Vertical Sampling Ensures Accurate Results

We all know that engineers and technicians use oscilloscopes to show what their signals look like—even when the scope signal display is not the primary point of interest. The scope could monitor the output of a function generator or the signal input to a frequency counter.

Most people use a scope with a counter almost as an insurance policy. They want to be certain that the number they see on the counter relates to the signal fed into the counter.

A new technique uses a frequency counter/timer to display waveforms on the frequency counter itself. This minimizes the need for a scope in most applications where the scope’s only function is to instill confidence in the measurements made on your counter.

In principle, the technique is very simple, analogous to transitional timing in some logic analyzers. This happens when the analyzer only records the moment when the signal changes state from low to high or high to low, which saves memory.

The implementation in the waveform-displaying frequency counter is similar. Only now, instead of two voltage levels as in the logic analyzer, the process depends on many accurately measured voltage levels that are precisely timed to a reference start-time point.

When compared to a scope, higher signal reproduction fidelity results can be achieved, particularly on signals with fast edges. When compared to a similarly priced digital storage oscilloscope (DSO), the timing resolution is phenomenal because the process relies on the timing resolution of the counter/timer, not on a DSO’s sample clock. The counter’s display resolution is typically 1 ns + 1 pixel.

Horizontal vs Vertical Sampling

A DSO operates by sampling the waveform at fixed intervals governed by the sample clock speed. The voltage of the sample is converted into a binary number and stored in the DSO’s acquisition memory. From there, it is repetitively read out through a digital-to-analog converter to the scope’s screen.

The controlling factor in the process is the sample rate. The sample rate is governed by the setting of the time base that is a time function on the horizontal or X axis. So DSOs could be said to have “horizontal sampling.”

If a signal has fast voltage transitions, then the DSO takes only a few samples on the transitions because it is limited by the time between samples given by the reciprocal of the sample rate. When a DSO time base is running at low speed, so is the digitizer and this limits the fastest transition that the DSO can display at that time-base speed.

Figure 1 shows that the faster rising edge of the signal is sampled less than the slower falling edge in this typical charge/discharge curve. The sampling process controls the acquisition, and the sample clock operating on the X-axis controls the amount of signal detail acquired.

In the vertical sampling in Figure 2, the trigger voltage of the counter/timer is increased in small discrete steps as shown by the horizontal dotted lines. When the timer circuit is triggered at t_n, the elapsed time after t₀ is used to locate the sample for waveform reconstruction.

The timer function has a very high resolution, typically in the order of 1 ns. This is not affected by a time-base speed as in a DSO, and this very high resolution is available all the time. Because the samples are taken independently of a sample clock, there is a tremendous wealth of information acquired when there are voltage changes in the signal.

Compare the number of samples on the leading edge between Figure 1 and Figure 2. If you are making timing measurements, you would certainly choose Figure 2 for higher-resolution results. In fact, in Figure 1 there are 18 sample clock pulses; in Figure 2, there are only 10 trigger voltage levels used.

Putting Theory Into Practice

This new waveform display function works best on signals that are stable. This fits in well with the way counters are employed because counters are used primarily on stable signals. If they were not, then the typical eight to 10 digits of resolution would be meaningless and the count would change so fast that it would be unreadable. Remember, seeing the waveform with a counter gives you the confidence normally obtained by using a scope in parallel with the counter, and does not replace the scope for measurement purposes.

Figure 3 shows an example of the signal that would produce uncertain results on a conventional frequency counter. The large spikes superimposed on the sine wave have their own distinctive period. If the trigger level of a conventional counter is incorrectly set, it could easily count on a mixture of the signal’s period and the interfering spikes.

Because we can see the waveform, we already know that there is a chance of a counting error. Figure 3 is the screen from the waveform-displaying counter. As with a conventional counter, the frequency is given at the top—but that is as far as convention goes. The triggering circuitry is used to determine the peak-peak voltage shown in the top right corner of the waveform display.

The dotted band through the center of the waveform shows the trigger hysteresis for the counter. This is set by the counter to remove the chance of false counts. To register a count, the signal must rise fully through the band, fall through it and rise through it again.

Looking at Figure 3, this means that the spikes now have no influence on the final reading because the hysteresis band is too wide for them to have any effect. All this is plain to see without a scope, and the waveform display settings are shown just under the waveform itself.

Figure 4 and Figure 5 show an excellent example of why the waveform display assures you of a valid measurement result. The difference between the figures is the trigger level. Without the waveform, there is no clue to which reading is correct. With the waveform display, there is no possibility of error and it is obvious that the correct pulse width reading is given in Figure 5.

When implemented in a frequency counter, this new vertical sampling technique gives you more information about the signal you are measuring and confidence in your results. It also makes life easier on the bench and especially in the field because you no longer have to tote a scope to do your work effectively.

About the Author

Charles Holtom is a Product Marketing Manager at Fluke. Previously, he worked in sales, marketing and product-planning capacities for Fluke and the Fluke/Philips alliance in England and The Netherlands. Mr. Holtom studied at the British Royal Naval Air Engineering School and later worked on helicopter avionics systems. Fluke, P.O. Box 9090, Everett, WA 98206-9090, (206) 347-6100.

August 1996