arXiv:2609.31837
Measuring Clock Precision Without an Ideal Time Reference
Pablo Á. Domínguez, Christopher Chubb, Mischa P. Woods
Abstract
Characterising physical devices typically requires comparison with more precise reference standards. For clocks, however, this approach is problematic because no physically perfect time reference exists; real clocks have finite, model-dependent uncertainty. We address whether multiple imperfect clocks can collectively determine their precision without an absolute reference. This paper provides analytic bounds and protocols for quantifying clock precision using only multiple copies of the same clock, which need not be synchronised and may have manufacturing variations or environmental coupling. Beyond clock characterisation, these methods enable precise measurements f sub-period intervals and relativistic effects such as time dilation and Doppler shifts, providing tools for quantum metrology and fundamental physics.
AI Summary
Can the dimensionless precision of a ticking clock, $R=\mu^2/\sigma^2$, be inferred without an ideal time reference? This paper studies protocols using multiple imperfect clocks, modelled as independent renewal processes with i.i.d. inter-tick intervals. The observables are tick counts and tick ordering rather than externally measured timestamps. The main two-clock protocol starts equal-frequency clocks with an offset $\Delta=\delta\mu$, then counts their combined ticks until one produces two consecutive ticks. Writing this count as $\tau$, it establishes $\mathbb{E}[\tau]/R=\delta(1-\delta)+O(R^{-1/2+1/p})$ and $P(\tau\text{ odd})=\delta+O(R^{-1/2+1/p})$, assuming uniformly bounded standardized absolute moments of order $p>2$. Thus the unknown offset can be inferred from the same data, giving $\mathbb{E}[\tau]/R=P_{\rm odd}P_{\rm even}+O(R^{-1/2+1/p})$. Higher stopping-time moments obey a polynomial hierarchy corresponding to Brownian exit-time moments. Proofs use martingale stopping arguments and overshoot bounds. A many-clock protocol instead estimates $R$ through the mean-to-variance ratio of long-window tick counts, using renewal asymptotics. Gaussian Monte Carlo simulations illustrate the two-clock relations for $R=10^2$–$10^4$. Extensions give asymptotic equations for jointly inferring precision, initial offset and frequency mismatch when two clocks have equal $R$ but different periods. These require additional distributional assumptions, including asymptotically mesokurtic fluctuations. The stated practical limitations include long runs and many repetitions for precise clocks, and difficulty manufacturing clocks with equal precision. The framework assumes i.i.d. ticking; its two-clock error bounds decrease with increasing intrinsic $R$, not simply with additional trials.
Central claim may not hold: The central two-clock proof replaces the parity-specific double-tick event with unrestricted first exit, and an explicit bounded-moment lattice example contradicts its claimed exit-parity estimate. The unequal-frequency derivation drops the clock-frequency drift and uses a non-martingale in optional stopping, while the conclusion additionally claims an uncertainty rate not proved for that extension. Theorem 1 claims an inverse-duration error under sublinear duration-variance growth, although its own variance decomposition and an explicit Poisson-count example show that this assumption permits slower convergence. The stated Allan-variance formula omits nominal-frequency normalisation: for fractional frequency the renewal-model result is asymptotically mu/(R t), not 1/(R mu t). The operational guarantees are narrower than advertised: they assume i.i.d. ticking and equal precision across copies, and the two-clock asymptotic model-error bounds are not finite-sample guarantees or arbitrary-accuracy estimates at fixed R.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 19 pages, 4 figures