Table of contents

Based on very exact measurements of emf and temperature in the range from 0 °C to 40 °C the effect of temperature on the electromotive force of saturated standard cells has been redetermined. The most satisfactory mathematical relation was found to be the following fourth-order equation: E_t = E₂₀ - [39.83 (t - 20) + 0.930 (t - 20)² - 0.0090 (t - 20)³ + 0.00006 (t - 20)⁴] · 10^-6.

In the temperature range from 10 °C to 30 °C the third-order equation published in 1971 may continue to be used as in this range the results of the two equations agree to within ± 0.4 μV.

It is shown that the Ramsey resonance curves for most atomic beam machines can be conceived as depending on two distributions of velocity, ρ(V) and ξ(V), the second being a correction for beam width.

An analysis and computer program are described which permit one to obtain ρ, ξ and the nominal microwave power parameter from three or more measured Ramsey resonance curves at properly spaced power levels whose ratios are known. The determination from the functions (ρ, ξ) of bias errors due to second order Doppler shift, cavity phase difference, and cavity pulling is described.

The method may also be used to improve an experimentally obtained velocity distribution (i.e., one obtained through the pulse technique); to provide the proper function ξ; and to provide diagnostic checks of the measurement technique and the validity of the model chosen for the transition probability.

The method is applied to the NBS frequency standard. Error estimates indicate that it is feasible by microwave power shift measurements to evaluate the total bias error due to the above sources to within one part in 10¹³.

A new method for measuring the activity of strong ⁶⁰Co sources is described. It is based on the precise experimental determination by gamma spectrometry of the activity ratio for two sources which are very different in strength but of identical geometry. After dissolution of the weak source its activity is measured by the absolute 4π β-γ coincidence method. The activity of the strong source is obtained from the activity ratio and the activity of the weak source. In this way, a 1 Ci ⁶⁰Co source has been compared with a 2 mCi source. The activity of the 1 Ci source is determined with an accuracy of about 0.2%.

The results of the activity measurement of a 1 Ci ⁶⁰Co source made at the Bureau International des Poids et Mesures are compared to those of the power measurement of a similar ⁶⁰Co source performed at the National Bureau of Standards. For this comparison, the ratio of the activities of the two sources is determined; the method used is described and the errors analysed. The agreement between the activity and power measurements is excellent: the relative difference is less than 1 × 10^-3, which is not significant according to the estimated errors.

The method of generating the NRC atomic time scales UTC(NRC) and AT(NRC) is outlined, and estimates of their uniformity are derived from intercomparisons of the NRC primary standard, Cs III, and the group of atomic clocks maintained at NRC. Comparisons between UTC(NRC) and other national scales through reception of television and Loran C signals are outlined, and it is shown that conflicting results arise from these two methods. In addition, the uniformity of UTC(NRC) deduced from these comparisons appears to be significantly worse than that determined from internal clock comparisons. Variation in signal propagation delay for both TV and Loran C transmissions is proposed as a possible explanation both for the disagreement between TV and Loran C time scale comparisons and for the apparent degradation in the uniformity of UTC(NRC).

This note presents revised results of optical pyrometry, carried out at NPL and elsewhere, following recalculation taking into account the refractive index of air.