A GNSS cycle slip is a discontinuity in the accumulated carrier-phase measurement: the receiver no longer preserves the same whole-cycle count across the event. It changes the ambiguity associated with that satellite signal. Ordinary measurement noise perturbs a continuous observation; loss of lock describes a tracking condition; re-fixing describes a later positioning-solution decision. These events can be related, but they are not the same event or timestamp.
This distinction matters when a log shows a phase jump while position continues, or a receiver reacquires satellites before an ambiguity-fixed solution returns. The useful question is which measurement became discontinuous, how processing handled it, and which output subsequently recovered.
A slip changes the phase arc
While tracking remains continuous, carrier-phase processing follows the changing phase with an initially unknown integer offset. A continuous interval is often called a phase arc. If cycles are missed, the ambiguity after the interruption may differ from the one before it. Treating both intervals as one unchanged arc can introduce an offset into the measurement model.
ESA-hosted Navipedia's cycle-slip reference describes full-cycle slips as integer-wavelength discontinuities. For the underlying measurement, see the existing pseudorange and carrier-phase guide. That guide explains the initial ambiguity; the additional issue here is a change during an observation sequence.
Consider an illustrative model, with geometric motion, clocks and propagation already accounted for. A phase residual before an event fluctuates around one fitted offset. If the ambiguity changes by two full cycles, the subsequent measurements carry an additional 2λ, where λ is that signal's wavelength. This persists until processing accounts for the new offset; it is not merely one isolated noisy sample. It also does not imply that the position output moves by 2λ, because the solution combines measurements and can reject or reset an affected one.
Why tracking continuity can break
Obstruction or attenuation can interrupt usable carrier tracking. In its Waypoint cycle-slip plots, NovAtel discusses low-elevation tracking, blockage and antenna attenuation, and explains why slip patterns can help investigate poor post-processing results. Those observations belong to the documented software workflow; a slip count alone does not identify a defective antenna.
Propagation can also disturb tracking. NovAtel's ionospheric-activity bulletin describes scintillation-driven amplitude and phase fluctuations that can introduce noise or cause loss of lock. Thus a continuity event is not, by itself, evidence of a jammer or spoofing attack. Use signal conditions and receiver records to investigate the cause rather than assigning one from the word “slip”.
A short interruption may fall between logged epochs. Seeing a phase-lock flag set at the next recorded epoch therefore does not establish uninterrupted lock throughout the preceding interval. Conversely, a missing log record can be a transport gap rather than proof that RF tracking failed. Measurement continuity and recording continuity need separate evidence.
Separate four observations in the log
| Observation | What changed | What to conclude from it |
|---|---|---|
| Increased measurement noise | Scatter around the measurement model increased | Assess uncertainty and residual behavior; do not automatically create a new ambiguity arc for every outlier |
| Loss of phase lock | The tracking loop no longer reports valid phase lock | The tracking condition changed; inspect continuity when phase becomes usable again |
| Cycle slip or declared phase discontinuity | The accumulated phase relationship changed, or processing cannot safely carry it forward | Repair with justified evidence or start a new ambiguity arc; do not silently join the intervals |
| Return to an ambiguity-fixed solution | The positioning engine again accepted its ambiguity solution under its criteria | Record the solution recovery separately from signal reacquisition; it does not identify every channel's history |
The distinctions are observable in real interfaces. NovAtel's OEM7 RANGE documentation separately defines code-lock and phase-lock flags, continuous lock time and carrier-phase fields. It also documents half-cycle ambiguity considerations. These are NovAtel field meanings, not ResiNav diagnostics. Decode the receiver's specified conventions before treating every phase-field jump as a whole-cycle slip; unit, sign, half-cycle and status handling matter.
A large jump needs a model, not only a threshold
A raw carrier-phase value normally changes as the satellite and receiver move. Its change between epochs is not a cycle-slip test by itself. A detector instead looks for behavior inconsistent with an appropriate prediction or combination of observations.
Navipedia's single-frequency detector examples use code/carrier combinations and show how code noise, multipath, ionospheric changes and initialization affect detection. A small slip can be hidden in noisy evidence, while a changing non-slip contribution can trigger a false alarm. A universal “jump greater than this many metres” rule loses that context.
With multiple frequencies, combinations can remove large common terms. But Navipedia's multi-frequency examples explain that some simultaneous slips can nearly cancel in one combination. Independent evidence improves the test. A detector flag is consequently evidence about its test and assumptions, not a complete explanation of the event's physical cause.
Receiver behavior is another source of discontinuities. NovAtel explicitly notes in CLOCKADJUST documentation that changing the clock-adjustment mode during tracking can produce jumps in pseudorange, carrier phase and Doppler. This is a configuration example to recognize in a log, not a recommendation to change that setting. A coordinated jump across fields should be investigated with clock and configuration records before being labelled a signal-specific slip.
Reacquisition does not complete ambiguity recovery
After tracking resumes, processing can exclude the affected observations, begin new ambiguities, or repair a discontinuity when sufficient evidence supports the correction. It must not keep an old ambiguity merely because the same satellite identifier reappears.
GNSS-SDR's observables documentation provides a concrete implementation example: continuity flags reach downstream processing so affected ambiguities are not carried across the event, and carrier smoothing restarts. These are documented GNSS-SDR operations, not a universal receiver algorithm. Separately, Navipedia's ambiguity-fixing explanation describes the estimation and bias handling needed for an integer interpretation. Restoring a tracked measurement and accepting a fixed solution remain distinct tasks.
For a hypothetical interrupted run, record three milestones: carrier tracking becomes usable again, the estimator accepts its new ambiguity solution, and the platform meets its own return-to-use criteria. A position may remain available from other measurements between those milestones. There is no reason to force the three times to coincide or to infer a universal recovery duration from this example.
Turn the event into interpretable evidence
Retain the satellite and signal identity, epoch, phase units, lock/continuity flags, observation gaps, relevant configuration changes and solution state around the event. Compare like signals at like epochs. Keep the original records when later software repairs or rejects observations so the processing decision remains traceable.
Use the receiver interface guide to identify available fields, and the receiver evaluation guide to distinguish receiver recovery from platform re-entry. This article adds the measurement-continuity mechanism to that existing evaluation framework; it does not assign cycle-slip detection, raw-observation export or recovery performance to a ResiNav model.
The result should be a specific account of what broke and what recovered: a noisy observation, an interrupted phase arc, a restored tracking state, or a newly accepted navigation solution. Each supports a different engineering conclusion.
NEXT STEP
Turn available platform information into an engineering review.
Use the Technology centre to frame the discussion, review relevant application scenarios, then send the available platform and receiver details for confirmation.