A signal flow graph uses nodes for signals and branches with gains for the relations between them. It gives the transfer function in one step, without repeated reduction.
Terms
- Forward path: a path from input node to output node that touches no node twice.
- Loop: a closed path that returns to its starting node.
- Non-touching loops: loops that share no node.
Mason's gain formula
$$T = \frac{1}{\Delta}\sum_k P_k\Delta_k$$
where $P_k$ is the gain of the $k$-th forward path and
$$\Delta = 1 - \sum L_i + \sum L_iL_j - \sum L_iL_jL_k + \dots$$
($\sum L_i$: all loop gains; $\sum L_iL_j$: products of all pairs of non-touching loops; and so on). $\Delta_k$ is $\Delta$ computed with the loops that touch the $k$-th forward path removed.
Example
For $G_1$ in series with an inner loop $G_2, H_2$, and outer feedback $H_1$: one forward path $P_1 = G_1G_2$; loops $L_1 = -G_2H_2$, $L_2 = -G_1G_2H_1$, which touch. So $\Delta = 1 + G_2H_2 + G_1G_2H_1$ and $\Delta_1 = 1$, matching Lesson 5.