Determining if clock deviation is fixed is very simple; just repeat the step test three times. The core logic is: if the deviation value fluctuates within half a sampling period across the three tests, it's a fixed deviation; otherwise, it's a drift deviation. The specific steps are as follows:
1. Three-Step Test Operation
Prepare a quickly triggered step signal (e.g., a rapid pressure valve switch), clear previous compensation settings, and restore the original clock state;
Trigger three consecutive step signals, with a one-minute interval between each trigger. Record the step time of both sensors each time, and calculate the single deviations ΔT₁, ΔT₂, and ΔT₃;
Calculate the range of the three deviations: Range = Maximum Deviation - Minimum Deviation.
2. Judgment Criteria
Taking a commonly used 10Hz sampling frequency (sampling period 100ms, half a period = 50ms) as an example:
|
Range Range |
Deviation Nature |
Conclusion |
|
Range ≤ 50ms (½ sampling period) |
Fixed Deviation |
Can be directly corrected using the "fixed offset compensation method," done in 5 minutes. |
|
Range > 50ms (½ sampling period) |
Drift Deviation |
Deviation changes over time; fixed compensation is useless, daily automatic time synchronization is required. |
For other sampling frequencies, simply replace ½ sampling period with the corresponding value. For example, for 50Hz sampling (period 20ms), the threshold is 10ms.
3. Quick Auxiliary Judgment
If the device has been running for a period of time, the following characteristics can also indicate a fixed deviation:
The false alarm rate is roughly the same each day and does not increase.
The deviation after each power-on is consistent with the previous one and does not increase with operation.
If the deviation is increasing and the false alarm rate is increasing, it is a drift deviation, requiring periodic time synchronization.
