Timing jitter of a COUNT^® module is an uncertainty of time difference between the photon arrival time and the output electric pulse. It is measured by accumulating this time difference over many photons into a histogram.

Timing jitter measurement setup consists of a pulsed laser, attenuator that reduces laser pulses to single-photon ­level, the COUNT^® under test, and the Time Tagger operates here as an electronic correlator. The ­correlator ­records time differences between the laser synchronization signal and the COUNT^® output signal and ­accumulates a histogram. Since the laser sync and COUNT^® output is strongly correlated the histogram has a shape of a peak. The width of this peak contains information about the timing jitter. Note that the width of the correlation peak is also influenced by the timing jitter of the electronic correlator, laser sync signal and the laser pulse duration. The total r.m.s. value of these contributions is given by the following expression:

Hence the measurement system is able to measure timing jitter of COUNT^® modules down to 50 ps.

Dark count rate of a COUNT^® module is a rate of false output pulses that are generated in absence of the input photons. This parameter can be easily measured by counting a number of pulses (N_photons) during a well-defined time interval ΔT:

The Time Tagger provides built-in functionality for the count rate measurements.

After detecting a photon, a photon counting module requires certain reset time before it can detect a next ­photon, namely the dead-time. In COUNT^® modules the dead-time is a fixed value defined by the reset circuit.
In contrast to the dark counts that occur spontaneously, the after-pulsing effect leads to generation of time ­correlated false events. For instance, in COUNT^® modules such false detection events appear due to release of the trapped charges created by a previous avalanche. The after-pulsing manifests itself as an increased ­probability of generating a false signal pulse immediately after the dead-time. 
The measurement setup for an after-pulsing measurement is essentially the same as shown in Figure 2, except ignoring the laser synchronization signal. The laser pulses of 1 MHz repetition-rate are attenuated such that there is < 1 photon per pulse which means that there is at most 1 photon per microsecond. These photons illuminate the module under test and the Time Tagger accumulates a histogram of the time differences between each of the detected photons. This is essentially an autocorrelation histogram. Both, the deadtime and the after-pulsing probability can be determined from such histogram. Figure 3 shows a sample histogram recorded with a real COUNT^® module.

We perform the characterization of the after-pulsing by recording an autocorrelation histogram for the output signal. An example autocorrelation histogram recorded with a real COUNT^® module is shown in Figure 3.
With an ideal detector that has no after-pulsing and no dark counts, one would observe that all counts are ­added only to the bin at zero time-difference (an event against itself). With a real signal , however, the after-­pulsing leads to a distribution spanning a certain time range. Furthermore, a section of the histogram at small time differences remains empty due the deadtime.
The recorded autocorrelation histogram, contains all information necessary for calculation of the integrated after-pulsing probability as:

Where N_ap is the number of pulses which are separated in time at most by 500 ns (bins N₁…N_m) and N_∑ is the total number of all pulses recorded (bin N₀). The above-mentioned expression gives an accurate estimate of the after-pulsing probability as long as the dark rates remain << 10⁶ counts/s.c

Photon detection efficiency is an important characteristic of a photon counting module. It is a probability with which the input photon is converted to a signal pulse. The PDE is defined as a ratio of the detection rate R_COUNT® to the incident photon rate R_in:

Generally, the PDE depends on the material used for light sensing component as well as design parameters, ­like dimensions of the sensitive area and presence of antireflection coatings of the device input windows.

where 𝜆 is laser wavelength, h is the Planck constant, c is the speed of light, and K is the factor that relates ­powermeter reading (P_mon) and the power incident at the COUNT^® module (P_in).
Taking into account the dark countrate R_dark and the deadtime T_dead of the COUNT^® module it is crucial to accurately determine the factor K. Therefore, we replace the COUNT^® module with an accurate and ­calibrated ­powermeter and record the readings from both powermeters simultaneously. Usually, the laser power is ­increased to the level detectable by the calibrating powermeter.