How Does an Oscilloscope Work? A Bench Guide to Capturing PWM and Ripple

At its core, an oscilloscope works by translating continuous analog voltage into discrete digital samples using an Analog-to-Digital Converter (ADC), then mapping those voltage-time coordinates to a display grid. But understanding the block diagram isn't enough to debug a noisy power supply or a failing microcontroller. To actually use the instrument, you need to understand how the probe attenuates the signal, how the trigger system freezes time, and what a healthy waveform looks like in hard numbers.

This guide moves past the textbook theory of how does oscilloscope work and onto the workbench. We will trace the signal path, configure the front panel for a real-world power electronics test, and define the exact numerical thresholds that separate a good circuit from a failing one.

The Signal Path: From Probe Tip to Pixel

When you touch a probe to a test point, the signal doesn't just flow into the screen. It passes through a precision voltage divider. A standard 10X passive probe contains a 9 MΩ series resistor that works with the oscilloscope’s internal 1 MΩ input impedance to divide the voltage by 10. This protects the sensitive front-end amplifiers and reduces the capacitive loading on your circuit.

Once attenuated, the signal hits the ADC. If your oscilloscope has an 8-bit ADC and the vertical scale is set to 5.00 V/div (with an 8-division screen height, totaling 40V full-scale), the ADC resolves that 40V range into 256 discrete steps. That means each digital step represents roughly 156 mV. Any noise or ripple smaller than the quantization error might be masked, which is why zooming in (reducing V/div) is critical for measuring millivolt-level ripple on a 12V or 48V rail.

Bench Setup: Probe Placement & Expected Readings

To ground the theory, let's look at a common benchmark test: capturing the PWM gate drive and output ripple on a 12V-to-120V inverter or switching power supply. Below is the exact instrument configuration and the numerical thresholds you should expect.

Probe Placement Technique

Where and how you place the ground reference dictates your signal integrity. For Test Point A (PWM Gate), place the probe tip directly on the gate resistor pad. For the ground connection, never use the 6-inch alligator clip lead when measuring fast switching edges; its inductance will create false ringing. Instead, slide the plastic sleeve off the probe tip and use the ground spring to connect directly to the MOSFET source pin. For Test Point B (Output Ripple), place the tip on the output capacitor's positive terminal and the ground spring on the capacitor's grounded metal can or negative pad.

Expected Readings: Good vs. Bad Waveforms

Here is the data-dense reference table for evaluating your capture. A "good" reading means the circuit is operating within safe semiconductor limits and EMI specifications.

Measurement Point	Good Reading (Numerical)	Bad Reading (Failure Mode)	Underlying Cause of Bad Reading
PWM High State (Gate Drive)	11.8V flat, <50mV droop	9.2V with 1.5V droop during ON time	Weak gate driver IC, insufficient bootstrap capacitor, or excessive gate resistance.
PWM Rise/Fall Time	<80ns (10% to 90%)	>300ns with severe Miller plateau step	Driver cannot source enough peak current to charge the MOSFET's Miller capacitance (Cgd).
Output DC Ripple (AC Coupled)	<50mV peak-to-peak	>250mV p-p with high-frequency spikes	Degraded output electrolytic capacitors (high ESR) or poor PCB layout causing ground bounce.
Switching Node (Drain) Ringing	Overshoot <20% of bus voltage	Overshoot >60%, exceeding Vds max rating	Excessive parasitic trace inductance; requires an RC snubber circuit across the drain-source.

Common Mistakes That Yield Misleading Readings

Even with a high-end Tektronix or Keysight oscilloscope, operator error can fabricate waveforms that don't exist in the physical circuit. Watch out for these three traps:

1. The 1X / 10X Menu Mismatch

Passive probes have a physical switch on the probe body (1X or 10X). The oscilloscope has a corresponding digital menu setting for each channel. If your physical probe is set to 10X (attenuating the signal by 10), but the scope's digital menu is accidentally set to 1X, the oscilloscope will multiply the incoming signal by 1 instead of 10 to "compensate." A 12V battery will suddenly read as 120V on the screen. Always verify the physical switch and the digital menu match before taking measurements.

2. The Ground Clip Antenna Effect

As mentioned in the setup block, using the standard 6-inch alligator ground clip creates a loop antenna with roughly 100nH of inductance. When you measure a fast PWM edge (which contains high-frequency harmonics), this inductance resonates with the probe's input capacitance, displaying massive 50MHz+ ringing on the screen. The circuit isn't actually ringing that badly; your probe is. Switch to the ground spring (reducing inductance to ~10nH) to see the true signal. For a deeper dive into probe parasitics, refer to Keysight's probing best practices.

3. AC Coupling on DC-Biased Control Signals

If you are measuring a 3.3V microcontroller PWM signal, but you accidentally leave the channel coupling on "AC", the oscilloscope inserts a series capacitor in the signal path. This blocks the DC component and centers the waveform around 0V. Your 0V to 3.3V logic signal will suddenly appear as a -1.65V to +1.65V signal, which can lead you to falsely diagnose a negative voltage fault in your logic circuit. Use DC coupling for all digital logic and gate drive measurements.

Verifying the Capture: Triggering and Cursors

Understanding how an oscilloscope works requires mastering the trigger system. The trigger is the "start gun" for the ADC. Without a stable trigger, the scope continuously sweeps random slices of time, resulting in a blurred, rolling mess on the screen.

For repetitive signals like PWM, use Edge Triggering. Set the trigger level to a voltage that the signal actually crosses (e.g., 2.5V for a 5V signal). More importantly, change the trigger mode from Auto to Normal. In "Auto" mode, if the signal drops out or falls below the trigger threshold, the oscilloscope will force a sweep anyway, making it look like a signal is still present. In "Normal" mode, the screen freezes on the last valid capture until a new trigger event occurs. If your circuit intermittently shuts down, "Normal" mode will freeze the exact moment of failure on the screen, allowing you to use the oscilloscope's cursor tools to measure the exact voltage droop or timing delay that caused the fault.

By combining proper probe attenuation, strict attention to ground inductance, and a disciplined trigger setup, you transition from simply looking at squiggly lines to extracting precise, actionable engineering data from your circuits.