Skip to main content
  • Pricing
Tutorials

Mason's gain formula: from signal flow graph to transfer function, step by step

By Oakgen TeamPublished 3 min read

Share
Mason's gain formula: from signal flow graph to transfer function, step by step

Mason's gain formula turns a signal flow graph into an overall transfer function in one pass: add up every forward path from input to output, weight each by a cofactor that removes the loops it touches, and divide by a determinant built from all the loops. It replaces repeated block diagram reduction with a checklist. This guide states the formula, the definitions it needs, a procedure, and two worked examples.

The formula and its terms

The KFUPM EE 380 notes and the MSOE EE371 technical note give the same statement:

T(s) = (sum over k of Pk x delta_k) / delta

  • Node: a junction that denotes a variable or a signal. Branch: a one-way path between nodes carrying a gain.
  • Forward path: a continuous path from input to output that meets no node twice. Pk is its gain, the product of the branch gains along it.
  • Loop: a closed path that starts and ends at the same node and meets no node twice. Its gain is the product of the branch gains around it.
  • Non-touching: loops with no common nodes. The Benha course file words this as no common nodes and/or branches, so check which definition your course uses.
  • delta: 1 - (sum of all loop gains) + (sum of products of every pair of non-touching loops) - (sum of products of every non-touching triple) + ...
  • delta_k: delta recomputed for the part of the graph that does not touch the kth forward path.

The KFUPM notes credit the method to Mason. MIT's repository also holds a 1953 Research Laboratory of Electronics technical report titled "Feedback theory", subtitled "Some properties of signal flow graphs", which was also issued as an MIT thesis in 1951; the record page does not name an author.

Turning a block diagram into a graph

The MSOE note gives a short recipe: add an input source node and an output sink node, add a node for each summing junction and each pickoff point, then connect them with branches. A block between two nodes supplies the branch gain, and a plain wire is a branch of gain 1. A minus sign at a summing junction becomes a negative branch weight. In the MSOE example the unity feedback path is drawn with weight -1.

The five-step procedure

The MSOE note lists these steps:

  1. List all forward paths.
  2. List all feedback loops.
  3. List which loops do not touch each other.
  4. List which loops touch each path.
  5. Compute delta and every delta_k, then evaluate the formula.

Worked example 1: two paths, four loops

The KFUPM notes include a graph with two forward paths, P1 = G1G2G3G4 and P2 = G5G6G7G8, and four loops, L1 = G2H2, L2 = G3H3, L3 = G6H6 and L4 = G7H7. Loops L1 and L2 do not touch L3 or L4, and no other loops are non-touching. So:

  • delta = 1 - (L1 + L2 + L3 + L4) + (L1L3 + L1L4 + L2L3 + L2L4)
  • delta_1 = 1 - (L3 + L4), because L1 and L2 touch the first path
  • delta_2 = 1 - (L1 + L2), because L3 and L4 touch the second path
  • T = (P1 delta_1 + P2 delta_2) / delta

A numeric check by Oakgen, using only that structure: let both forward paths have gain 1 and give each loop a gain of 0.2. Then delta = 1 - 0.8 + 4(0.04) = 0.36, and delta_1 = delta_2 = 1 - 0.4 = 0.6. T = (0.6 + 0.6) / 0.36 = 3.33. If you instead forgot the non-touching products, delta would be 0.2 and T would be 6, a useful sign that the pair terms matter.

Worked example 2: one loop that touches everything

The KFUPM notes also treat an armature-controlled DC motor. There is one forward path, P1 = (1/s)(Km/(Ra + La s))(1/(Js + b)), and one loop, L1 = -Kb(Km/(Ra + La s))(1/(Js + b)). With a single loop that touches the only path, delta_1 = 1, and the result is theta(s)/Va(s) = Km / (s[(Ra + La s)(Js + b) + Kb Km]), matching the answer the notes obtained earlier by block diagram reduction.

Common mistakes

  • Counting a path or loop that visits a node twice.
  • Forgetting a loop sign: negative feedback contributes a negative loop gain, so it adds to delta.
  • Treating all loop pairs as non-touching. Check node by node.
  • Using the full delta for delta_k instead of removing the loops that touch path k.

As a sanity check, a single forward gain G with one negative feedback loop of gain -GH gives delta = 1 + GH, delta_1 = 1 and T = G / (1 + GH), the standard closed-loop result.

Take this further with Oakgen Agent

Agent can explain the method, write a quiz with solutions, run numeric checks in a Python sandbox, and create a downloadable pdf or docx study sheet.

Coming soon: Drawing and auto-solving a signal flow graph from an uploaded diagram image is not available in Agent yet.

Looking for a recap of one specific YouTube video? The YouTube Summarizer does that today.

Open this task in Oakgen Agent

The prompt is filled in for you. Nothing is sent until you press send.

Open in Oakgen Agent

Frequently asked questions

What is Mason's gain formula?

It gives the overall gain of a signal flow graph as T = (sum over k of Pk x delta_k) / delta, where Pk is the gain of the kth forward path from input to output, delta is the graph determinant and delta_k is the cofactor of the kth path. KFUPM's EE 380 notes and the MSOE EE371 technical note both state it in this form.

How is delta calculated?

Delta = 1 minus the sum of all individual loop gains, plus the sum of gain products of every pair of non-touching loops, minus the sum of products of every non-touching triple, and so on with alternating signs. The MSOE note and the KFUPM notes give this definition.

What is delta_k?

It is the value of delta for the part of the graph that does not touch the kth forward path. In practice you compute delta again after discarding every loop that shares a node with that path. If every loop touches the path, delta_k equals 1.

When is Mason's formula better than block diagram reduction?

The KFUPM notes say block diagram reduction becomes cumbersome for a complex system and that the signal flow graph route follows a straightforward procedure without moving pickoff points or redrawing the diagram. The MSOE note makes the same point: it is especially useful when a block diagram would need many reduction steps.

Sources

  1. King Fahd University of Petroleum and Minerals, EE 380 class notes, Lecture 8: Signal flow graph model and Mason's rule (read 2026-11-19)
  2. Milwaukee School of Engineering, EE371 Technical Note: Mason's Gain Formula (read 2026-11-19)
  3. Benha University, Dr. Ahmed M. Hussein, Chapter 4: Signal Flow Graph course file (read 2026-11-19)
  4. MIT DSpace, Feedback theory (MIT Research Laboratory of Electronics Technical Report 153, 1953) record (read 2026-11-19)
Share