Electrical Engineering › Section E · Methods of analysis (dc) › Topic E.2

Branch-current analysis

Put two sources in different branches and series-parallel reduction no longer works. Branch-current analysis is the most direct systematic cure: give every branch an unknown current, write KCL and KVL, and solve the equations together.

When reduction stops working

Series-parallel reduction needs a single source, or sources that can be combined, because it works back from one total current. Place a second source in another branch, as when two batteries share a load or a charger feeds a battery and a lamp at once, and no element is simply in series or in parallel with the rest. You need a method that works for any network. Branch-current analysis is the most direct one: it uses nothing beyond Ohm's law, KCL and KVL, applied in a fixed order. The later methods, "Mesh analysis" and "Nodal analysis", are refinements that need fewer equations, but this one shows most plainly where every equation comes from.

Branches, nodes and loops

A branch is a path between two nodes that carries one current, such as a source in series with a resistor. A node, as before, is a point where branches meet. A network with \(b\) branches has \(b\) unknown branch currents, so it needs \(b\) independent equations. KCL supplies \(n - 1\) of them in a network with \(n\) nodes: the last node's equation is always a combination of the others and adds nothing new. KVL supplies the rest, \(b - (n - 1)\) equations, from independent loops. The easy way to choose independent loops is to take the window panes of the drawing, each loop enclosing no other.

The procedure

Step 1: assign a current, with a direction, to every branch. The direction is a guess; it does not need to be right. Step 2: mark the polarity of each resistor's voltage from its assumed current, + where the current enters. Source polarities are fixed by the sources themselves. Step 3: write KCL at \(n - 1\) nodes. Step 4: write KVL around each independent loop, using \(V = IR\) for every resistor. Step 5: solve the equations together. A negative answer means the current actually flows opposite to the guess; its size is correct, and it is best left negative in later working so that every equation stays consistent.

Worked example: two sources sharing a load

Node a is at the top and ground at the bottom. Branch 1: \(E_1 = 12\,\text{V}\), − at ground, in series with \(R_1 = 2.2\,\text{k}\Omega\) up to a. Branch 2: \(E_2 = 9\,\text{V}\), − at ground, with \(R_2 = 1.0\,\text{k}\Omega\) up to a. Branch 3: \(R_3 = 4.7\,\text{k}\Omega\) from a to ground. Working in kΩ, mA and V: 1. Guess \(I_1\) and \(I_2\) up into a and \(I_3\) down through \(R_3\). 2. KCL at a: \(I_1 + I_2 = I_3\). 3. KVL, left pane: \(12 - 2.2I_1 - 4.7I_3 = 0\). 4. Right pane: \(9 - 1.0I_2 - 4.7I_3 = 0\). 5. Substituting \(I_3\): \(6.9I_1 + 4.7I_2 = 12\) and \(4.7I_1 + 5.7I_2 = 9\). 6. Solving: \(I_1 = 1.51\,\text{mA}\), \(I_2 = 0.331\,\text{mA}\) and \(I_3 = 1.84\,\text{mA}\), so \(V_a = 8.67\,\text{V}\).

Solving with determinants

Two equations in the standard form \(a_1x + b_1y = c_1\) and \(a_2x + b_2y = c_2\) can be solved quickly by Cramer's rule:

x = \frac{c_1b_2 - b_1c_2}{D}, \quad y = \frac{a_1c_2 - c_1a_2}{D}

where \(D = a_1b_2 - b_1a_2\). In the example, \(D = 6.9(5.7) - 4.7(4.7) = 17.24\), so \(I_1 = \frac{12(5.7) - 4.7(9)}{17.24} = 1.51\,\text{mA}\) and \(I_2 = \frac{6.9(9) - 12(4.7)}{17.24} = 0.331\,\text{mA}\). Three or more equations can be solved by elimination, by 3 × 3 determinants or by the equation solver on a calculator or computer. Whatever the tool, write the equations in a tidy standard form first.

Worked example: a current that comes out negative

Use the same layout with \(E_1 = 10\,\text{V}\), \(R_1 = 1.0\,\text{k}\Omega\), \(E_2 = 5\,\text{V}\), \(R_2 = 2.0\,\text{k}\Omega\) and \(R_3 = 3.3\,\text{k}\Omega\), with the same guessed directions. 1. KCL: \(I_3 = I_1 + I_2\). 2. KVL gives \(4.3I_1 + 3.3I_2 = 10\) and \(3.3I_1 + 5.3I_2 = 5\). 3. \(D = 4.3(5.3) - 3.3(3.3) = 11.90\). 4. \(I_1 = \frac{10(5.3) - 3.3(5)}{11.90} = 3.07\,\text{mA}\). 5. \(I_2 = \frac{4.3(5) - 10(3.3)}{11.90} = -0.966\,\text{mA}\). 6. \(I_3 = 3.07 - 0.966 = 2.10\,\text{mA}\), so \(V_a = 6.93\,\text{V}\). The minus sign says 0.966 mA really flows down branch 2, into the + terminal of the 5 V source. Node a sits above 5 V, so \(E_2\) is being charged.

Checking the answer

Two checks catch nearly every error. First, KVL around a loop you did not use: in the first example, going from a down branch 2, \(V_a = 9 - 1.0(0.331) = 8.67\,\text{V}\), matching \(4.7 \times 1.84 = 8.67\,\text{V}\) from branch 3. Second, a power balance. With the passive sign convention, the sources deliver \(12(1.51) + 9(0.331) = 21.1\,\text{mW}\), and the resistors absorb \(5.04 + 0.11 + 16.0 = 21.1\,\text{mW}\). Delivered power must equal absorbed power. In the second example, \(E_2\) appears on the absorbing side, since current enters its + terminal.

Strengths and limits

Branch-current analysis always works for linear dc networks and gives every branch current directly, which is exactly what you want when the question asks for currents. Its weakness is size. Every branch adds an unknown, so a network with 6 branches and 4 nodes needs 6 equations: 3 from KCL and 3 from KVL. "Mesh analysis" builds KCL into its unknowns and needs only the 3 KVL equations; "Nodal analysis" needs only the 3 KCL equations. For circuits with two or three branches, though, branch currents are often the quickest route, and the method is a dependable fallback when the others confuse you.

Common mistakes to avoid

First, writing KCL at every node, including the last, and then finding the equations will not solve: one of them is redundant. Second, choosing loops that are not independent, such as two loops that together cover nothing new. Third, marking a resistor's polarity from the source rather than from the assumed branch current. Fourth, flipping the sign of a negative current halfway through; keep the minus sign and interpret it at the end. Fifth, mixing units, such as kilohms with amperes; working consistently in kΩ, mA and V keeps the numbers tidy.

Key takeaways

Branch-current analysis assigns one unknown current to each of the \(b\) branches, with a guessed direction. KCL at \(n - 1\) nodes and KVL around \(b - (n - 1)\) independent loops give exactly enough equations. Resistor polarities follow the assumed currents. Solve the system by substitution, Cramer's rule or a solver; a negative current flows opposite to the guess. Check with KVL around an unused loop and with a power balance. The method is general but needs one equation per branch, which the mesh and nodal methods reduce.

Flashcards

Flashcard

Branch

A path between two nodes carrying one current

One unknown per branch.

Test your understanding

Question 1 of 5

A network has 5 branches and 3 nodes. How many KCL and KVL equations does branch-current analysis need?

Question 2 of 5

Solve 3x + 2y = 12 and x + 4y = 14 using Cramer's rule. What is x?

Question 3 of 5

Branch-current analysis gives I2 = −1.5 mA for a branch whose current was assumed to flow upwards. What does this mean?

Question 4 of 5

Node a is joined to ground by three branches: 12 V (− at ground) with 2.0 kΩ, 6 V (− at ground) with 2.0 kΩ, and 2.0 kΩ alone. The currents in the two source branches are assumed to flow up into a. What is I1, in the 12 V branch?

Question 5 of 5

In a solved circuit of two sources and three resistors, both sources deliver power, 30 mW in total. What must the resistors absorb in total?

← Previous topicNext topic →