{c^2}\)
Mass of the neutron \(m_{\mathrm n}\) \(=\) \(1.674\,927\,471(21)\times10^{-27}\ \mathrm{kg}\)
\(=\) \(939.565\,413\,3(58)\ \dfrac{\mathrm{MeV}}{c^2}\)
Unified atomic mass unit \(u\) \(=\) \(1.660\,539\,040(20)\times10^{-27}\ \mathrm{kg}\)
Rydberg constant \(R_\infty\) \(=\) \(10\,973\,731.568\,508(65)\ \mathrm{m^{-1}}\)
Universal constant of gravitation \(G\) \(=\) \(6.674\,08(31)\times10^{-11}\ \mathrm{m^3\cdot kg^{-1}\cdot s^{-2}}\)
Acceleration due to gravity (in Zurich) \(g\) \(=\) \(9.81\ \mathrm{m\cdot s^{-2}}\)
Planck’s constant \(h\) \(=\) \(6.626\,070\,040\,(81)\times10^{-34}\ \mathrm{kg\cdot m^2\cdot s^{-1}}\)
Avogadro number \(N_{\mathrm A}\) \(=\) \(6.022\,140\,857\,(74)\times10^{23}\ \mathrm{mol^{-1}}\)
Molar gas constant \(R\) \(=\) \(8.314\,4598(48)\ \mathrm{kg\cdot m^2\cdot s^{-2}\cdot mol^{-1}\cdot K^{-1}}\)
Molar mass constant \(M_u\) \(=\) \(1\times10^{-3}\ \mathrm{kg\cdot mol^{-1}}\)
Boltzmann constant \(k_{\mathrm B}\) \(=\) \(1.380\,648\,52(79)\times10^{-23}\ \mathrm{kg\cdot m^2\cdot s^{-2}\cdot K^{-1}}\)
Stefan-Boltzmann constant \(\sigma\) \(=\) \(5.670\,367\,(13)\times10^{-8}\ \mathrm{kg\cdot s^{-3}\cdot K^{-4}}\)

Official English text of the problem.

Answers

  • A.1: \(R_{\mathrm{on}}=1.00\ \Omega\), \(R_{\mathrm{off}}=10.0\ \Omega\), \(I_0=6.00\ \mathrm A\), \(R_{\mathrm{int}}=2.00\ \Omega\).
  • A.2: For \(R=3.00\ \Omega\), there is one stationary state. For \(R=1.00\ \Omega\), there are three when \(8.00\ \mathrm V<\mathcal E<11.0\ \mathrm V\), two at either endpoint, and one otherwise, for the nonnegative source voltage shown.
  • A.3: \(I_{\mathrm{stationary}}=3.00\ \mathrm A\), \(V_{\mathrm{stationary}}=6.00\ \mathrm V\).
  • A.4: Stable; small current perturbations decay exponentially with time constant \(1.00\ \mu\mathrm s\).
  • B.1: Anticlockwise. In coordinates \((U/\mathrm V,I_X/\mathrm A)\), the cycle is \[ (4,0.4)\rightarrow(10,1)\rightarrow(10,10) \rightarrow(4,4)\rightarrow(4,0.4). \]
  • B.2: Defining \(t_1\) on the lower branch and \(t_2\) on the upper branch, \[ t_1=\frac{30}{13}\ln\frac{49}{10}\ \mu\mathrm s=3.67\ \mu\mathrm s, \qquad t_2=\frac34\ln25\ \mu\mathrm s=2.41\ \mu\mathrm s, \] \[ T=6.08\ \mu\mathrm s. \]
  • B.3: \(P\) is of order \(10^1\ \mathrm W\). Integration within the ideal model gives \(P=17.3\ \mathrm W\).
  • B.4: \(s=cT/4=456\ \mathrm m\).
  • C.1: For \(\tau<\tau_{\mathrm{crit}}\), the current makes a small, continuous excursion from \(0.923\ \mathrm A\), remaining below \(1.00\ \mathrm A\), and returns to \(0.923\ \mathrm A\). For \(\tau>\tau_{\mathrm{crit}}\), it produces one large spike: a jump \(1.00\to10.0\ \mathrm A\), decay to \(4.00\ \mathrm A\), a jump to \(0.400\ \mathrm A\), and recovery to \(0.923\ \mathrm A\).
  • C.2: \[ \tau_{\mathrm{crit}}=\frac{30}{13}\ln\frac32\ \mu\mathrm s =0.936\ \mu\mathrm s. \]
  • C.3: Yes, because \(1.00\ \mu\mathrm s>\tau_{\mathrm{crit}}\), the pulse triggers one complete excursion.

Solution

A.1

The upper branch passes through the origin when extended and through \((U,I)=(4.00\ \mathrm V,4.00\ \mathrm A)\). Hence

\[ R_{\mathrm{on}}=\frac UI=\frac{4.00}{4.00}\ \Omega =\boxed{1.00\ \Omega}. \]

The lower branch passes through the origin and \((10.0\ \mathrm V,1.00\ \mathrm A)\), giving

\[ R_{\mathrm{off}}=\boxed{10.0\ \Omega}. \]

The middle branch joins \((4.00\ \mathrm V,4.00\ \mathrm A)\) to \((10.0\ \mathrm V,1.00\ \mathrm A)\), so

\[ -\frac1{R_{\mathrm{int}}} =\frac{1.00-4.00}{10.0-4.00}\ \frac{\mathrm A}{\mathrm V} =-0.500\ \Omega^{-1}. \]

Therefore

\[ \boxed{R_{\mathrm{int}}=2.00\ \Omega}, \qquad I_0=4.00\ \mathrm A+\frac{4.00\ \mathrm V}{2.00\ \Omega} =\boxed{6.00\ \mathrm A}. \]

Here \(R_{\mathrm{int}}\) is positive; the middle branch’s differential resistance is \(-R_{\mathrm{int}}\).

A.2

Let \(U\) denote the voltage across \(X\). Kirchhoff’s voltage law gives

\[ \mathcal E=RI+U+L\frac{dI}{dt}. \]

In a stationary state,

\[ \mathcal E=RI+U. \]

Thus stationary states are intersections of the characteristic with the load line \(I=(\mathcal E-U)/R\).

Equivalently, using the characteristic as \(U(I)\),

\[ \mathcal E(I)= \begin{cases} (R+10.0\ \Omega)I, &0\le I\le1.00\ \mathrm A,\\[2mm] 12.0\ \mathrm V+(R-2.00\ \Omega)I, &1.00\ \mathrm A\le I\le4.00\ \mathrm A,\\[2mm] (R+1.00\ \Omega)I, &I\ge4.00\ \mathrm A. \end{cases} \]

For \(R=3.00\ \Omega\), all three pieces increase with \(I\). They join continuously at \(\mathcal E=13.0\ \mathrm V\) and \(16.0\ \mathrm V\). Consequently, every nonnegative source voltage has exactly one stationary state.

For \(R=1.00\ \Omega\), the middle piece decreases with \(I\). Its endpoints occur at

\[ \mathcal E(1.00\ \mathrm A)=11.0\ \mathrm V, \qquad \mathcal E(4.00\ \mathrm A)=8.00\ \mathrm V. \]

The number of distinct stationary states is therefore

Source voltage Number of stationary states
\(0\le\mathcal E<8.00\ \mathrm V\) 1
\(\mathcal E=8.00\ \mathrm V\) 2
\(8.00\ \mathrm V<\mathcal E<11.0\ \mathrm V\) 3
\(\mathcal E=11.0\ \mathrm V\) 2
\(\mathcal E>11.0\ \mathrm V\) 1

At either boundary, two branch intersections coincide at a corner of the characteristic and count as one state.

A.3

For \(R=3.00\ \Omega\) and \(\mathcal E=15.0\ \mathrm V\), the stationary state lies on the middle branch because \(13.0<15.0<16.0\ \mathrm V\). Hence

\[ 15.0\ \mathrm V =12.0\ \mathrm V+(3.00-2.00)\ \Omega\,I_{\mathrm{stationary}}. \]

Therefore

\[ \boxed{I_{\mathrm{stationary}}=3.00\ \mathrm A}, \]

and

\[ \boxed{ V_{\mathrm{stationary}} =\mathcal E-RI_{\mathrm{stationary}} =6.00\ \mathrm V}. \]

A.4

Write \(I=I_{\mathrm{stationary}}+\delta I\). A sufficiently small perturbation remains on the middle branch, where

\[ U=R_{\mathrm{int}}(I_0-I). \]

Substituting into Kirchhoff’s law and subtracting the stationary equation gives

\[ L\frac{d(\delta I)}{dt} =-(R-R_{\mathrm{int}})\delta I. \]

Thus

\[ \delta I(t)=\delta I(0) \exp\left[-\frac{R-R_{\mathrm{int}}}{L}t\right]. \]

Since \(R-R_{\mathrm{int}}=1.00\ \Omega>0\), the perturbation decays. The stationary state is therefore stable, with relaxation time

\[ \boxed{ \frac{L}{R-R_{\mathrm{int}}} =\frac{1.00\ \mu\mathrm H}{1.00\ \Omega} =1.00\ \mu\mathrm s}. \]

The external resistor’s positive differential resistance exceeds the magnitude of the element’s negative differential resistance.

B.1

The capacitor and \(X\) have the same voltage \(U\). Kirchhoff’s current law gives

\[ \frac{\mathcal E-U}{R}=I_X+C\frac{dU}{dt}, \]

or

\[ C\frac{dU}{dt}=\frac{\mathcal E-U}{R}-I_X(U). \]

On a branch of resistance \(r\), \(I_X=U/r\), so

\[ C\frac{dU}{dt}=\frac{\mathcal E}{R} -\left(\frac1R+\frac1r\right)U. \]

On the lower branch, the voltage towards which the circuit would relax is

\[ U_{\infty,\mathrm{off}} =\frac{\mathcal E R_{\mathrm{off}}}{R+R_{\mathrm{off}}} =\frac{150}{13}\ \mathrm V =11.538\ldots\ \mathrm V. \]

This exceeds \(U_{\mathrm{th}}=10.0\ \mathrm V\). Consequently, throughout the lower-branch part of the cycle, \(dU/dt>0\): the capacitor charges from \(4.00\) to \(10.0\ \mathrm V\).

At \(U_{\mathrm{th}}\), the lower branch ends. The capacitor voltage cannot jump, so \(X\) switches vertically on the \(I-V\) plot:

\[ (U,I_X)=(10.0\ \mathrm V,1.00\ \mathrm A) \longrightarrow(10.0\ \mathrm V,10.0\ \mathrm A). \]

On the upper branch, the corresponding limiting voltage is

\[ U_{\infty,\mathrm{on}} =\frac{\mathcal E R_{\mathrm{on}}}{R+R_{\mathrm{on}}} =3.75\ \mathrm V. \]

This lies below \(U_{\mathrm h}=4.00\ \mathrm V\). Thus \(dU/dt<0\): the capacitor discharges from \(10.0\) to \(4.00\ \mathrm V\). At the holding voltage, the current jumps from \(4.00\) to \(0.400\ \mathrm A\), again at constant voltage.

The requested sketch therefore has these successive segments:

Segment Motion in the \(I-V\) plane
Lower branch From \((4.00\ \mathrm V,0.400\ \mathrm A)\) rightwards and slightly upwards to \((10.0\ \mathrm V,1.00\ \mathrm A)\)
Threshold switch Vertically upwards to \((10.0\ \mathrm V,10.0\ \mathrm A)\)
Upper branch Leftwards and downwards to \((4.00\ \mathrm V,4.00\ \mathrm A)\)
Holding-voltage switch Vertically downwards to \((4.00\ \mathrm V,0.400\ \mathrm A)\)

The direction is anticlockwise.

The middle-branch equilibrium is unstable in this capacitor circuit. Indeed, near \(U=6.00\ \mathrm V\),

\[ C\frac{d(\delta U)}{dt} =\left(\frac1{R_{\mathrm{int}}}-\frac1R\right)\delta U =\frac{\delta U}{6.00\ \Omega}. \]

Voltage perturbations grow, so the middle branch is not a sustained part of the oscillation cycle.

B.2

For a branch of resistance \(r\), define

\[ \theta_r=C\frac{Rr}{R+r}, \qquad U_{\infty,r}=\frac{\mathcal E r}{R+r}. \]

The differential equation from B.1 becomes

\[ \frac{dU}{dt}=\frac{U_{\infty,r}-U}{\theta_r}, \]

with solution

\[ U(t)=U_{\infty,r} +\bigl(U(0)-U_{\infty,r}\bigr)e^{-t/\theta_r}. \]

Define \(t_1\) as the time on the lower branch and \(t_2\) as the time on the upper branch.

For charging from \(U_{\mathrm h}\) to \(U_{\mathrm{th}}\),

\[ \boxed{ t_1=\theta_{\mathrm{off}} \ln\left( \frac{U_{\infty,\mathrm{off}}-U_{\mathrm h}} {U_{\infty,\mathrm{off}}-U_{\mathrm{th}}} \right)}. \]

Here

\[ \theta_{\mathrm{off}} =\frac{30}{13}\ \mu\mathrm s, \]

so

\[ t_1 =\frac{30}{13}\ \mu\mathrm s \ln\left(\frac{150/13-4}{150/13-10}\right) =\frac{30}{13}\ln\frac{49}{10}\ \mu\mathrm s =\boxed{3.67\ \mu\mathrm s}. \]

For discharging from \(U_{\mathrm{th}}\) to \(U_{\mathrm h}\),

\[ \boxed{ t_2=\theta_{\mathrm{on}} \ln\left( \frac{U_{\mathrm{th}}-U_{\infty,\mathrm{on}}} {U_{\mathrm h}-U_{\infty,\mathrm{on}}} \right)}. \]

Since

\[ \theta_{\mathrm{on}}=0.750\ \mu\mathrm s, \]

we obtain

\[ t_2=0.750\ \mu\mathrm s \ln\left(\frac{10.0-3.75}{4.00-3.75}\right) =\frac34\ln25\ \mu\mathrm s =\boxed{2.41\ \mu\mathrm s}. \]

Neglecting the switching durations,

\[ \boxed{T=t_1+t_2=6.08\ \mu\mathrm s}. \]

B.3

On either conducting branch, the instantaneous power dissipated in \(X\) is

\[ p_X=UI_X=\frac{U^2}{r}. \]

We neglect energy dissipated during the ideal instantaneous switches. Using the voltage evolution derived above, the energy dissipated on each branch is

\[ W_{\mathrm{off}} =\frac{\theta_{\mathrm{off}}}{R_{\mathrm{off}}} \int_{U_{\mathrm h}}^{U_{\mathrm{th}}} \frac{U^2}{U_{\infty,\mathrm{off}}-U}\,dU =23.2\ \mu\mathrm J, \]

and

\[ W_{\mathrm{on}} =\frac{\theta_{\mathrm{on}}}{R_{\mathrm{on}}} \int_{U_{\mathrm h}}^{U_{\mathrm{th}}} \frac{U^2}{U-U_{\infty,\mathrm{on}}}\,dU =82.3\ \mu\mathrm J. \]

Thus

\[ P=\frac{W_{\mathrm{off}}+W_{\mathrm{on}}}{T} =\frac{105.5\ \mu\mathrm J}{6.0816\ \mu\mathrm s} =\boxed{17.3\ \mathrm W}. \]

The requested order-of-magnitude estimate is therefore

\[ \boxed{P\sim10^1\ \mathrm W}. \]

B.4

The emitted fundamental has wavelength

\[ \lambda=cT =(299\,792\,458\ \mathrm{m\,s^{-1}}) (6.08162\times10^{-6}\ \mathrm s) =1.82322\times10^3\ \mathrm m. \]

In the ideal antenna model, the free end is a current node, while efficient excitation places a current antinode at the driven end. Their separation is an odd multiple of a quarter wavelength:

\[ s=(2n+1)\frac{\lambda}{4}, \qquad n=0,1,2,\ldots \]

The shortest resonant length is

\[ \boxed{s=\frac{\lambda}{4}=\frac{cT}{4}=456\ \mathrm m}. \]

The next possibility is \(3\lambda/4=1.37\ \mathrm{km}\), which exceeds the limit. Thus the quarter-wave length is the permitted optimum.

C.1

First find the initial stationary state at \(\mathcal E'=12.0\ \mathrm V\). It lies on the lower branch:

\[ I_{\mathrm r}=\frac{\mathcal E'}{R+R_{\mathrm{off}}} =\frac{12}{13}\ \mathrm A =\boxed{0.923\ \mathrm A}, \]

\[ U_{\mathrm r}=R_{\mathrm{off}}I_{\mathrm r} =\frac{120}{13}\ \mathrm V =\boxed{9.23\ \mathrm V}. \]

Let \(u=t-t_0\). Before the pulse, \(I_X=I_{\mathrm r}\). While the source is at \(15.0\ \mathrm V\) and the element remains on the lower branch,

\[ I_X(u) =\frac{\mathcal E}{R+R_{\mathrm{off}}} +\left(I_{\mathrm r} -\frac{\mathcal E}{R+R_{\mathrm{off}}}\right) e^{-u/\theta_{\mathrm{off}}}, \]

or numerically,

\[ I_X(u)=\frac{15-3e^{-u/\theta_{\mathrm{off}}}}{13}\ \mathrm A. \]

The critical duration is the time required for this current to reach \(1.00\ \mathrm A\), corresponding to \(U=U_{\mathrm{th}}\).

For \(\tau<\tau_{\mathrm{crit}}\), the graph is a small continuous bump:

  • Before \(t_0\), the current is constant at \(0.923\ \mathrm A\).
  • During the pulse, it rises exponentially, with decreasing slope, but remains below \(1.00\ \mathrm A\).
  • At \(t_0+\tau\), the current is continuous and its slope changes sign.
  • Afterwards it decays exponentially back to \(0.923\ \mathrm A\):

\[ I_X(u)=I_{\mathrm r} +\bigl[I_X(\tau)-I_{\mathrm r}\bigr] e^{-(u-\tau)/\theta_{\mathrm{off}}}, \qquad u\ge\tau. \]

There is no branch switch.

For \(\tau>\tau_{\mathrm{crit}}\), the graph contains one large spike:

  1. The current initially follows the same gradual rise from \(0.923\) to \(1.00\ \mathrm A\).
  2. At \(t=t_0+\tau_{\mathrm{crit}}\), it jumps vertically from \(1.00\) to \(10.0\ \mathrm A\).
  3. It then decreases on the upper branch. While the source remains at \(15.0\ \mathrm V\), \[ I_X(u)=3.75\ \mathrm A +6.25\ \mathrm A\, e^{-(u-\tau_{\mathrm{crit}})/\theta_{\mathrm{on}}}. \]
  4. On reaching \(4.00\ \mathrm A\), it jumps vertically to \(0.400\ \mathrm A\).
  5. It subsequently rises on the lower branch and ultimately approaches \(0.923\ \mathrm A\) from below.

If the source returns to \(12.0\ \mathrm V\) while the element is still on the upper branch, the current remains continuous at that instant, but its decay becomes faster. Its subsequent expression is

\[ I_X(u)=3.00\ \mathrm A +\bigl[I_X(\tau)-3.00\ \mathrm A\bigr] e^{-(u-\tau)/\theta_{\mathrm{on}}}. \]

The asymptote \(3.00\ \mathrm A\) lies below the holding current, so the current necessarily reaches \(4.00\ \mathrm A\) and switches off. If this occurs at \(u=u_d\), the following recovery is

\[ I_X(u)=I_{\mathrm r} +\bigl(0.400\ \mathrm A-I_{\mathrm r}\bigr) e^{-(u-u_d)/\theta_{\mathrm{off}}}. \]

For a longer pulse, the upper-to-lower switch can occur before the pulse ends. If the source stays at \(15.0\ \mathrm V\), that switch occurs at

\[ u=\tau_{\mathrm{crit}}+t_2=3.35\ \mu\mathrm s. \]

The current then recovers on the lower branch under the \(15.0\ \mathrm V\) source until \(u=\tau\), after which it approaches the \(12.0\ \mathrm V\) rest state. The source changes produce slope changes, not instantaneous changes in \(I_X\).

There cannot be a second spike during the prescribed pulse: with \(15.0\ \mathrm V\) maintained, the second threshold crossing would occur at \(u=\tau_{\mathrm{crit}}+T>T>\tau\).

C.2

At the critical duration, the capacitor has just charged from \(U_{\mathrm r}\) to \(U_{\mathrm{th}}\):

\[ U_{\mathrm{th}} =U_{\infty,\mathrm{off}} +\bigl(U_{\mathrm r}-U_{\infty,\mathrm{off}}\bigr) e^{-\tau_{\mathrm{crit}}/\theta_{\mathrm{off}}}. \]

Solving,

\[ \boxed{ \tau_{\mathrm{crit}} =\theta_{\mathrm{off}} \ln\left( \frac{U_{\infty,\mathrm{off}}-U_{\mathrm r}} {U_{\infty,\mathrm{off}}-U_{\mathrm{th}}} \right)}. \]

In terms of the circuit parameters,

\[ \boxed{ \tau_{\mathrm{crit}} =\frac{CRR_{\mathrm{off}}}{R+R_{\mathrm{off}}} \ln\left[ \frac{(\mathcal E-\mathcal E')R_{\mathrm{off}}} {\mathcal E R_{\mathrm{off}} -(R+R_{\mathrm{off}})U_{\mathrm{th}}} \right]}. \]

Numerically,

\[ \tau_{\mathrm{crit}} =\frac{30}{13}\ \mu\mathrm s \ln\left( \frac{150/13-120/13}{150/13-10} \right) =\frac{30}{13}\ln\frac32\ \mu\mathrm s. \]

Therefore

\[ \boxed{\tau_{\mathrm{crit}}=0.936\ \mu\mathrm s}. \]

C.3

The specified pulse satisfies

\[ \tau=1.00\ \mu\mathrm s >\tau_{\mathrm{crit}}=0.936\ \mu\mathrm s, \qquad \tau<T. \]

It therefore triggers the switch to the upper branch. The circuit completes one large current excursion and then returns to its original stationary state under the \(12.0\ \mathrm V\) source.

\[ \boxed{\text{Yes, the circuit behaves as a neuristor.}} \]

The solution was checked against the official solution.