I'm trying to figure out how to construct a differential equation for the natural response of the following circuit and I am having trouble. The capacitor is initially charged with a voltage \$v_0\$ and I want to solve for \$v(t)\$, the voltage across the capacitor.


simulate this circuit – Schematic created using CircuitLab

Here is my attempt so far:

By Kirchhoff's laws, \$v_L=v_2\$, \$v=v_1+v_L\$, and \$i_1+i_2+i_3=0\$. The currents and voltages are defined so that \$i_1=-\frac{v_1}{R_1}=C\frac{dv}{dt}\$, \$i_2=\frac{1}{L}\int{v_Ldt}\$, and \$i_3=\frac{v_2}{R_2}=\frac{v_L}{R_2}\$. $$C\frac{dv}{dt}+\frac{1}{L}\int{v_Ldt}+\frac{v_L}{R_2}=0$$ And I cannot come up with an equation the exclusively solves for \$v\$. Is there even an existing solution to this problem?

  • \$\begingroup\$ You need to have an initial value for the inductor current as well as the capacitor voltage. \$\endgroup\$ – The Photon Mar 31 '18 at 5:18
  • \$\begingroup\$ It's self-explanatory. The inductor acts like an open circuit at t=0, by the time the capacitor is connected to the circuit. The capacitor is initially charged with \$v_0\$ as stated above and cannot change abruptly. But, how can this help me form the correct differential equation? \$\endgroup\$ – mjtsquared Mar 31 '18 at 5:25
  • \$\begingroup\$ nowhwere did you say the capacitor was connected to the circuit at t=0, you just stated it's charge at that time. To solve this kind of circuit you need to know the initial conditions of each energy storage element --- both capacitors and inductors. \$\endgroup\$ – The Photon Mar 31 '18 at 5:30
  • \$\begingroup\$ I know that initial conditions are required to solve the differential equation. That's not what I wanted to do; I wanted to construct the correct differential equation, if any. \$\endgroup\$ – mjtsquared Mar 31 '18 at 5:44

You just need to add one more equation

$$v_1 = -i_1 R_1$$ $$v_1 = -CR_1\frac{dv}{dt}$$

Thus $$ v_L = v - v_1 = v + C R_1 \frac{dv}{dt}$$

Now you can substitute in your integro-differential equation to get an equation with just one variable

$$C\frac{dv}{dt} + \frac{1}{L}\int_{-\infty}^t (v + CR_1\frac{dv}{dt})dt +\frac{v+CR_1\frac{dv}{dt}}{R_2} = 0$$

From here you can rearrange terms and take derivatives until you reach a differential equation in one variable, that can be solved by the usual methods (recognizing a known form, Laplace transform, etc.)

  • \$\begingroup\$ If I had more patience for Mathjax I'd rearrange the 2nd i-d equation to push the dv/dt back into the integral until the next step where it would fall out again properly. But it's late here so hopefully this is clear enough. \$\endgroup\$ – The Photon Mar 31 '18 at 5:50
  • \$\begingroup\$ I see, but the coefficient to the integral is \$\frac{1}{L}\$ though. \$\endgroup\$ – mjtsquared Mar 31 '18 at 6:04
  • \$\begingroup\$ I got almost that, but (C R1 / L) not (L C R1) and also v/L and not L v. Oh, well. \$\endgroup\$ – jonk Mar 31 '18 at 6:54
  • \$\begingroup\$ @jonk, probably related to the transcription error that mjt already pointed out. \$\endgroup\$ – The Photon Mar 31 '18 at 16:32

Just so you have my approach to consider, as well. (I know you've already selected an answer.) Here's the redrawn schematic which I prefer:


simulate this circuit – Schematic created using CircuitLab

I apply nodal analysis and get these two equations from the two nodes:

$$\begin{align*} \frac{V_\text{C}}{R_1}+C\frac{\text{d} V_\text{C}}{\text{d} t}&= \frac{V_\text{L}}{R_1}\label{n1}\tag{node $V_\text{C}$}\\\\ \frac{V_\text{L}}{R_1}+\frac{V_\text{L}}{R_2}+\frac{1}{L}\int V_\text{L}\:\text{d} t&=\frac{V_\text{C}}{R_1}\label{n2}\tag{node $V_\text{L}$} \end{align*}$$

Just solve the \$\ref{n1}\$ equation for \$V_\text{L}\$:

$$\begin{align*} V_\text{L}&=V_\text{C}+R_1\:C\frac{\text{d} V_\text{C}}{\text{d} t}\label{n3}\tag{solved for $V_\text{L}$} \end{align*}$$

, and then substitute that into the above for \$\ref{n2}\$:

$$\begin{align*} \frac{V_\text{C}+R_1\:C\frac{\text{d} V_\text{C}}{\text{d} t}}{R_1}+\frac{V_\text{C}+R_1\:C\frac{\text{d} V_\text{C}}{\text{d} t}}{R_2}+\frac{1}{L}\int \left[V_\text{C}+R_1\:C\frac{\text{d} V_\text{C}}{\text{d} t}\right]\:\text{d} t&=\frac{V_\text{C}}{R_1} \\\\ \frac{V_\text{C}}{R_1}+\frac{V_\text{C}}{R_2}+C\left(1+\frac{R_1}{R_2}\right)\frac{\text{d} V_\text{C}}{\text{d} t}+\frac{1}{L}\left[\int V_\text{C}\:\text{d} t+R_1\:C\int\text{d} V_\text{C}\right]&=\frac{V_\text{C}}{R_1}\\\\ \frac{V_\text{C}}{R_2}+C\left(1+\frac{R_1}{R_2}\right)\frac{\text{d} V_\text{C}}{\text{d} t}+\frac{1}{L}\left[\int V_\text{C}\:\text{d} t+R_1\:C\int\text{d} V_\text{C}\right]&=0 \end{align*}$$

, now take everything with respect to the derivative of time:

$$\begin{align*} \frac{1}{R_2}\frac{\text{d} V_\text{C}}{\text{d} t}+C\left(1+\frac{R_1}{R_2}\right)\frac{\text{d}^2 V_\text{C}}{\text{d} t^2}+\frac{1}{L}\left[V_\text{C}+R_1\:C\frac{\text{d} V_\text{C}}{\text{d} t}\right]&=0\\\\ C\left(1+\frac{R_1}{R_2}\right)\frac{\text{d}^2 V_\text{C}}{\text{d} t^2}+\left(\frac{1}{R_2}+\frac{R_1\:C}{L}\right)\frac{\text{d} V_\text{C}}{\text{d} t}+\frac{V_\text{C}}{L}&=0\\\\ \end{align*}$$

You can easily put that into standard form and solve using the usual 2nd order diff-eq approach or else use Laplace.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.