Root Locus Compensator Design: A Jigsaw Activity

Tania Demonte Gonzalez, Michigan Technological University (tsdemont@mtu.edu)

Author Profile
Initial Publication Date: October 6, 2026
DOI | Cite this

Summary

This is an in-class activity where students use MATLAB and root-locus design to develop three compensators for a high-speed electromechanical positioning system, then participate in a jigsaw activity to compare the designs from both a controls and client-value perspective.

The three strategies are: (A) one lead compensator, (B) two cascaded lead compensators, and (C) a custom design using pole cancellation and lag compensation. Each expert group develops and verifies one controller in a MATLAB Live Script. Students then reorganize into mixed groups to share results, compare transient-response performance and implementation tradeoffs, and recommend the design that creates the most value for the client.

Topics: Feedback control, root locus, lead compensation, lag compensation, pole-zero cancellation, transient response, engineering decision-making

Share your modifications and improvements to this activity through the Community Contribution Tool »

Learning Goals

By the end of the activity, students should be able to:

  1. Translate client transient-response requirements into approximate second-order design targets and a desired dominant pole location.
  2. Use root-locus angle and magnitude conditions to design lead/lag compensators that target the desired pole region.
  3. Use MATLAB and Control System Toolbox to construct transfer functions, evaluate root-locus conditions, form closed-loop models, locate poles, and measure step-response characteristics.
  4. Compare MATLAB-measured overshoot and settling time with second-order predictions and explain important discrepancies.
  5. Evaluate transient performance, controller complexity, implementation burden, and robustness, then justify a client recommendation using engineering evidence.
  6. Reflect on how the jigsaw comparison and MATLAB evidence changed or strengthened their understanding of controller-design tradeoffs.

Context for Use

This activity is intended for a feedback controls course after students have learned the root-locus angle and magnitude conditions and have been introduced to lead and lag compensation. Students should be comfortable using MATLAB transfer functions and rlocus, but the Live Script provides the computational structure needed for the activity.
The activity can replace a traditional homework assignment or be used as an in-class application after compensator design has been introduced. A 50–60 minute class works well if students have already seen the design procedure. A 75-90 minute class allows additional time for interpretation, robustness discussion, and short group presentations.
Recommended group structure:
- Individual prediction: 5–10 min
- Expert groups: 20–25 min
- Jigsaw groups: 15–20 min
- Client recommendation/share-out: 5–10 min

Description and Teaching Materials

The student Live Script guides students through five phases.
1. Individual prediction: Students locate the desired pole from the damping-ratio and natural-frequency requirements, inspect the uncompensated root locus, and estimate the phase deficiency/excess at the target location.
2. Expert-group design: Each student is assigned to one expert group. The group designs only one of the three controller architectures and uses MATLAB to determine compensator parameters, calculate the gain, and verify the closed-loop response.
3. Implementation analysis: Students identify the order and fastest pole of their controller. The shortest controller time constant is used as a simple proxy for how demanding the design may be to simulate or implement digitally. This is intentionally presented as an implementation indicator rather than a complete processor-load model.
4. Jigsaw comparison: Students reorganize into mixed groups containing experts from Designs A, B, and C. Each expert teaches the other students how the assigned controller works and presents the MATLAB evidence. The group builds a common comparison table containing:
- compensator structure/order,
- dominant closed-loop poles,
- rise time,
- 2% settling time,
- percent overshoot,
- fastest controller pole magnitude,
- shortest controller time constant,
- implementation concerns, and
- robustness concerns.
5. Client-value recommendation: The mixed group recommends a design for the client. The client values meeting the transient-response target, fast response with limited overshoot, practical implementation, and reasonable robustness to modeling error. No numerical weighting is prescribed, so students must state their assumptions and explain the gains and sacrifices associated with each design.
Root_Locus_Jigsaw_Student.mlx (MATLAB Live Script 9kB Oct27 26)

Root_Locus_Jigsaw_Instructor.mlx (MATLAB Live Script 712kB Oct27 26)

Teaching Notes and Tips

Teaching Notes and Tips

Phase wrapping

A common MATLAB issue is interpreting the phase returned by angle(evalfr(...)). MATLAB reports a principal angle between -180deg and 180deg, whereas the root-locus angle condition is interpreted modulo 360deg and must satisfy an odd multiple of 180deg.

Students may therefore obtain an angle that appears different from a hand calculation by 360deg. Encourage them to calculate the phase deficiency relative to an odd multiple of 180deg and verify the sign with an angle sketch rather than relying only on the numerical MATLAB result.

Design C and properness

Design C is particularly useful for highlighting that controller implementation details can affect the root-locus design itself. Before a high-frequency pole is added, a cancellation-plus-lag design can satisfy the graphical angle condition. Once an additional fast pole is included to make the controller proper, that pole contributes phase at the desired design point. The lag portion must therefore be re-checked and, if necessary, retuned.

This provides an opportunity to emphasize that a controller should be evaluated in its final implementable form rather than only as an intermediate graphical design.

Robustness extension

For Design C, keep the controller fixed and perturb the nominally canceled plant pole at \(-15\) by approximately ±10%. Overlay the nominal and perturbed step responses.

Ask students:

  • Why does exact cancellation in the model not remain exact in the physical system?
  • How does this affect the closed-loop response?
  • Does this change the value proposition of Design C relative to the other controllers?

This extension makes the robustness cost of pole-zero cancellation visible without requiring a full robust-control treatment.

Preventing "fastest is best" reasoning

Students often focus immediately on rise time or settling time. Before the mixed-group recommendation, remind them that client value includes more than a single transient-response metric. Controller complexity, fast internal dynamics, sensitivity to model error, and practical implementation all matter.

Require explanation, not just plots

The jigsaw is most effective when experts must teach the controller logic, not merely report numbers. A useful facilitation rule is to have each receiving student ask at least one question before the group records that controller in the comparison table.


Assessment

The assessment is aligned directly to the learning outcomes and includes both group and individual evidence of learning.

Criterion Learning outcomes assessed Exemplary Proficient Developing Beginning Weight
Client requirements and root-locus reasoning 1, 2 Correctly translates the client requirements into an approximate pole target, applies the angle and magnitude conditions correctly, and clearly justifies the compensator design. Minor calculation or explanation issue, but the design method and result are substantially correct. Partial reasoning is shown, but target calculations or compensator choices require significant correction. Little evidence of correct transient-response or root-locus reasoning. 25%
MATLAB construction and verification 3 Correctly constructs transfer functions, evaluates the design point, computes the gain, forms the closed-loop system, verifies poles, and presents readable response plots/results. MATLAB workflow is mostly correct with minor errors or presentation issues. Several MATLAB steps are incomplete, incorrect, or insufficiently verified. MATLAB evidence is missing or does not support the claimed design. 20%
Prediction and response interpretation 4 Correctly connects the second-order design targets to the measured complete-system response and explains important discrepancies using poles, zeros, and dominance assumptions. Makes a correct comparison with a reasonable but incomplete explanation. Reports results but gives limited interpretation of differences. Comparison or interpretation is missing or conceptually incorrect. 15%
Jigsaw comparison and tradeoff analysis 5 Uses quantitative evidence to compare all three designs and gives technically sound discussion of performance, complexity, implementation, and robustness. Identifies the major tradeoffs with adequate quantitative support. Discusses some tradeoffs but with limited evidence or conceptual gaps. Treats one metric as sufficient or omits major implementation/robustness considerations. 15%
Client-value recommendation 5 Recommendation is clearly supported by numerical evidence, states priorities, and explains what the client gains and gives up across the three designs. Recommendation is supported and acknowledges multiple tradeoffs. Recommendation is plausible but weakly supported or overlooks a major tradeoff. Recommendation is unsupported or based primarily on a single metric. 15%
Individual reflection 6 Clearly identifies new learning from the share-out, connects a MATLAB result to a change or confirmation in understanding, and proposes a meaningful next analysis tied to a decision. Addresses all prompts with reasonable reflection. Reflection is mostly descriptive or addresses only part of the prompts. Reflection is missing or does not demonstrate individual learning. 10%

Total: 100%

The rubric may be simplified to a four-point formative scale when the activity is used primarily for in-class practice.

References and Resources

Root_Locus_Jigsaw_Instructor.mlx (MATLAB Live Script 712kB Oct27 26)

Root_Locus_Jigsaw_Student.mlx (MATLAB Live Script 9kB Oct27 26)