1. Frequency/Period
There are several functions suited for measuring Frequency or Period. Next sections provide details about each particular function and when one should be preferred over another.
1.1. Frequency/Period Average measurements
These are the most universal measurement functions for Frequency and Period. In this mode each sample is a Frequency/Period value averaged over sample interval (which acts as gate).
This is back-to-back measurement with no dead-time between the samples (see chapter 4.6 Arming for exceptions). 4 input signals can be measured in parallel. Minimal sample interval is 50 ns. Up to 32 million samples total can be measured in a single measurement session. Resolution is 12 digits per 1 s of gate time (Sample Interval).
If the signal period is greater or equal to the set Sample Interval – each signal period can be captured. When measuring Frequency/Period Average on inputs A, B, D, E and Trigger Mode is set to Auto or Relative, wide hysteresis (see details below) is used to improve noise tolerance. In this mode 2 comparators with different trigger levels are used for each input. First trigger level (e.g. Trigger Level A) defines the upper limit of wide hysteresis band and the second one (e.g. Trigger Level A2) defines the lower limit. Trigger Mode Auto sets trigger levels to 60% and 40% of signal’s voltage range and Relative allows modifying them to fine tune the hysteresis band.
Without wide hysteresis, the signal needs to cross the approx. 20 mV in case of 1x Attenuation (200 mV in case of 10x) input hysteresis band before triggering occurs. This hysteresis prevents the input from self-oscillating and reduces its sensitivity to noise. If signal noise is comparable or higher than hysteresis band – it can result in false extra triggering producing erroneous counts. These could ruin the measurement.
Figure 11 shows how spurious signals can cause the input signal to cross the trigger or hysteresis window more than once per input cycle and give erroneous counts. Figure 12 shows that a wide enough hysteresis prevents false counts.
1.1.1. Frequency C measurement.
With an optional RF input prescaler the Analyzer can measure up to 3, 10, 15, 20, or 24 GHz on Input C. These RF inputs are fully automatic, and no trigger setup is required. Set Sample Interval to achieve optimal compromise between resolution (long Sample Interval) and speed (short Sample Interval). The optional RF input C contains a prescaler that divides the RF signal with an integer value (Prescaler factor), to enable the normal counting circuitry to measure the frequency. The Option 10 (3 GHz) divides by 16, and the option 110/xx (10 to 24 GHz) divides by 64.
Figure 14 shows the effect of the 3 GHz prescaler. For each 16 input cycles, the prescaler gives one square wave output cycle. An input frequency of let’s say 1.6 GHz is divided down to 100 MHz and measured by the normal counting circuitry. The display shows the correct input frequency since the microcomputer compensates for the effect of the division factor.
Prescalers do not reduce resolution. The relative quantization error is the same;12-13 digits for 1s Sample Interval (Gate Time). See Table 1 to find the prescaler factors.
| Function | Prescaling Factor |
| All input A, B, D, E, EA, ER functions (up to 400 MHz) | 1 |
| Frequency C (3 GHz) | 16 |
| Frequency C (10, 15, 20, 24 GHz) | 64 |
1.2. Smart Frequency/Period
Smart Frequency/Period is based on the same principle as normal Frequency/Period Average. Measurement gate is divided into 1000 sub-gates, giving additional samples and statistics resolution enhancement algorithm is applied on top of it. Thanks to that it allows to get up to 1 extra digit of resolution per 1 s of gate time (depending on input signal and measurement settings).
This comes with expense of additional constraints though. Minimal possible Sample Interval for Smart Frequency/Period is 50 us with possibility to measure up to 32000 samples per measurement session. Setting Sample Interval to 40 ms or more allows to get up to 4290000 samples per measurement session.
This is back-to-back measurement with no dead-time between the samples (see chapter 4.6 Arming for exceptions). 4 signals can be measured in parallel. Wide Hysteresis is used in Trigger Modes Auto and Relative.
Smart Frequency/Period is the best choice when one needs maximal possible resolution and can bear with lower sampling rate and sample count per session.
Please, note: resolution enhancement algorithm is based on the assumption that signal frequency is static. If it is not the case – the algorithm won’t be effective and it might make sense to fall back to normal Frequency/Period Average. Please note: section Frequency C measurement. Applies to Smart Frequency C as well.
1.3. Period Single
This measurement function is handy if one needs to capture individual periods of continuous signals or single cycles which are less than 50 ns. Individual periods starting from 2.5 ns can be captured.
This is not a back-to-back measurement, meaning that there is a dead-time of 50 ns between the samples if High-Speed license is installed, 1 us otherwise. 2 signals can be measured in parallel. Up to 16 million samples total can be measured in a single measurement session.
Unlike Frequency/Period Average and their Smart versions, Period Single doesn’t use wide hysteresis. So only one comparator is used on A, B, D, E and Trigger Mode Auto sets trigger level to the middle of signal voltage range (Relative Trigger Level 50%).
Please, note: when measuring non-continuous signal or single cycles, Trigger Mode Auto/Relative might fail to find proper trigger level. One need to fall back to Manual Trigger Mode in this case.
1.4. Frequency Ratio
In Frequency Ratio mode, the instrument measures Frequency Average of up to 4 signals in parallel and then divides the resulting frequencies of any two signals. Since the instruments measures on four input channels in parallel, two frequency ratio values can be displayed in parallel.
This is back-to-back measurement with no dead-time between the samples. 4 input signals can be measured in parallel. Minimal sample interval is 50 ns if High-Speed license is installed, 1 us otherwise. Up to 16 million samples total can be measured in a single measurement session.
2. Time Interval and Phase
2.1. Time Interval
This function allows to measure phase delay between clock signals with the same nominal frequency. At least N + 1 signal cycles are needed on each measurement input to get N Time Interval samples.
Time intervals in the range [-1000 s .. 1000 s] can be measured. Resulting values are normalized to always be in the range [-0.5 * Period .. Period].
4 input signals can be measured in parallel. Minimal sample interval is 50 ns. Up to 16 million samples total can be measured in a single measurement session.
Please note: Time Interval Continuous is not suitable for measuring time interval between single shot events, use Time Interval Single instead.
When using Sample Interval close to 50 ns, actual sample interval might vary between 50 ns and 100 ns. In case of Time Interval measurement – it can result in generating less samples than was ordered by Sample Count setting. To avoid this please consider setting Sample Interval to 0 if you need minimal sample interval between samples. Setting Time Interval to 0 disables Sample Interval clock, meaning taking samples as fast as possible (close to 50 ns if signal period is 50 ns or greater).
2.2. Accumulated Time Interval
Accumulated Time Interval is useful for comparing phase delay between signals with the same nominal frequencies, but when frequencies of individual signals have small constant offset to each other. Time Interval will gradually increase over time and then drop after reaching value equal to signal Period, thus forming a sawtooth like graph. Accumulated Time Interval corrects this by adding or subtracting signal Period to Time Interval values when necessary. Other than that, it is exactly the same measurement as Time Interval Continuous.
As can be seen on Figure 26 and Figure 27, Accumulated Time Interval gives much better view on relative clock drift over time.
2.3. Time Interval Single
This function should be used to measure Time Interval between single events. Sample Interval setting is discarded, samples are captured as fast as it is possible.
Time intervals in the range [-1000 s .. 1000 s] can be measured. Resulting values are not normalized.
4 input signals can be measured in parallel. Minimal sample interval is 50 ns. Up to 16 million samples total can be measured in a single measurement session.
2.4. Measuring Time Interval between different trigger points of the same signal
Thanks to the presence of 2 comparators on each of A, B, D, E inputs it is possible to measure Time Interval, Accumulated Time Interval, Time Interval Single, Phase and Accumulated Phase between 2 trigger points inside the same signal.
This can be useful for measuring intervals inside multi-level a signal, e.g. TDR measurement.
Another example is measuring time interval from start on input A positive slope to input A negative slope to input B positive slope to input B negative slope. This can be achieved by selecting Time Interval A, A2, B, B2 and specifying trigger levels and slopes accordingly (see Figure 30).
However, Time Interval A to A to A to A is not possible since that would require 4 different trigger conditions on input channel A, while only 2 comparators are present.
2.5. Phase
Phase is similar to Time Interval but with phase delay expressed as angle. This measurement assumes same nominal frequency on all measured inputs. At least N + 1 signal cycles are needed on each measurement input to get N Phase samples.
During this measurement, the Analyzer estimates continuous Time Interval and clock Period and calculates Phase as following:
Phase= 360°×((Time Interval)/Period)
where
Time Interval=TSCH3-TSCH1Period= TSCH2-TSCH1
Resulting Phase values are normalized to always be in the range [-180° .. 360°].
2 input signals can be measured in parallel. Minimal sample interval is 50 ns. Up to 16 million samples total can be measured in a single measurement session.
The typical measurement case is to measure the phase shift in various electronic components or systems, for example, filters or amplifiers. In this case, the input A signal is the input signal to the filter/amplifier, and the input B signal is the output signal from the filter/amplifier. That means that the input A and B signals are typically sine waves, with exactly the same frequency per test point, and the phase should be constant with zero drift (per test point).
Another typical use case is to compare two ultra-stable signals from different sources, but with the same nominal frequency, and express their phase difference in degrees. Then the signal shape could be both sine or pulse, and there is a possibility for a small phase drift between the signals.
2.6. Accumulated Phase
The same as for Time Interval, there is an Accumulated version of Phase measurement function to ease drift visualization over time when clock signals in comparison have same nominal frequency with a slow phase drift. But it has no meaning for phase measurements on sources with a more erratic behavior, or when the two frequencies are not the same.
3. Time Interval Error (TIE)
Please note: license is needed to unlock TIE option.
TIE measurement uses continuous back-to-back time-stamping to observe slow phase shifts (wander) in nominally stable signals during extended periods of time. The measurement itself is performed the same way as Frequency/Period Average but different processing is applied.
TIE is only applicable to clock signals, not data signals. Monitoring distributed PLL clocks in synchronous data transmission systems is a typical application.
The nominal frequency of the signal under test can be either manually or automatically set. Auto detects the frequency from the first samples, and rounds to number of digits set by the user (5 by default). TIE is measured as the period deviation of the input signal from the “ideal” reference period, and the accumulated deviation, up or down, is calculated for each Sample Interval, and displayed.
4 input signals can be measured in parallel. Minimal sample interval is 50 ns. Up to 32 million samples total can be measured in a single measurement session. Resolution is 12 digits per 1 s of gate time (Sample Interval).
TIEA(B/D/E)(i) = TSCHx (i)-TSCHx (1)-(((EVENT_CNTCHx (i)-EVENT_CNTCHx (1))/F_ref)
Figure 36. TIE measurement
4. Pulse characterization
4.1. Positive and Negative Pulse Width
Positive pulse width measures the time between a rising edge and the next falling edge of the signal. Negative pulse width measures the time between a falling edge and the next rising edge of the signal.
The selected trigger slope is the start trigger slope. The Analyzer automatically selects the inverse polarity as stop slope.
This is not a back-to-back measurement, meaning that there is a dead-time of 50 ns between the samples. 2 signals can be measured in parallel. Up to 16 million samples total can be measured in a single measurement session.
4.2. Positive and Negative Duty Cycle
Duty cycle (or duty factor) is the ratio between pulse width and period time. The Analyzer determines this ratio by simultaneously making a pulse width measurement and a period measurement, and calculates the duty factor as:
This is not a back-to-back measurement, meaning that there is a dead-time of 50 ns between the samples. 2 signals can be measured in parallel. Up to 16 million samples total can be measured in a single measurement session.
4.3. Rise Time, Fall Time, Rise-Fall Time
By convention, rise/fall time measurements are made with the trigger levels set to 10% (start) and 90% (stop) of the maximum pulse amplitude. For ECL circuits, the reference levels are instead nominally 20 % (start) and 80 % (stop). In this case one can use Relative Trigger Levels mode and set trigger levels to 20% and 80% respectively.
These are not a back-to-back measurement, meaning that there is a dead-time of 50 ns between the samples.
Rise Time and Fall Time functions can measure up to 2 signals in parallel with up to 16 million samples per session total.
Rise-Fall Time function can measure only 1 signal but provides both rise and fall time at once. Up to 8 million samples total can be measured in one measurement session.
4.4. Positive and Negative Slew Rate
Slew rate is the speed of voltage change on pulse positive or negative edge. Hence, Positive and Negative Slew Rate are based on Rise Time and Fall Time measurements, the following formulae are applied:1.1.1. Positive and Negative Slew Rate
5. Totalize
Totalize functions count the number of trigger events on Analyzer inputs. There are few modes Totalize can operate in. See next sections for details. In each of these modes the user can choose between Totalize, Totalize X+Y, Totalize X-Y or Totalize X/Y.
5.1. Manual Totalize
If neither Start nor Stop Arming is used, Totalize operates in so-called Manual Totalize mode.
In this mode RESTART button start counting trigger events, while HOLD/RUN button is used to pause and resume the counting. Sample Interval setting has no effect, samples are generated each 100 ms.
5.2. Timed Totalize
Setting Start Arming Source to Off and Stop Arming Source to Timer enables so-called Totalize Timed Mode.
In this mode RESTART button start counting trigger events for time duration set by Sample Interval (which defines the length of Totalize Gate). Single sample is generated after the end of the Gate.
5.3. Armed Totalize
If both Start Arming Source and Stop Arming Source are set, then start and stop arming events define start and stop of Totalize Gate for each Sample (Arm On setting is ignored, the Analyzer uses arming in Sample implicitly).
See 4.6 Arming for details on this Totalize mode.
6. Voltage
The Analyzer measures the voltage by searching the minimum and maximum signal levels. See 4.2 Voltage measurement principles for details.
Vmin, Vmax and Vpp functions allow measuring voltage on 4 inputs in parallel, Vminmax allows only one input but provides both – min and max – at the same time. Resolution is 1 mV, up to 16 million samples can be acquired in one measurement session.