Week 10
Second-order examples and systems
This week is practice with complete second-order transient problems: find the initial conditions before and just after a switch moves, the final conditions, the initial slope, and then the full iL(t) or vC(t). It then generalizes transfer functions into system block diagrams with summers and feedback loops, ending with a model of the lab servo. The servo lab shows the same overdamped, ringing and oscillating behavior in an electromechanical system as you turn up the loop gain.
Objectives
- For a switched circuit with one inductor and one capacitor, find iL and vC before the switch moves, the circuit just after (L as a current source, C as a voltage source), and the final conditions, then write the full iL(t) or vC(t).
- Write X(t) with all constants given only the characteristic equation and a table of initial and final conditions, using the overdamped, critically damped or underdamped canned solution.
- Find the poles and zeros of a transfer function and plot them on the complex s-plane.
- Reduce cascaded blocks, summers and a feedback loop to one transfer function, using AB for series blocks and A/(1 + AB) for negative feedback (A/(1 − AB) for positive).
- Find the forward or feedback gain that makes a closed-loop system critically damped, and predict whether a larger gain makes it underdamped or overdamped.
- Solve the characteristic equation's discriminant b² − 4k = 0 for the resistor value that makes a multi-resistor RLC circuit critically damped.
- Relate servo gain settings to overdamped, critically damped, ringing and sustained-oscillation responses, and measure the amplifier gain and the potentiometer constant Kp in V/deg and V/rad.
Lectures
| 1 | Examples 1–2 | 42:32 |
| 2 | Example 3 | 34:04 |
| 3 | Lecture 18 | 1:03:09 |
| 4 | Lecture 19Open original recording | 45:50 |
| 3:05:35 |
Notes
From the notes pack:
- Lecture 18, Ex. 1: transfer function, characteristic equation, differential equation and roots for vL in a circuit with C = 0.22 µF in series feeding R = 120 Ω in parallel with L = 0.5 mH; the complex roots mean underdamped
- Ex. 2: switch closes in a 15 V circuit with R1 = 200 Ω, R2 = 300 Ω, L = 100 mH, C = 0.5 µF; overdamped iL(t) from initial, final and slope conditions
- Ex. 3: switch opens in a 10 V circuit with R1 = 40 Ω, L = 80 mH, R2 = 60 Ω, C = 0.125 µF; underdamped vC(t) and its decaying envelope
- Ex. 4: Ex. 3 run backwards (switch closes), which is overdamped; full vC(t)
- Ex. 5 and Ex. 6: only the characteristic equation and tables of initial and final conditions are given; write iL(t) for an underdamped case and a critically damped case
- Lecture 19, Ex. 1: in a 24 V circuit with 120 Ω, 80 Ω, 40 Ω, L = 0.5 mH and C = 1.5 µF, find the final conditions and the initial conditions and slopes of iL and vC without solving for the constants
- Systems: transfer functions as blocks, a potentiometer as a block (Kp = 0.7 V/rad), blocks in series multiply, summer blocks add or subtract
- Feedback loops: positive feedback gives A/(1 − AB), negative feedback gives A/(1 + AB); AB is the loop gain; open loop versus closed loop
- Block diagram of the lab servo: input potentiometer Kp, summer, circuit gain G, motor-and-gears transfer function, feedback potentiometer Kp, and the overall H(s)
Homework
2ndTr3: Complete second-order responses
- 1a–c. An 8 V source and a switch feed R1 = 2200 Ω in series with C = 40 µF in parallel with a branch of R2 = 150 Ω and L = 20 mH in series: find the characteristic equation, its roots, and the damping type
- 1d. Find the final conditions
- 1e. The switch has been open a long time and closes at t = 0: find the initial conditions
- 1f. Write the full expression for vC(t) with all constants
- 2a–b. A 100 mA current source feeds three parallel branches: a switch in series with R1 = 40 Ω, R2 = 120 Ω, and C = 0.2 µF in series with L = 4 mH; given H(s) = IL/Iin = (R2/L)s / (s² + (R2/L)s + 1/(LC)), find the roots and the damping type
- 2c. The switch opens at t = 0: find the final and initial conditions of iL
- 2d. Write the full expression for iL(t)
2ndTr4: Initial conditions, poles and zeros, feedback systems
- 1. A switch moves from a 12 V source to a 4 V source feeding R1 = 10 Ω, then R2 = 150 Ω in parallel with C = 40 µF, then L = 12 mH to ground; given the characteristic equation s² + 1000s + 2.222×10⁶ = 0, find the initial conditions and write the full iL(t)
- 2. Find the R1 that would make the circuit of problem 1 critically damped
- 3a. The circuit of Hw2ndTr3 problem 1 with R1 = R2 = 50 Ω and the output taken across R1 has H(s) = (s² + 2500s + 1.25×10⁶)/(s² + 3000s + 2.5×10⁶); find its poles and zeros
- 3b. Plot the poles and zeros on the complex plane
- 4a. A negative-feedback system has forward block G/(s + 30) and feedback block 10/(s + 60); find the closed-loop H(s) with a simple polynomial denominator
- 4b–c. With G = 5, find the poles and zeros and the type of damping
- 4d. Find the G that makes the system critically damped
- 4e. Predict the damping if G is doubled from that value
- 5a. Input gain 20 feeds a summer, then forward block 50/(s + 25); the output returns to the summer's minus input through a gain −F and a block −4/(s + 40); find the closed-loop H(s)
- 5b. Find the largest F that keeps the system from becoming underdamped
- 5c. Find H(s) with F = 0.2
- 5d. With F = 0.2, find the s values that make H(s) infinite (output with no input)
- 5e. Decide whether H(s) has a zero, and find it if it does
Lab: Control system (servo) (Lab 9)
Drive the lab's position servo with a square wave and turn its gain knob to see sluggish, well-damped, ringing and oscillating motion, the mechanical version of over-, critically and underdamped responses. You find the gains where overshoot, ringing and sustained oscillation begin, measure the amplifier gain at each, measure the potentiometer constant Kp, and compare with the appendix model.
Before the lab
- Bring a book that shows overdamped, critically damped and underdamped second-order responses if you aren't sure what they look like
- Review how you set the power supply to ±6 V in Lab 1
- Read the appendix: motor model, system block diagram and the predicted responses at G = 1.7, 14 and 40
Equipment
- Servo (from the stockroom)
- Triple-output DC power supply (Agilent E3631A) set to ±6 V
- Function generator (Agilent 33120A)
- Oscilloscope with one 10× probe
- BNC T connector and two BNC-to-BNC cables
- Digital multimeter (voltmeter)
Procedure
- Power supply hookup
- Turn off the servo's power switch.
- Set the power supply to ±6 V as in Lab 1 (pressing Recall twice may bring the setting back).
- Connect the red lead to +, green to common, and blue to − on the supply's ±25 V outputs, and to the servo's matching binding posts.
- Set up the servo to minimize friction
- Turn the gears by hand (the red gear is easiest) and make sure they turn freely.
- The usual cause of binding is the rubber connector between the gear train and the motor position sensor pot. Pushed on too far, the whitish-clear output gear rubs the metal frame; too little, it pushes the red gear against the white gear. Adjust until the gears turn smoothly.
- Turn the servo on to check that it works, then turn it off. Note what you did in the notebook.
- Set up the function generator and scope
- Put a BNC T on the function generator output. Run one BNC cable from the T to the servo's BNC input and the other to the scope.
- Connect a 10× probe to the other scope channel; it will measure the servo amplifier output. Clip the scope ground to the big green connector marked Ground on the servo board (where the supply common connects).
- Configuration 1 (used only to measure amplifier gain): sine wave, 1 kHz, amplitude set to 50 mVpp (actual output 100 mVpp). Store it: Shift, Store, turn the knob to STORE 1, Enter.
- Configuration 2 (used to move the servo): square wave (it must be square), 0.5 Hz, amplitude set to 500 mVpp (actual 1 Vpp). Store it as configuration 2.
- Record both configurations in the notebook.
- Observe how the gain knob affects the response
- With configuration 2 running, turn the gain fully counterclockwise (minimum) and turn the servo on. It should jerk back and forth once a second, commanded to about ±40° around center (the intended positions).
- Sweep the gain slowly through its whole range to see the kinds of motion.
- Return to minimum gain. Note how little and how sluggishly it moves. Does it overshoot? Does it reach the intended position?
- Raise the gain slowly. The motion gets snappier and closer to the intended position (you can hear it). Keep going until you first see or hear overshoot; just below that is the optimal gain.
- Raise the gain further until it rings (more than one overshoot). Relate what you see at each gain to overdamped, critically damped and underdamped responses.
- Disconnect the function generator, turn the gain all the way up, and set the input position knob to the center. Describe what the servo does and whether the second-order circuits from class can do that.
- Find the minimum gain for oscillation
- Turn the gain down very slowly until the oscillation stops, then back up a hair.
- Try to restart the oscillation by nudging the INPUT POSITION knob. Repeat until you are confident you have the minimum gain for oscillation.
- Measure the circuit gain
- Unplug the motor connector from the circuit board. Clip the 10× probe (scope CH2) to the + side of that connector.
- Reconnect the function generator and recall configuration 1 (Recall, turn the knob to RECALL 1, Enter).
- If CH2 doesn't show a clean sine wave, turn the red gear by hand (this moves the motor position sensor) until you see a full, unclipped sine wave.
- Check that the scope channel is set for the probe (1× or 10×). Gain = Vopp / 0.1 V = 10 × Vopp.
- Disconnect the scope and plug the motor back in.
- Find the minimum gain for ringing and for a single overshoot
- Recall configuration 2. Turn the gain down very slowly to find the minimum that still gives ringing (in either direction of motion). Measure that gain as above.
- Repeat to find the minimum gain for a single overshoot (passes the intended position once, turns back and stops). This is the better estimate of where ringing really starts, since overshoot is almost always ringing.
- Confirm with gain measurements that the range runs from about 1.7 to 133.
- Compare to calculations
- Compare your observations and gains with the appendix: G = 1.7 is overdamped and creeps to 90° in about 1 s, G = 14 rings and decays, G = 40 is a growing oscillation near 5 Hz.
- Note whether your servo oscillates that fast and whether the oscillation keeps growing. The model is linearized: it ignores supply limits and amplifier clipping (which slow the real servo and stop the growth) and fakes time delays with an artificially large motor inductance. Comment on this.
- Find the system block diagram in the appendix. The pot block is the one you measure next; you may already have the voltages from the DC or Thévenin labs.
- Find Kp, the position-sensor potentiometer constant
- The INPUT POSITION pot sits in a divider across 12 V: 11 kΩ, then the 10 kΩ pot (red, yellow wiper, black leads), then 11 kΩ.
- Connect a voltmeter from the pot's center lead (wiper) to circuit ground. Measure the voltage at both ends of the pot's rotation.
- The pot turns about 270°. Compute Kp = change in volts / change in angle, in V/deg and in V/rad.
- Conclusion
- Conclude as always: very different systems can show the same second-order behavior, which is why the transient circuit tools are worth learning.
Record
- What you did to reduce friction in the gear train
- The two stored function generator configurations
- Descriptions of the servo motion at low, optimal, overshooting and ringing gains, related to overdamped, critically damped and underdamped responses
- What the servo does at maximum gain with no input, and whether a second-order circuit could do that
- Minimum gain for sustained oscillation, minimum gain for ringing and minimum gain for a single overshoot, each with the measured amplifier output Vopp and gain = 10 × Vopp
- Comparison with the appendix predictions at G = 1.7, 14 and 40, and reasons for the differences
- Pot voltages at both extremes and Kp in V/deg and V/rad
- Conclusion
Files
- Appendix: DC permanent-magnet motor and gears: Derives the motor-and-gears transfer function θ(s)/Va(s) = KT / (s[JLa s² + (JRa + BmLa)s + (BmRa + KTKV)]) with KT = KV = 0.12, Ra = 2.704 Ω, La = 45 mH, Bm = 0.012 N m s, J = 1.161×10⁻³ N m s². Combines it with Kp = 0.717 V/rad and gain G into the closed-loop servo transfer function and its cubic characteristic equation, then solves for θ(t) at G = 1.7 (overdamped), G = 14 (decaying oscillation) and G = 40 (growing oscillation).