Facultatea de Electronică şi Telecomunicaţii
Departamentul Măsurări şi Electronică Optică
/
COUNTER TIMER HM 8122
1.Introduction
Counter timer machines are devices that measure physical quantities on the size of time: period, frequency, time intervals, the two frequencies ratio, the number of units at a time, etc. In this class of products is included Programmable Counter Timer HM 8122, which is build by Hameg Company. Initially, the operation of such a device was based on the principle of classical count of pulses over a period of time, as in figure 1.
Fig.1 Counter timer principle
In time interval T2, all the pulses with period T1 are counted.The number of pulses can be noted as Nx, then:
Nx = T2/T1 (1)
or
Nx=T2f1 (2)
In (2), if the period T2is known, you can find frequency, f1, and if you know frequency f1you can find the period, T2. In this case measurements are made with a sampling error equal with ±1/Nx, because in the time T2 does not always fit a whole number of pulses. When low frequencies are measured, Nx has a small value and the sampling error could have large values and this thing is unacceptable.
To avoid this disadvantage, the device presented in this paper uses the principle presented below with one change. The idea is to increase the time T2 until it contains an integer number of periods T1. For this to be possible, the pulse with duration T2 is triggered by an edge (eg. the rising edge) of the signal with period T1. The Pulse width T2 is established at the beginning. When this period ends, the signal is maintained on high until the first edge of T1signal (identical to the one that triggered the T2 pulse – the rising edge in our example) appears. Obviously, this increase in the pulse width can not be done in infinitely small steps. At Hameg HM8122 this step is 10 ns, corresponding to the frequency of 100Mhz, which is the internal clock.This principle is presented in figure 2.
The new pulse width is referred as T2' in the following equations:
T2' = Nx' Tc, (3)
T2' = NxT1 (4)
where, Tc is the step of 10 ns, and Nx' represents the number of Tc pulses from the T2' period. So, through Nx' ,the device can calculate the T2, Nx' is not affected by sampling errors. Because T2' can’t be modified in infinitely small steps in (4) Nx is affected by a sampling error equals to Tc/T2'. At HM8122 T2' has a minimum value of 1ms, the maximum sampling error is 10-7.
Fig.2. Counter Timer modified principle
Always T1 is the signal to be measured. Using (4), the frequency or the period of this signal may be calculated:
f1= Nx/( Nx' Tc) (5)
T1= (Nx' Tc)/Nx (6)
The resolution is the smallest variation of measured signal that could be visible on device’s display or the smallest unit of the displayed result. The result from (5) or (6) must be truncated to a number of significant digits to give the correct resolution.Fair resolution can be achieved through the interpretation of relations (5) and (6).
Example:
1. T2 is established smaller than T1.Using the principle below, T2will be increased to T2' until will be equal with T1, meaning that Nx =1, and T1 = Nx' Tc, so the device can distinguish variations of 10 ns of the measured signal. Tc = 10 ns is the measurement resolution. It follows that the result should be displayed with a number of digits corresponding to theresolution of 10 ns.
2. If T2 is established to a value of 10 periods of measured signal, T2 will be increased at a value of 10* T1 , or Nx =10. The result will be T1 = Nx' Tc/10, the measured resolution is Tc/10 = 1ns and the displayed result will have an extradecimal digit.
Similar reasoning can be made in the case of measuring frequency.
Conclusion: the resolution is betterby increasing the width T2, called “gate time”.
The measurement error of counter has two components:
- sampling error of Nx , denoted ec which in absolute value is the resolution
- time-base error eo, is the quartz oscillator error; for HM8122 is equal with 5x10-7
Total error is:
e = ±(ec+eo) (7)
2. Universal Counter HM 8122 specifications
The device has 3 inputs for measuredsignals, marked A, B, C. Some functions of the device use only a signal and other functions usetwo signals, the entries from A and B.
The measured parameter is displayed on a 9-digits display (function of resolution not all can be visible). The measurement unit is indicated by a lit LED (Hz or seconds), located below display and if necessary the exponent (in the right of display).
The function selection is made with one of the two FUNCTION pushbuttons.
One of the parameters required for almost all functions is the gate time. Setting it is made by potentiometer called GATE, at a value between 1 ms and 10s. By pressing READ pushbutton the measurement is stopped and gate time is displayed and could be modified. By pressing READ pushbuttonagain the measurement can be continued. Changing the gate time can be done during normal measurement, but you don’t know its value. Like mentioned below the real time of measurement differs from that set, being generally higher, depending on the period of measured signal.
Frequency measurement can be made using FREQ A or FREQ B functions for signals with frequency between 0 and 150 MHz (the signal is applied to terminal A or B), or using FREQ C functionfor signals with frequency between 100MHz and 1.6 GHz (the signal is applied to terminal C).
Period measurement is made using function PER A.
Frequency ratio measurement can be made using RATIO A:B function. The signal with higher frequency is applied to the input A. COM A/B pushbutton must be off.
Time interval measurement is made using TI A->B or TI AVG functions. The time interval can be measured between two edges of the same signal, in which case the signal is applied to the A input and COM A/B pushbutton is push connecting the two inputs together. If the pushbutton COM A/B remains off, the time may be determined by two fronts of different signals (the signals must be synchronous!). In both cases the active edge is selected from the proper switch (“” for rising edge and “” for falling edge). Always the selected edge of the signal at A input represent the beginning of the measurement and the selected edge of the signal at B input represent the end of the measurement. For this function, adjusting the gatetime has no effect, the gatetime is given by the two selected fronts. In this time interval, the device counts internal oscillator’s pulses (with period = 10ns), so the resolution is 10 ns. For a better resolution it can be useTI AVG function, which made a sequence of measurements in that time interval, after which made the arithmetic mean with the results. The number of measurements made is given by the gatetime relative to the period of the measured signal. With the increasing number of measurements, the resolution become better, may reach the value of 1ps for 25.000.000 measurements.
TOTL A function measure the number of pulses applied to A input, beginning from 0. The measurement can be stopped by pressing HOLD pushbutton and the value of the measured parameter is displayed. With RESET pushbutton, the result will be deleted and the measurement is re-run.
RPM A function display the value of fA*60/NPR, where fA is the frequency of the signal applied to A input (displayed in Hz), and NPR represents the number of pulses per revolution and is a number entered by the user between 1 and 65535. RPM means “revolution per minute”. In this way the number of revolutions per minute of an electric motor can be calculated based on the signal from a rotary transducer. NPR setting can be done by depressing READ, then by pressing ARM/EXT pushbuttonincrements the digit. Next digit can be selected by pressing FUNCTION pushbuttons.
With DC/AC pushbutton the signal measured can be measured directly or by a capacitor to eliminate the DC voltage component (offset).
With LEVEL potentiometer for both inputs, TTL signals can be generated from the input signals. The potentiometers must be rotated until the TRIG LED is blinking. If AUTO TRIG is pressed and DC/AC is set to AC, TTL are automatically generated, independently of the position of LEVEL potentiometer.
L. PASS 50 kHz pushbuttons related to inputs A and B eliminates any perturbations with higher frequencies of 50 kHz, and it is appropriate to use in measuring low-frequency signals.
When the OFFSET pushbutton is pressed, the signal value from that moment is saved as a reference and from that moment the difference from this value is displayed as a positive or negative number.
The two ATENUATOR 1:10 pushbuttons related to A and B inputs realize a signal attenuation of 10 times, or 100 times (if both pressed). This is used for high-amplitude signals.
The OFL LED is lit when the resolution is high and the result doesn’t fit on the display.
HM8122 has two operating modes when the measurement is made only when some external signals is applied:
- External Arming mode (ARM/EXT button is pressed until ARM/EXT LED is lit). Measurement begins when an external signal applied to ARMING/EXT.GATE back-panel terminal has a low-high transition.
- External Gate mode (ARM/EXT button is pressed until GATE/EXT LED is lit). The gate time is equal with the time in which the signal applied to ARMING/EXT.GATE back-panel terminal is on “high”.
3.Exercises
1. Signal from a signal generator is brought at the counter input. Check the functions FREQ. A, FREQ. B and PER. A. Adjust the gate time. What is happening?
2. Set the function RATIO A:B. First, you have to bring the same signal at both inputs (how can be done easier?). Then bring different signals at each input.
3. Establish the gate time to 1 ms. Bring a signal with 100 Hz frequency at the input A and measure the period. What value will have the gate time (see chapter 1)? Please verify this value by viewing on the oscilloscope the signal from the GATE VIEW OUTPUT on the back front of the instrument (at this output are available square pulses that hade the width equal to the real gate time). What is the resolution in this case (see chapter 1)? Check it practically. What value must be set for the gate time to obtain a resolution of 1 ns?
4. Using the function TI A->B measure the pulse width, the pause interval between the pulses and the period for a pulse waveform (just one signal!). After, measure the period by the dedicated function and compare the results. What is the resolution in each case? How the resolution can be improved?
5. Test the function RPM A. First, you must theoretical calculate the result. You have to know the frequency for the input signal and to introduce the NPR value. Compare the result with the value displayed by the instrument.
6. By using the function TOTL A find a way to implement the counter principle described by figure 1 and relation (2) for measuring the frequency. For T2 establish the value 10 s, respectively 100s by measuring time with a clock. For each case calculate the sampling error. Then measure the frequency by the dedicated function. What are the conclusions?
7. Determine the sensitivity for the input A. Bring to the input a sine wave. Decrease the amplitude of the signal until the counter doesn’t display anymore the correct value for the frequency (the LED corresponding to input A is not blinking anymore!). Then increase slowly the amplitude until the frequency is correctly displayed. The sensitivity represents the minimum value (Vrms – the root mean-square voltage) of the input voltage for which the counter still measure correctly. Use an oscilloscope to measure the voltage.
8. Take a single pulse from a pulse generator and view it on the oscilloscope. For this pulse, measure the pulse widthwith the counter.
9. The same single pulse is connected to the ARMING/EXT.GATE input. What is happening in the external arming mode?
10. Find a situation when overflow occurs for measuring the frequency. Explain the overflow by evaluating the resolution depending on the gate time and the signal period.
.
1