Short answer: sampling rate is how often the acquisition system measures or updates a value; logging interval is how often a durable record is written. They may be equal, but they do not have to be. Choose each from the fastest event that matters, the process dynamics, filtering and the evidence you need later.
A responsive HMI, an alarm engine and a ten-year archive have different jobs. Recording every internal conversion may create millions of redundant rows, while archiving too slowly can hide a short excursion or make ramp calculations unreliable.
Keep the terms separate
| Term | Meaning | Design question |
|---|---|---|
| Sensor/instrument conversion | How the input hardware obtains a measurement | What conversion and channel-scan behaviour does the instrument provide? |
| Application sample/update | How often software receives or evaluates a value | Can the alarm/rate logic see the fastest relevant change? |
| Display refresh | How often the HMI redraws | Is it readable and responsive without implying false resolution? |
| Archive/log interval | How often a durable time-series point is stored | Can the later report reconstruct the required event and statistics? |
Choose from process evidence, not habit
- Identify the shortest excursion, rate change or stage transition that matters.
- Confirm the sensor and sheath response; fast software cannot recover dynamics the sensor never captured.
- Include instrument filtering, multiplexer scan time and communication latency.
- Select alarm evaluation and display updates.
- Select the durable interval, retention and any event-triggered higher-resolution window.
- Test timestamp consistency, gaps and restart behaviour.
A simple example
A furnace may update the HMI every second, evaluate alarms on each fresh input and archive routine soak data every five or ten seconds. A configured event can preserve a denser window around a deviation. Those numbers are illustrative, not a universal recommendation; a rapid process, survey workflow or customer specification may require a different design.
Watch for misleading averages
If several samples are collapsed into one stored value, document whether it is the latest, mean, minimum, maximum or a set of summary statistics. A single average can erase the excursion an alarm detected. Preserve event flags and, where needed, min/max within each archive interval.
Use the logging interval planner to compare record counts, then the storage calculator to estimate capacity. Neither replaces process-specific sampling analysis.