The fundamental relationship
The time a capture covers is the memory depth divided by the sample rate. Conversely, the sample rate the instrument can maintain is the memory divided by the time the screen spans. If you want to see a long interval in fine detail, many samples are needed, which is why memory is a limited resource.
What happens when the time base is extended
When the visible interval is lengthened, the memory runs out and the instrument lowers the sample rate. With little memory, this drop happens early: a fast event inside a long record can no longer be seen well, or it appears with aliasing. With deep memory, the instrument maintains a high rate for longer, and you can zoom in on the image afterward to see detail.
When it matters most
- Observing a high-frequency pulse within a long window, such as a switching burst in a mains cycle.
- Analyzing protocols, startup sequences or sporadic events.
- Capturing a signal and then reviewing several parts at maximum time resolution.
The trade-off
Processing more samples takes more time, so with very deep memory, the screen update and the response of the controls can be slower. In addition, each sweep leaves a dead time during which nothing is captured. Some instruments offer segmented memory: they divide the memory into many shorter segments, each triggered by an event, so that the intervals without activity between them are skipped.
How to use it well
- Define the total interval you want to observe and the minimum detail you need.
- Calculate the resulting sample rate or verify it on the screen.
- Use deep memory only when you need it, and stay with short capture times in all other cases.
Depth is specified per channel and, on some instruments, is shared among the active channels, so it varies with the configuration.