TL;DR
- ▸Quartz crystal oscillators use the piezoelectric resonance of carefully cut crystals as a frequency reference.
- ▸Modern OCXOs deliver stability of a few parts in 10⁹, sufficient for most production timing requirements at a fraction of the cost of atomic alternatives.
- ▸When the holdover requirement exceeds OCXO capability, the next steps are double-OCXO (DOCXO) or rubidium.
Why quartz still wins most of the time
Quartz crystal oscillators have been the foundation of electronic timekeeping for almost a century. A precisely cut crystal of quartz has a piezoelectric resonance at a specific frequency determined by its physical dimensions, and the resonance frequency is stable to a few parts in 10⁹ over short timescales. For timing applications where this stability is enough, quartz is unbeaten on cost, power, size and reliability.
Modern PTP grandmasters mostly use temperature-compensated crystal oscillators (TCXO) or oven-controlled crystal oscillators (OCXO). The temperature compensation or oven control removes the dominant short-term noise source (temperature variation), pushing achievable stability into the sub-nanosecond range at second-scale measurement windows. For deployments where GNSS is reliably available and the holdover requirement is moderate, an OCXO is genuinely sufficient and there's no engineering reason to spend more on atomic alternatives.
When you need to step up
OCXO holdover degrades over hours and days rather than seconds and minutes. A good OCXO drifts roughly 1-10 microseconds per 24 hours of free-run, which is fine for most enterprise deployments where GNSS outages are short and rare. When the holdover requirement exceeds this — typically because the deployment is in a regulated environment with explicit holdover obligations or a contested environment with credible multi-hour GNSS denial — the next step is double-OCXO (two OCXOs cross-disciplined for better stability) or rubidium (atomic frequency reference, much better holdover).
Both DOCXO and rubidium are real upgrades from a single OCXO, both cost more, and both should be specified against a documented holdover requirement rather than picked because they sound more impressive. The most common procurement mistake we see is buying rubidium when DOCXO would be sufficient — paying for capability that the deployment doesn't actually use.
Frequently asked questions
What is an OCXO?+
How much holdover does an OCXO provide?+
When should I choose rubidium over OCXO?+
Related reading
Blog · Foundations
Atomic Clocks and GPS Timing
How atomic clocks define UTC, how GPS distributes that time globally, and what the relationship between the two means for precision timing infrastructure on the ground.
Blog · Hardware
PTP Accuracy with Caesium and Rubidium Oscillators
How atomic-grade oscillators — caesium and rubidium — change the accuracy and holdover profile of a PTP grandmaster, and when their additional cost is justified.