In short. Event analysis needs a common clock. If relays, generator controllers and SCADA disagree by seconds, the sequence of events cannot be trusted. Choose a time source and protocol that matches the accuracy you need, monitor it and flag time quality on records.
Accuracy needs
| Use | Typical accuracy | Common method |
|---|---|---|
| Alarm and operator logs | Around a second | NTP |
| Sequence of events on relays | About a millisecond | IRIG-B, PTP, SNTP with good network |
| Sampled values and synchrophasors | Microseconds | PTP or IRIG-B with GPS |
Architecture
- A master clock with a GPS or other reliable source and holdover.
- Distribution by IRIG-B, PTP or NTP to devices.
- Redundant master clocks for critical sites.
- Monitoring of clock status and offset.
- A defined fall-back if GPS is lost, with an alarm.
Time quality
Records should carry a flag for time quality: synchronised, drifting or unsynchronised. A mismatch between devices is easy to spot if flags are recorded. Daylight saving and time zone settings must be consistent; use UTC internally in many systems.
Correlation workflow
- Gather event lists and disturbance files from all devices.
- Align times and check offsets with a known event.
- Build a combined timeline.
- Identify the first event and follow the chain.
- Compare to the cause-and-effect design. See cause and effect.
See sequence of events.
Generator specifics
- Generator controllers often have lower time accuracy; know what they provide.
- Time jumps can disturb logs and schedules; handle step changes carefully.
- Test periodic functions that use time, such as exercise runs, after resynchronisation.
Illustrative example
After a trip, two relays report the breaker failure signal 3 seconds apart, which seems impossible. Investigation shows one relay lost its time source and drifted. With clock monitoring and time-quality flags, the discrepancy would have been obvious and the correlation could have been fixed in minutes.
Why clocks matter
Event analysis needs a common clock. If device clocks differ by a few seconds, the order of events is unreliable and root-cause analysis becomes guesswork.
| Method | Accuracy | Use |
|---|---|---|
| NTP | Milliseconds to tens of ms | SCADA servers, HMIs, general devices |
| SNTP in devices | Tens of ms | Basic relay clocks |
| IRIG-B | Microseconds to 1 ms | Protection relays, recorders |
| PTP (IEEE 1588, power profile) | Sub-microsecond to microsecond | Sampled values and precise correlation |
Architecture
- GPS-based master clock, with a holdover oscillator for GPS loss.
- Distribution by IRIG-B, PTP or NTP, according to device capability.
- Alarm on loss of sync and on offset beyond limit.
Local time and UTC
Record in UTC and display in local time (India Standard Time). Avoid manual clock setting, which introduces drift.
Practical impact
Two relays that disagree by 40 ms may seem to show a downstream breaker tripping before the fault was detected. With synchronised clocks, the timeline is trustworthy and misoperation can be pinpointed.
Commissioning check
Compare device times against the master after setup, and verify behaviour on GPS loss. See sequence of events.
The master engineering question
If a disturbance happens, can every record from every device be placed on one timeline with a known accuracy?
What to send us
Send device lists and event timing needs. Send your application details.
Questions engineers ask
Is NTP enough?
For logs, often yes. For millisecond event analysis, use better methods.
What is holdover?
The ability of a clock to keep good time without its reference for a while.
Why UTC?
It avoids daylight saving confusion.
How do I check sync?
Compare device times against the master, and alarm on offset.
This guide gives general principles. Confirm device capabilities.