Block diagram reduction and signal flow graphs

Reduces block diagrams with the series, parallel, feedback and point-moving rules, and finds transfer functions from signal flow graphs with Mason's gain formula.

Drafted with Aria, reviewed by the AiCanCode.org team. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

A real instrument loop is drawn as many blocks — controller, actuator, process, transmitter, sometimes an inner cascade loop. To analyse stability or error you need one overall transfer function. Block-diagram reduction and Mason's gain formula are the two standard ways to get it quickly and without algebra mistakes, and GATE tests both every year in some form.

Key ideas

Block diagram elements

  • Block: multiplies its input signal by its transfer function.
  • Summing point: adds or subtracts signals; the sign on each arrow matters.
  • Take-off (pick-off) point: copies a signal to another path without loading it.

Reduction rules

  1. Blocks in series (cascade): multiply, G1·G2 — valid only if the second block does not load the first.
  2. Blocks in parallel (feeding the same summing point): add, G1 ± G2.
  3. A feedback loop: G/(1 ± GH) (plus sign in the denominator for negative feedback).
  4. Moving a summing point ahead of a block G: put 1/G in the moved branch. Moving it after the block: put G in the moved branch.
  5. Moving a take-off point ahead of a block G: put G in the moved branch. Moving it after the block: put 1/G in the moved branch.
  6. Adjacent summing points can be interchanged; adjacent take-off points can be interchanged. A summing point and a take-off point next to each other cannot be swapped without adding a compensating branch. Reduce the innermost loop first, and move points only to un-nest interlocked loops.

Signal flow graph (SFG)

  • Node: a system variable. Branch: a directed line with a gain (transmittance) from one node to another. A node's value is the sum of all incoming branch signals.
  • Source node has only outgoing branches; sink node has only incoming branches.
  • Forward path: from source to sink, visiting no node more than once.
  • Loop: a closed path that starts and ends on the same node without passing any other node twice.
  • Non-touching loops: loops that share no node.
  • A block diagram converts to an SFG by making every summing point and take-off point a node; the gains of SFG branches carry the signs that the summing points had.

Mason's gain formula gives the transfer function directly from the graph. It is exact for linear graphs and usually faster than reduction when there are several forward paths or interlocked loops.

Formulas

  • T = C(s)/R(s) = (Σ Pk·Δk) / Δ — Mason's gain formula.
  • Δ = 1 − ΣLi + ΣLiLj − ΣLiLjLk + … — graph determinant: 1 − (sum of all individual loop gains) + (sum of gain products of every pair of non-touching loops) − (sum for every triple of mutually non-touching loops) + …
  • Δk = Δ evaluated after removing every loop that touches the k-th forward path.
  • Pk = gain of the k-th forward path.
  • T = G/(1 + GH) — single negative-feedback loop (loop gain in SFG form L = −GH, so Δ = 1 + GH).
  • Series: G1·G2; Parallel: G1 + G2.

All gains are transfer functions in s (or constants); T, Pk, Li carry the units of output/input of the path they describe, and Δ, Δk are dimensionless. The loop gain Li includes the sign of the feedback branch: for a negative-feedback loop, Li is negative.

Worked examples

Example 1 — minor loop, by reduction and by Mason Given: G1 = 10 in series with G2 = 5; a negative-feedback branch H = 2 from the output of G2 back to a summing point just before G2. No outer loop.

  1. Reduction: inner loop G2/(1 + G2H) = 5/(1 + 10) = 5/11.
  2. Series with G1: T = 10 × 5/11 = 50/11.
  3. Mason: one forward path P1 = G1G2 = 50; one loop L1 = −G2H = −10 (the minus sign comes from the negative summing point).
  4. Δ = 1 − L1 = 1 − (−10) = 11; the loop touches the path, so Δ1 = 1.
  5. T = P1Δ1/Δ = 50/11.

Result: T = 50/11 = 4.55. A common wrong answer is 50/(1 − 10) = −5.56, which comes from forgetting the negative sign of the loop gain.

Example 2 — two forward paths and non-touching loops (GATE level) Given an SFG with nodes R, x1, x2, x3, C and branches: R→x1 gain 1, x1→x2 gain G1 = 2, x2→x3 gain G2 = 3, x3→C gain G3 = 4, x2→x1 gain −H1 = −0.5, C→x3 gain −H2 = −0.25, and a feed-forward branch R→x3 gain G4 = 1. Find C/R.

  1. Loops: L1 = G1·(−H1) = −1 (nodes x1, x2); L2 = G3·(−H2) = −1 (nodes x3, C). They share no node, so they are non-touching.
  2. Δ = 1 − (L1 + L2) + L1L2 = 1 − (−2) + 1 = 4.
  3. Forward path 1: R→x1→x2→x3→C, P1 = G1G2G3 = 24. It touches both loops, so Δ1 = 1.
  4. Forward path 2: R→x3→C, P2 = G4G3 = 4. It touches L2 but not L1, so Δ2 = 1 − L1 = 2.
  5. T = (P1Δ1 + P2Δ2)/Δ = (24 × 1 + 4 × 2)/4 = 32/4.

Result: C/R = 8. (Solving the four node equations simultaneously gives the same value, which is a good way to check Mason answers.)

Common mistakes

  • Dropping the minus sign of a negative-feedback branch when computing loop gains in Mason's formula.
  • Counting two loops as non-touching when they share even one node.
  • Using Δk = 1 automatically; check which loops the k-th path leaves untouched.
  • Treating a branch that only enters the source node as a loop, or forgetting that the source node must have only outgoing branches (add a unity branch if needed).
  • Multiplying blocks in series when the second loads the first (e.g. two RC sections connected directly).
  • Moving a take-off point the wrong way and putting G where 1/G belongs.

For GATE IN

Expect questions that give a block diagram or SFG and ask for the overall transfer function, the number of forward paths or non-touching loop pairs, or the value of a gain that makes the output independent of a disturbance. Practise converting block diagrams into SFGs, listing loops systematically, and checking Mason results by writing node equations.

Quick check

  1. Two blocks 4/s and s/(s + 2) are in cascade. What is the equivalent?
  2. An SFG has a single loop of gain −0.2 touching the only forward path of gain 6. What is T?
  3. Two loops share exactly one node. Are they non-touching?
  4. When a summing point is moved from the input of block G to its output, what gain goes in the moved branch?

Answers: 1. 4/(s + 2); 2. 6/1.2 = 5; 3. No; 4. G.

Try answering each one aloud before you open it.

  1. 1.What is a block diagram in control systems?Concept

    A block diagram is a graphical representation of a control system, showing the system's components and their interconnections. Each block represents a system component or process, and arrows indicate the direction of signal flow. Block diagrams help in visualizing the functional relationships between different parts of a system.

  2. 2.Explain the purpose of signal flow graphs in control systems.Concept

    Signal flow graphs are used to represent the flow of signals in a control system. They consist of nodes and directed branches, where nodes represent system variables and branches represent functional relationships between these variables. Signal flow graphs are useful for analyzing complex systems and deriving transfer functions using Mason's Gain Formula.

  3. 3.How do you reduce a block diagram to its simplest form?Concept

    Combine cascaded blocks by multiplying and parallel blocks by adding, then replace each feedback loop by G/(1 ± GH), working from the innermost loop outward. When loops are interlocked, first move summing or take-off points across blocks, inserting G or 1/G in the moved branch so every signal stays the same. Repeat until one block remains; for diagrams with many paths, Mason's gain formula on the equivalent signal flow graph is usually quicker.

  4. 4.Why is block diagram reduction important in control systems?Application

    Block diagram reduction is important because it simplifies complex control systems into a single transfer function, making analysis and design easier. It helps in understanding the overall system behavior and facilitates the design of controllers and compensators by providing a clear view of the system dynamics.

  5. 5.How does Mason's Gain Formula help in analyzing signal flow graphs?Application

    Mason's Gain Formula provides a systematic way to calculate the overall transfer function of a system represented by a signal flow graph. It considers all forward paths, loops, and their interactions. The formula is T = Σ(P_kΔ_k)/Δ, where P_k is the gain of the k-th forward path, Δ is the determinant of the graph, and Δ_k is the cofactor of the k-th path.

  6. 6.What is the difference between a block diagram and a signal flow graph?Concept

    A block diagram uses blocks and arrows to represent system components and signal flow, focusing on the functional relationships. A signal flow graph uses nodes and directed branches to represent variables and their interactions, emphasizing the mathematical relationships. Signal flow graphs are often more suitable for deriving transfer functions using Mason's Gain Formula.

  7. 7.Why might an engineer choose to use a signal flow graph over a block diagram?Application

    An engineer might choose a signal flow graph over a block diagram when they need to analyze complex systems with multiple feedback loops and interactions. Signal flow graphs provide a clearer mathematical representation and facilitate the use of Mason's Gain Formula to derive the system's transfer function efficiently.

  8. 8.Given a block diagram with two blocks in series with transfer functions G1(s) = 2/s and G2(s) = 3/(s+1), what is the equivalent transfer function?Numerical

    For blocks in series, the equivalent transfer function is the product of the individual transfer functions. Therefore, the equivalent transfer function is G_eq(s) = G1(s) * G2(s) = (2/s) * (3/(s+1)) = 6/(s(s+1)).

  9. 9.A signal flow graph has two forward paths with gains 5 and 3 and a single loop of gain −0.5 that touches both paths. What is the overall transfer function?Numerical

    Δ = 1 − L = 1 − (−0.5) = 1.5. The loop touches both forward paths, so Δ1 = Δ2 = 1. By Mason's formula T = (5 × 1 + 3 × 1)/1.5 = 8/1.5 = 5.33. If the loop did not touch a path, that path's Δk would be 1.5 instead of 1.

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?