In addition to the physics package, passive atomic frequency references require two other components: a local oscillator (LO) to generate the initial periodic signal that is locked to the atoms and some control electronics to implement the various servo systems needed to run the clock. Schematically, these three components are related according to the diagram below.

In a passive atomic frequency standard, the atoms do not generate a stable periodic signal themselves. Rather, a relatively unstable periodic signal from an external source drives the atoms, whose response changes as a function of the LO frequency. By monitoring the atomic response (by measuring the absorption of light through the vapor cell, for example), it is possible to determine how the LO frequency compares with the atomic resonance frequency and whether the LO frequency should be increased or decreased to bring it closer to the atom frequency. In most commercial and laboratory frequency standards, the LO is based on a quartz crystal oscillator. The frequency of the quartz crystal (typically in the range of 5 - 10 MHz) is multiplied up to near the atomic resonance frequency (several gigahertz) using a nonlinear electronic device such as a step-recovery diode or a direct digital synthesizer (DDS). This synthesis chain is also designed to enable frequency modulation of the output RF signal.

Microfabricated atomic clocks are designed for small size and low power consumption. It is likely (although not certain) that traditional local oscillator designs based on quartz crystal oscillators will be too large and power-hungry for application to microfabricated atomic clocks. As a result, significant research to develop small, low-power local oscillators based on a variety of high-frequency resonator technologies, such as thin-film resonators (TFRs), microcoaxial resonators and nanomechanical resonators, is underway. NIST is actively collaborating with organizations developing these advanced oscillators to assess their suitability for use in microfabricated atomic clocks.

In order to control physical parameters of the physics package, as well as implement the correction of the local oscillator relative to the atomic transition, some control electronics is required. In table-top experiments carried out at NIST, four control systems were used as outlined below.

1. 1.Laser wavelength: In order for the laser to be able to excite the atomic resonance, the laser must be tuned to an appropriate optical transition in the atomic spectrum. The wavelength of most semiconductor lasers depends on both the laser temperature and the injection current. Since the wavelength of the light emitted by vertical-cavity surface-emitting lasers (VCSELs) is highly sensitive to the injection current, this parameter is used to control the laser wavelength. The control system implements this by monitoring the absorption of light through the atomic vapor cell and locking the laser wavelength to the peak of the absorption profile using a phase-sensitive detection. By modulating the laser injection current (and therefore the laser wavelength) at a frequency of about 100 kHz, and then demodulating the detected optical power after the light has passed through the cell, the lock can be realized.
   

2. 2.Laser temperature control: This is needed to ensure that the laser wavelength can be locked to the optical transition in the atoms without disturbing the optical power emitted by the device. Since the laser wavelength depends on its temperature as well as its injection current, temperature-induced changes in the wavelength will be compensated by correcting the injection current, which has the side-effect of changing the optical power emitted by the laser. Since changes in optical power cause instabilities in the clock output frequency related to the AC Stark shift (light shift), the optical power should be as stable as possible. To accomplish this, the temperature of the laser is monitored directly and stabilized with a localized heater.
   

3. 3.Cell temperature control: The frequency of the atomic resonance depends on the temperature of the cell that contains the atoms. This is because collisions of the alkali atoms that define the clock frequency with buffer gas atoms cause a shift in the alkali atom resonance frequency. This shift depends on temperature because the properties of the collisions (velocity, energy, etc.) do also. Combinations of buffer gases are used to reduce the temperature coefficient of the cell but even with this compensation, cell temperature stabilization at the level of 100 mK is required to achieve a long-term clock fractional frequency instability below 10-11.
   

4. 4.LO frequency control: This control system implements the basic correction signal that stabilizes the local oscillator onto the atomic transition. Typically, the LO frequency is modulated at a Fourier frequency of several hundred hertz to several kilohertz. The LO is tuned to the CPT resonance and the optical power transmitted through the cell is demodulated to generate an error signal proportional to the difference in frequency between the LO and the atomic resonance. This error signal is used to correct the frequency of the LO.
   

Additional servos that may be necessary to achieve good long-term frequency stability could stabilize the laser optical power and  RF modulation power.

In the table-top experiments set up at NIST, the control circuits are implemented according to the diagram below.