Method of measuring frequency using time stamp counter

Frequency counters were once considered basic measurement tools at the level of voltmeters, capable of measuring a CW signal with high resolution depending on modulation, and were widely used to calibrate oscillators. As frequency counters evolved, the measuring instruments gained the ability to measure pulses and had wider measurement bandwidths capable of handling measurements on modulated waveforms. Modern frequency counters have reached new levels of performance, capable of discerning dynamic signals that continually change frequency.

Frequency counters have undergone tremendous changes in design and functionality over the past decade. Today's frequency counters, such as the CNT-90 from Pendulum Instruments (www.pendulum-instruments.com), use advanced measurement principles to achieve very high sampling rates and higher resolution than previous test instruments. This particular frequency counter has a graphical display that shows signal jitter and modulation, as well as other important characteristics of the signal.

Since the first frequency counter appeared 25 years ago, the improvement of frequency counters has been slowly and steadily progressing. The table shows the change in frequency resolution in the continuous upgrading of frequency counters. CNT-90 is the fourth generation product, which can store 250,000 measurements in the built-in memory and has a frequency resolution of 12 bits. Because it combines the functions of timer, counter and analyzer, it can also measure time and signal phase with a resolution of 100ps and 0.001deg. respectively.

One of the most common measurement methods used by frequency counters today is a technique known as reciprocal counting. This method is based on measuring the time (clock pulses) between two identical trigger points in the input signal. During these start and stop trigger events, the instrument measures the number of signal cycles (N) and the number of clock pulses from the internal reference oscillator that are generated during the T _N period. The internal microprocessor then calculates the frequency as "N/T _N " or the period as "T _N /N"; where N is an integer number of measured signal cycles from the two trigger points.

In the countdown, the counter measures a multiple period T _N (in N cycles), which is converted into a frequency by the microprocessor. Of course, this frequency is the inverse of the period (f=T ^-1 ). It should be emphasized that N is an integer, because the measurement (start/stop of the two counting chains) is synchronized with the input signal (not with the clock pulse) (Figure 2). However, the measurement and the clock pulse are completely out of sync. Therefore, it is possible that part of the clock pulse period is measured, while part is not.

Therefore, for a reciprocating counter, the resolution in the time measurement T _{N is one clock pulse period. Most counters incorporate a 10MHz reference oscillator that translates into a 100ns period. This reference enables a relative resolution (equivalent to the absolute resolution divided by the measured time) of 7 bits within a measurement time of 1s (100ns/1s=10} ^-7 ), regardless of the input signal frequency.

What if this resolution is not enough for some applications? There are many ways to improve the 7-bit resolution of a standard reciprocal counter. One way is to increase the base clock frequency. The resolution can also be improved by increasing the clock frequency of the time base oscillator from 10MHz to 100MHz in 10ns. After this improvement, for a measurement time of 1s, the resolution will be 8 bits (10ns/1s=10-8 ⁾ .

The measurement resolution of interpolating counters is further improved. Interpolating counters have the basic performance of reciprocal counters, plus the uncertainty of time measurement of one time period. This is because the start and end points of the measurement in the clock pulse period are unknown. With the help of a special interpolation circuit, the phase angle of the clock pulse at the beginning and end of the measurement can be determined. At this time, there are always two identical interpolators running at the same time: one for the start trigger event and one for the end trigger event.

Such an interpolator can be constructed in a variety of ways. One of the more common ones is the analog interpolator, where the time difference between the trigger event and the clock signal is converted into an analog voltage that can be measured by an analog-to-digital converter (ADC). Through interpolation, the resolution of the time measurement is theoretically improved from the 100ns of the "digital clock pulse counting" method to the ratio between a clock period and the interpolation factor of the interpolator count.

In practice, achieving this level of accuracy is difficult because there are several sources of error, including interpolation linearity, that must be accurately located and controlled. Interpolation factors are typically between 100 and 500. With a clock frequency of 10 MHz, typical time resolution would be 200 ps to 1 ns (100 ns for a conventional 10-MHz reference clock), and 9 to 10 bits of resolution can be achieved over a 1-second test period. The CNT-81 timer/counter/analyzer combines interpolation with a 100-MHz reference clock oscillator to achieve 50 ps of time resolution and 11 bits of frequency resolution over a 1-second measurement period.