Checking Oscilloscope Sample Rate Across a Long Capture

A startup capture can show the entire power sequence while missing the brief disturbance that caused a reset. The oscilloscope’s advertised maximum sample rate does not establish the rate used for every timebase, enabled channel combination, or memory setting. Evaluate the stored acquisition before treating a smooth trace as evidence of a smooth rail.

Calculate the required record

Write down two different requirements: total observation time and the shortest feature whose shape matters. For uniformly sampled data, record duration is approximately sample count divided by sample rate. A hypothetical ten million point record spanning one second has samples about 100 ns apart. That record cannot describe a 10 ns pulse simply because the instrument can run faster at another timebase.

The Tektronix oscilloscope evaluation primer discusses the connection between record length, sample rate, and capture duration. Use the actual acquisition readout for the selected setup, including any resource sharing between channels.

Separate overview and detail

Start with an overview capture to identify the relevant operating transition. Then trigger a shorter, faster record around that transition, retaining enough pretrigger history to distinguish cause from response. A reset output makes a useful timing reference, but it may occur after the initiating fault. Include the suspected supply or enable signal in the same acquisition when timing relationships matter.

Use delayed acquisition or segmented memory only after understanding what those modes retain and omit. Gaps between segments are not monitored waveform data. A reconstructed sequence can establish event ordering without demonstrating that nothing happened between recorded windows.

Inspect saved samples

Zoom into the saved acquisition and compare the displayed interpolation with actual sample points. Changing the display scale after capture does not recover samples that were never stored. Check whether exported data contains the complete record or a reduced display record; both can look convincing in a spreadsheet.

Repeat with a shorter time span while preserving probe connection, vertical scale, bandwidth limit, and trigger condition. If the apparent edge or glitch changes, investigate the acquisition method before changing the board. For intermittent behavior, record how many relevant events were captured rather than reporting only elapsed bench time.

Make the result reproducible

Save record length, effective sample rate, acquisition mode, enabled channels, trigger settings, and the original waveform file. State separately what the long capture establishes and what the detailed capture establishes. Also distinguish analog bandwidth from sampling density: a high sample rate cannot restore frequencies removed by the probe or front end. This evidence lets another engineer repeat the measurement and understand the smallest event that could reasonably have been observed.