CHAPTER 13 / Signals and stability

Block diagrams: follow the signal, then reduce the blocks

Follow each signal through the diagram, and series, parallel and feedback connections reduce to one input–output ratio.

THE SHORT ANSWER

Block-diagram reduction replaces connected blocks with one equivalent transfer function. Blocks in series multiply, blocks in parallel add, and a negative-feedback loop with forward blocks G₁G₂ and feedback gain H reduces to G₁G₂/(1 + G₁G₂H), which follows from writing one equation per signal and solving them together. The rules hold for linear single-input, single-output blocks with zero initial state; stability must be checked separately.

YR=G1G21+G1G2H\frac{Y}{R}=\frac{G_1G_2}{1+G_1G_2H}
WHAT YOU WILL LEARN

Derive one closed-loop ratio by tracing the error signal through each block.

A single plant and gain gave one feedback formula. Block algebra lets us combine several forward blocks and a sensor without guessing the result.

Review the previous lesson: Transfer functions, poles and zeros
START WITH AN EXAMPLE

First, picture it

A block maps an input signal to an output. In this bench every block is a constant dimensionless gain, so the result is instantaneous. Set r=1, G₁=2 and G₂=3. In series the intermediate signal is 2 and the output is 6. In parallel both blocks see the same input and their outputs add to 5. Changing the wiring changes the system even when the numbers stay fixed.

THE IDEA

Build the idea step by step

Follow one signal around the loop. First subtract the sensor reading from the target to get error. The first forward block multiplies that error by its gain; the second multiplies the result again. The sensor measures the output and multiplies it by its own gain. Because output appears on both sides, solve the two signal equations together. The result is the forward gain divided by one plus the full loop gain. The worked example checks both the output and the remaining error.

Read the equations

YR=G1G21+G1G2H\frac{Y}{R}=\frac{G_1G_2}{1+G_1G_2H}

Two forward blocks multiply. The sensor gain closes a negative-feedback path in the denominator.

What the symbols mean

G1,G2G_1,G_2
forward-path block gains or compatible transfer functions
HH
sensor or feedback-path gain
RR
reference signal
YY
output signal

Read the denominator as one plus the product of every gain met once around the loop: G₁, G₂ and the sensor gain H.

BLOCK REDUCTION / three rules
SeriesRG1G2Y⇒G1·G2ParallelRG1G2++Y⇒G1 + G2Negative feedbackR+−G1G2YH⇒G1G2/(1+G1G2H)

With the worked example's static gains G₁ = 2, G₂ = 3 and H = 1, the feedback rule gives Y/R = 6/7 ≈ 0.857, and the error is E = R − HY = 1/7 ≈ 0.143 for R = 1.

WORKED EXAMPLE

Two gains and one sensor

  1. Use dimensionless static gains 2 and 3 with sensor gain 1. Set the reference to 1.

    G1=2,G2=3,H=1,R=1G_1=2,\quad G_2=3,\quad H=1,\quad R=1
  2. The forward path has gain 6. Negative feedback reduces the final output to six-sevenths.

    Y=2⋅31+2⋅3⋅1⋅1=67Y=\frac{2\cdot3}{1+2\cdot3\cdot1}\cdot1=\frac67
  3. The sensor reports six-sevenths, so the error is one-seventh. Doubling sensor gain would change both the error and output; it would not merely double the output.

    E=R−HY=1−67=17E=R-HY=1-\frac67=\frac17
More detail and model assumptions
A CLOSER LOOK

A sensor changes what the loop tracks

Increase H to 2. The loop reduces r−Hy, not necessarily r−y. At high forward gain, Hy approaches r, so y approaches r/H. A calibrated reference scale can compensate a known sensor gain. Blindly interpreting the summing-junction signal as physical tracking error is a common mistake. The bench displays both errors.

A CLOSER LOOK

Several loops: close the inner loops first

A diagram with several loops reduces the same way, one loop at a time, as long as no other signal enters or leaves an inner loop between its summing junction and its pickoff point. Take E = R − H₁X₂ − H₃Y, X₂ = G₁E, X₃ = G₂X₂ − H₂X₄, X₄ = G₃X₃ and Y = X₄ + G₄X₂, with G₁ = 1/(s+1), H₁ = 2, G₂ = 5, G₃ = 1/s, H₂ = 4, G₄ = 3 and H₃ = 1. First split the first summing junction into one that forms R − H₃Y and one after it that subtracts H₁X₂. The loop around G₁ then closes to G₁/(1+G₁H₁) = 1/(s+3), and the loop around G₃ to G₃/(1+G₃H₂) = 1/(s+4). The middle branch 5/(s+4) is in parallel with G₄ = 3, which gives (3s+17)/(s+4); in series with 1/(s+3), the forward path is (3s+17)/((s+3)(s+4)). Closing the outer loop through H₃ = 1 gives Y/R = (3s+17)/(s²+10s+29). The lecture on signal-flow graphs reaches the same answer with Mason's gain formula, without redrawing the diagram.

TRY IT

Connect the example to the lab

Predict series, parallel and feedback outputs before switching topology. Then change only H from 1 to 2. Which error does feedback reduce?

Open the concept bench
BE CAREFUL

The lab uses static gains, so this calculation has no delay or motion through time. The same signal algebra works for compatible single-input, single-output transfer functions with zero initial state. Dynamic stability is a separate check; a static gain calculation cannot establish it.

Check your understanding+

With sensor gain H=2, does feedback reduce r−y or r−2y?

It acts on r−2y, the difference at the comparison point. A target and a sensor reading must use matching scales to represent physical tracking error.
PRACTICE PROBLEMS

Solve these before opening the solutions

Work each problem on paper first. Open the answer to check yours, then the full solution if you need the steps, and mark how it went.

  1. A loop has static forward gains G1 = 4 and G2 = 5, sensor gain H = 0.5 and reference r = 2. Find the output y, the comparison-point signal r − Hy and the physical tracking error r − y. Then find y again with H = 1.

    Show answer

    Answery = 40/11 ≈ 3.64. The comparison signal is only 2/11 ≈ 0.18, yet the output is about 1.64 above the target. With H = 1, y = 40/21 ≈ 1.90.

    y=4011≈3.64,r−Hy=211≈0.18,r−y≈−1.64;H=1: y=4021≈1.90y=\frac{40}{11}\approx3.64,\qquad r-Hy=\frac{2}{11}\approx0.18,\qquad r-y\approx-1.64;\qquad H=1:\ y=\frac{40}{21}\approx1.90
    Show full solution

    Solution

    1. Write the two signal equations around the loop.

      e=r−Hy,y=G1G2 ee=r-Hy,\qquad y=G_1G_2\,e
    2. Solve them together for y.

      y=G1G21+G1G2H r=201+10×2=4011≈3.64y=\frac{G_1G_2}{1+G_1G_2H}\,r=\frac{20}{1+10}\times2=\frac{40}{11}\approx3.64
    3. The loop drives r − Hy toward zero, not r − y. With H = 0.5 the output heads toward r/H = 4 as the gain grows.

      r−Hy=2−2011=211≈0.18,r−y=2−4011≈−1.64r-Hy=2-\frac{20}{11}=\frac{2}{11}\approx0.18,\qquad r-y=2-\frac{40}{11}\approx-1.64
    4. With a unity sensor the output instead stays slightly below the target.

      H=1:y=2021×2=4021≈1.90H=1:\quad y=\frac{20}{21}\times2=\frac{40}{21}\approx1.90
  2. A proportional controller C = 5 drives the plant P(s) = 2/(s + 3), and the sensor has its own lag H(s) = 1/(0.1s + 1). Find the closed-loop transfer function Y/R, its poles and zero, and its DC gain. Compare the DC gain with that of an ideal sensor H = 1.

    Show answer

    AnswerY/R = 10(s + 10)/(s² + 13s + 130), with poles −6.5 ± j9.37 and a zero at −10, the sensor's pole. The DC gain is 10/13 ≈ 0.769, the same as with an ideal sensor.

    YR=10(s+10)s2+13s+130,s=−6.5±j9.37,z=−10,T(0)=1013≈0.769\frac YR=\frac{10(s+10)}{s^2+13s+130},\qquad s=-6.5\pm j9.37,\qquad z=-10,\qquad T(0)=\frac{10}{13}\approx0.769
    Show full solution

    Solution

    1. The negative-feedback rule holds for transfer functions as well as static gains.

      YR=CP1+CPH=10s+31+10(s+3)(0.1s+1)\frac YR=\frac{CP}{1+CPH}=\frac{\dfrac{10}{s+3}}{1+\dfrac{10}{(s+3)(0.1s+1)}}
    2. Multiply the numerator and the denominator by (s + 3)(0.1s + 1), then by 10.

      YR=10(0.1s+1)(s+3)(0.1s+1)+10=10(s+10)s2+13s+130\frac YR=\frac{10(0.1s+1)}{(s+3)(0.1s+1)+10}=\frac{10(s+10)}{s^2+13s+130}
    3. The sensor pole reappears as a closed-loop zero, and the closed loop is now second order.

      s=−13±169−5202=−6.5±j9.37s=\frac{-13\pm\sqrt{169-520}}{2}=-6.5\pm j9.37
    4. At s = 0 the sensor gain is 1, so the DC gain is unchanged. The lag alters the transient, not the steady value.

      T(0)=100130=1013≈0.769,H=1: 10s+13∣s=0=1013T(0)=\frac{100}{130}=\frac{10}{13}\approx0.769,\qquad H=1:\ \frac{10}{s+13}\Big|_{s=0}=\frac{10}{13}
  3. A position servo has an inner speed loop and an outer position loop. The inner loop applies gain K1 = 8 to the speed error around the motor model 1/(s + 2), with unity speed feedback. Speed is integrated (1/s) to give position, and the outer loop applies K2 = 5 to the position error with unity feedback. Reduce the diagram to one transfer function from position reference to position and find its poles.

    Show answer

    AnswerThe inner loop reduces to 8/(s + 10), and the whole servo to 40/(s² + 10s + 40), with poles −5 ± j3.87.

    Tin(s)=8s+10,ΘΘr=40s2+10s+40,s=−5±j3.87T_{\mathrm{in}}(s)=\frac{8}{s+10},\qquad \frac{\Theta}{\Theta_r}=\frac{40}{s^2+10s+40},\qquad s=-5\pm j3.87
    Show full solution

    Solution

    1. Reduce the inner loop first; the outer loop sees it as a single block.

      Tin=8/(s+2)1+8/(s+2)=8s+10T_{\mathrm{in}}=\frac{8/(s+2)}{1+8/(s+2)}=\frac{8}{s+10}
    2. The outer forward path is K2, the inner block and the integrator in series.

      Lout=5⋅8s+10⋅1s=40s(s+10)L_{\mathrm{out}}=5\cdot\frac{8}{s+10}\cdot\frac1s=\frac{40}{s(s+10)}
    3. Close the outer loop.

      ΘΘr=40s(s+10)+40=40s2+10s+40\frac{\Theta}{\Theta_r}=\frac{40}{s(s+10)+40}=\frac{40}{s^2+10s+40}
    4. Solve the characteristic equation. The inner loop had already moved the slow motor pole from −2 to −10 before the outer loop was closed.

      s=−10±100−1602=−5±j3.87s=\frac{-10\pm\sqrt{100-160}}{2}=-5\pm j3.87
  4. A controller K = 18 drives the plant G2(s) = 1/(s + 2) under unity negative feedback. A constant disturbance d = 0.5 adds to the plant input, and the reference is a unit step. (a) Write Y(s) in terms of R(s) and D(s). (b) Find the steady output and the steady error. (c) An open-loop alternative applies the constant input 2, which gives y = 1 without the disturbance. How large is the disturbance's effect there, and by what factor does feedback reduce it?

    Show answer

    Answery(∞) = 0.9 + 0.025 = 0.925, so the steady error is 0.075. Open loop, the disturbance shifts the output by 0.25; feedback cuts that to 0.025, a factor of 10.

    Y=KG21+KG2R+G21+KG2D,y(∞)=0.925,e(∞)=0.075Y=\frac{KG_2}{1+KG_2}R+\frac{G_2}{1+KG_2}D,\qquad y(\infty)=0.925,\qquad e(\infty)=0.075
    Show full solution

    Solution

    1. Follow each input to the output and superpose. Both paths share the denominator 1 + KG2.

      Y(s)=18s+20 R(s)+1s+20 D(s)Y(s)=\frac{18}{s+20}\,R(s)+\frac{1}{s+20}\,D(s)
    2. The closed-loop pole −20 is stable, so each constant input contributes its path's DC gain times its size.

      yr(∞)=1820=0.9,yd(∞)=0.520=0.025y_r(\infty)=\frac{18}{20}=0.9,\qquad y_d(\infty)=\frac{0.5}{20}=0.025
    3. Add the two contributions.

      y(∞)=0.925,e(∞)=1−0.925=0.075y(\infty)=0.925,\qquad e(\infty)=1-0.925=0.075
    4. Open loop, the disturbance passes through G2 alone. Feedback divides that path by 1 + KG2(0) = 10.

      ΔyOL=G2(0) d=0.25,ΔyCL=0.251+9=0.025\Delta y_{\mathrm{OL}}=G_2(0)\,d=0.25,\qquad \Delta y_{\mathrm{CL}}=\frac{0.25}{1+9}=0.025

Mark each exercise after you check your answer. Marks stay in this browser.

KEY TERMS

Terms this lecture teaches

All terms in the glossary →
QUESTIONS STUDENTS ASK

Frequently asked questions

What is the difference between blocks in series and in parallel?

In series the output of the first block is the input of the second, so the gains multiply: 2 and 3 give 6. In parallel both blocks receive the same input and their outputs are added, so the same two gains give 5. The numbers are identical, but the wiring makes a different system.

Why is the feedback denominator 1 + G₁G₂H and not 1 − G₁G₂H?

Because the measured signal is subtracted at the summing junction. Writing E = R − HY and Y = G₁G₂E and solving for Y gives Y/R = G₁G₂/(1 + G₁G₂H). If the feedback signal were added instead, as in positive feedback, the denominator would become 1 − G₁G₂H.

What happens if the sensor gain H is 2 instead of 1?

The loop drives the summing-junction signal R − HY toward zero, not R − Y. With forward gain 6 and R = 1, the output drops from 6/7 ≈ 0.86 to 6/13 ≈ 0.46, and at very high forward gain it approaches R/H = 0.5. Scale the reference to match the sensor if the physical output should track R.

How do I reduce a block diagram with several loops?

Write one equation for each summing junction and each block output, then eliminate the intermediate signals until only the input and output remain. Rules such as moving a pickoff point or collapsing an inner loop first are shortcuts for the same algebra, so you can always check a reduction by solving the signal equations directly.

Does a reduced block diagram show whether the loop is stable?

Not by itself. The reduction gives the input-output ratio; for dynamic blocks you still have to check that every root of 1 + G₁G₂H = 0 lies in the left half-plane. A pole-zero cancellation made during the reduction can also hide an internal mode, and static gains carry no information about motion over time.

Why the next lesson follows

Series, parallel and feedback rules close one loop at a time. When several loops interlock, a signal-flow graph and Mason's gain formula give the whole transfer function in one pass.

YOUR PROGRESS

Finished the reading and the exercises? Mark the lecture as done to tick it on the syllabus. Progress stays in this browser.

EIGENROOM · https://eigenroom.com/theory/block-algebra · Updated 2026-11-15