EIGENROOMLAB WORKSHEET 17ANSWER KEY · FOR INSTRUCTORSDC motor · Design challenges
Lead, lag or PI?
Goal
Compare PI, lead and lag on the same motor and load against several requirements at once.
Setup
Lab eigenroom.com/lab?lesson=lead-lag-pi · short link eigenroom.com/lead-lag-pi
Open it on a phone or a computer; there is nothing to log in to or install. The link loads run A's settings:
- Target speed
- 90 rad/s
- Loop
- Closed loop
- Controller type
- PID
- Proportional Kp
- 0.22
- Integral Ki
- 0.45
- Derivative Kd
- 0
- Anti-windup
- On
- Load torque
- 0.08 N·m
- Sensor noise
- 0
- Controller period
- 20 ms
- Disturbance events
- None
- Recording
- 12 s
- APreset, exactly as the link opens it.
- BChange only: Controller type → Lead compensator, Controller gain → 2, Zero frequency → 4 rad/s, Pole frequency → 8 rad/s. Select Lead in Controller type; the lab loads its default gain, zero and pole.
- CChange only: Controller type → Lag compensator, Controller gain → 0.25, Zero frequency → 8 rad/s, Pole frequency → 0.5 rad/s. Select Lag in Controller type.
- DChange only: Integral Ki → 0.9. (ext.)
1. Predict
Which controller best meets an error, speed and effort requirement under the same load?
PI can remove persistent error but can wind up; lead adds phase near a band; lag improves relative low-frequency gain without guaranteeing zero error. The design objective and actuator limit decide the tradeoff.
2. Observe
| Quantity | Where to read it | A | B | C | D ext. |
|---|---|---|---|---|---|
| Final error (rad/s) | Measured response | 0.00 rad/s | 8.88 rad/s | 2.33 rad/s | 0.00 rad/s |
| Settling, ±2% of target (s) | Measured response | 1.64 s | Not settled | Not settled | 2.20 s |
| Overshoot vs target (%) | Measured response | 0.96% | 0.00% | 25.22% | 3.49% |
| Control effort (V·s) | Measured response | 117.40 V·s | 110.02 V·s | 117.03 V·s | 117.48 V·s |
Values computed by DC motor and cart-pole engine 1.5.0 at the settings above. A replay readout can differ in the last digit if the slider is not exactly at the stated time; values read from a plot are approximate.
3. Explain
Only one run ends with zero error. Which one, and what in its structure guarantees it?
PI (run A, final error 0.00 rad/s): its integrator has unbounded gain at zero frequency. Lead and lag have finite low-frequency gain and leave 8.88 rad/s and 2.33 rad/s.
Lag leaves a smaller error than lead but overshoots far more. What did it trade for its low-frequency gain?
Phase near crossover: the lag overshoots 25.22%, the lead 0.00%. Low-frequency gain from a lag network comes with phase lag where the loop crosses over.
Rank the runs by control effort. Is the run with the least effort the best choice?
Lead uses the least (110.02 V·s), then lag (117.03 V·s) and PI (117.40 V·s). But lead leaves the largest error. The best run depends on which requirement matters most: zero error, little overshoot or low effort.
Extension (optional)
Raise the PI's Ki to 0.9 (run D). Does PI still overshoot less than the lag?
Yes, but it gets worse than run A: run D overshoots 3.49% and settles at 2.20 s, against 0.96% and 1.64 s, while the lag overshoots 25.22%. A stronger integrator is not automatically a faster loop.