Norton resistance is the equivalent internal resistance of a linear electrical network as measured across two specific output terminals when all independent power sources are deactivated. In practical electronics, this single value dictates how much your output voltage will sag under load, the maximum short-circuit current your network can deliver, and whether your microcontroller's analog-to-digital converter (ADC) will return accurate readings or garbage data. Think of it as the number of closed lanes on a highway off-ramp: even if the traffic lights (independent power sources) are turned off, the physical bottleneck (resistance) still dictates the maximum flow rate when cars are pushed through from the outside.

The Most Common Confusion: Beginners frequently confuse Norton resistance ($R_N$) with Thevenin resistance ($R_{Th}$). In reality, they are mathematically identical ($R_N = R_{Th}$). The only difference is the lens you use: Norton models the circuit as a current source in parallel with $R_N$, while Thevenin models it as a voltage source in series with $R_{Th}$. A second, more dangerous mistake is turning off dependent sources (like the output of an op-amp or a transistor's controlled source) during calculation. Independent sources get deactivated; dependent sources must remain active.

What Norton Resistance Actually Changes in Your Circuit

When you design a sensor interface or a biasing network, you are rarely building a standalone component; you are building a source that must drive a load. The Norton resistance of your source network determines the stiffness of that output.

If your Norton resistance is high, the circuit behaves like a weak current source. When you connect a load, the voltage at the terminals will drop significantly because the internal resistance forms a voltage divider with your load. If your Norton resistance is low, the circuit behaves like a stiff voltage source, capable of delivering high transient currents without the terminal voltage sagging.

This is not just theoretical. In mixed-signal PCB design, a high source resistance interacting with the switched-capacitor input of a modern SAR (Successive Approximation Register) ADC causes incomplete charging of the internal sampling capacitor. According to STMicroelectronics application note AN2834, if the source impedance exceeds the ADC's specified threshold during the acquisition window, you will see non-linear errors that masquerade as random noise, completely destroying your 12-bit or 16-bit resolution.

The Math: A Worked Numeric Example

Let us calculate the Norton resistance for a highly common real-world scenario: a precision voltage reference driving a microcontroller ADC through an isolation resistor.

The Circuit:

  • A 5.000V precision shunt reference (e.g., LM4040) acting as our independent voltage source.
  • A 100Ω series resistor ($R_S$) placed immediately after the reference to isolate capacitive loads and ensure stability.
  • A 10kΩ pull-down resistor ($R_P$) to ground, used to bleed off static charge and define a default zero-state if the reference fails open.
  • Our output terminals are taken across the pull-down resistor (where the ADC input will connect).

Step 1: Deactivate Independent Sources
We turn off the 5.000V LM4040 reference. Because it is an independent voltage source, we replace it with a short circuit (a wire to ground).

Step 2: Calculate Equivalent Resistance
Looking into the output terminals, the short circuit places the 100Ω series resistor in parallel with the 10kΩ pull-down resistor. The formula for two parallel resistors is:

$$R_N = \frac{R_S \times R_P}{R_S + R_P}$$

$$R_N = \frac{100 \times 10,000}{100 + 10,000} = \frac{1,000,000}{10,100} \approx 99.01Ω$$

Step 3: Evaluate Against the Load
Your Norton resistance is 99.01Ω. If you connect this directly to the ADC pin of an STM32F4 running at a 1MHz sampling rate, you will exceed the recommended 50Ω maximum source impedance. The internal 12pF sampling capacitor will not have enough time to charge to the true voltage during the ~3μs acquisition window. The fix is not to change the 100Ω resistor (which is needed for stability), but to lower the Norton resistance seen by the ADC by inserting a unity-gain op-amp buffer.

Where You Meet This in Practice

While Norton's Theorem is often taught purely as an academic exercise in DC circuit analysis, it maps directly to physical design challenges across multiple disciplines:

Audio Amplifier Output Stages: In audio, the Norton (or Thevenin) resistance of the amplifier output stage defines the damping factor. A very low output resistance (high damping factor) allows the amplifier to tightly control the speaker cone's movement, stopping it from ringing after a transient bass note.
  • Sensor Signal Conditioning: Wheatstone bridges used in strain gauges have a relatively high Norton resistance (often 350Ω to 1kΩ). They cannot drive ADCs directly and require instrumentation amplifiers to transform that high resistance into a near-zero output impedance.
  • Solar Panel Strings: A photovoltaic array can be modeled as a Norton equivalent circuit where the current source represents photon-generated current, and the Norton resistance represents the parallel leakage and shunt resistance of the cells. As panels degrade, this parallel resistance drops, reducing maximum power output.
  • 4-20mA Current Loops: Industrial transmitters act as high-impedance Norton current sources. The Norton resistance of a healthy loop transmitter is typically >500kΩ, ensuring that the current remains constant regardless of the voltage drop across the receiving shunt resistor.

Decision Tree: Sizing Your Load Against Norton Resistance

Do not default to "making the load as high as possible." The correct relationship between your load resistance ($R_L$) and your source's Norton resistance ($R_N$) depends entirely on what physical quantity you are trying to transfer. Use this decision matrix to select your exact component values.

Design Goal Mathematical Rule Why It Matters Concrete Component Pick / Value
Maximize Voltage Transfer (ADC inputs, sensor buffers) $R_L \ge 100 \times R_N$ Prevents voltage divider sag and ensures ADC sampling capacitors charge fully within the acquisition window. Pick: OPA333 unity-gain buffer. (Transforms a 100Ω source into a <0.1Ω source, allowing a 10kΩ ADC input to see a perfect voltage).
Maximize Power Transfer (RF antennas, audio speakers) $R_L = R_N$ The Maximum Power Transfer Theorem dictates that power delivery peaks exactly when load matches source impedance. Efficiency is only 50%, but signal strength is maximized. Pick: 8Ω nominal speaker for an amplifier measured at 8Ω output impedance, or a 50Ω termination resistor for a 50Ω RF trace.
Maximize Current Transfer (4-20mA loops, biasing current mirrors) $R_L \ll R_N$ (Ideally $R_L \le 0.01 \times R_N$) Ensures the load does not restrict the current flow dictated by the source network, keeping the loop current independent of wire length. Pick: 250Ω precision shunt resistor (e.g., Vishay Y1487) for a 4-20mA receiver, assuming the transmitter has a Norton resistance >25kΩ.

Frequently Asked Questions

Can Norton resistance be negative?

Yes, but only in circuits containing active dependent sources (like op-amps, transistors, or switching regulators with specific feedback loops). A negative Norton resistance means the circuit actively pushes more voltage as you draw more current, effectively canceling out parasitic resistances. This is a deliberate design technique used in negative impedance converters (NICs) to cancel out the resistance of long transmission lines or to create high-Q active filters. You will never measure a negative resistance with a standard multimeter on a passive network.

How do I measure Norton resistance on a physical PCB with a multimeter?

You cannot just put a multimeter in resistance mode across a live circuit; the internal voltage sources will damage the meter or yield false readings. To measure it physically:
1. Power down the board and discharge all capacitors.
2. Short out any independent voltage sources (or remove batteries/power supplies and place a jumper wire across the input terminals).
3. Remove or open any independent current sources.
4. Set your multimeter to the Ohms (Ω) range and probe the two output terminals.
Note: This only works for purely passive or linear networks. If your circuit relies on active switching or dependent semiconductor biasing to establish its output impedance, you must calculate it analytically or measure it dynamically by applying two different known loads and calculating the slope of the V-I curve.

What happens to Norton resistance if I add a capacitor in parallel?

Resistance is a DC concept; when you introduce capacitors or inductors, you transition from Norton resistance to Norton impedance ($Z_N$), which varies with frequency. A capacitor placed in parallel with your output terminals will lower the Norton impedance at high frequencies (acting as a bypass or decoupling capacitor), which is exactly why we place 100nF MLCC capacitors across the power pins of microcontrollers to provide a low-impedance source for high-frequency transient switching currents.