Agilent Technologies Electronic Counter User Manual

This literature was published years prior to the establishment of Agilent Technologies as a company independent from Hewlett-Packard
and describes products or services now available through Agilent. It may also refer to products/services no longer supported by Agilent.
We regret any inconvenience caused by obsolete information. For the latest information on Agilent’s test and measurement products go to:
www.agilent.com/find/products
Or in the US, call Agilent Technologies at 1-800-452-4844 (8am–8pm EST)
Fundamentals of Time Interval
Measurements
Application Note 200-3
Electronic Counter Series
1
2
Table of Contents
Precision Time Interval Measurement
Using an Electronic Counter ..................................................... 5
Preface ........................................................................................................... 5
Time Interval Measurement Using an Electronic Counter ........ 6
Introduction .................................................................................................. 6
What can be Measured ................................................................................ 6
How Measurement is Made ......................................................................... 7
Resolution ...................................................................................... 8
One-Shot Measurements ............................................................................. 8
TI Averaging .................................................................................................. 8
Minimal Interval, Dead Time and Pulse Width ......................................... 9
Start and Stop Signal Input Channels ....................................... 10
General ........................................................................................................ 10
Desirable Characteristics .......................................................................... 10
Controls Associated with Time Interval Measurements ....................... 11
Input Circuit Operation as it Affects the User ........................................ 11
Input Signal Conditioning Controls and
Trigger Circuit Operation ........................................................ 13
Signal Conditioning Controls Set the Trigger Point............................... 13
Other Input Controls .................................................................................. 13
Trigger Operation ....................................................................................... 16
Determining the Hysteresis Window and Triggering at Zero Volts ..... 24
Hysteresis Compensation.......................................................................... 27
Polarity Control .......................................................................................... 29
Input Attenuators for Measuring Higher Amplitude Signals ................ 30
Trigger Lights .............................................................................................. 30
Markers ........................................................................................................ 31
Delay ! Control ......................................................................................... 34
Time Interval Averaging ............................................................. 36
Reduce +1 or –1 Count Error by N on Repetitive Signals ................ 36
Time Interval Averaging is Useful When ................................................. 36
Synchronizers Needed for True TI Averaging ........................................ 37
Extending Time Interval Measurements to Zero Time .......................... 38
Disadvantages of Time Interval Averaging ............................................. 39
Time Interval Error Evaluation ................................................. 40
+1 Count Error or –1 Count Error ........................................................... 40
± Trigger Error ........................................................................................... 40
± Time Base Error ...................................................................................... 48
± Systematic Error ..................................................................................... 49
Time Intervals by Digital Interpolation .................................... 50
Digital Interpolation ................................................................................... 50
Phase-Startable Phase-Lockable Oscillators (PSPLO) .......................... 50
The Dual Vernier Method .......................................................................... 50
3
The Generation of Precise Time Intervals ................................ 52
Phase-Startable Phase-Lockable Oscillators .......................................... 52
Time Interval Measurement Using the
HP 5363B Time Interval Probes ............................................... 54
Solve Troublesome TI Measurements Problems .................................... 54
Level Calibration ......................................................................................... 56
Time Zero Calibration ................................................................................. 58
To Make a TI Measurement Using the HP 5363B .................................... 59
Some Applications of Time Interval Measurements .................. 60
Simple Timing System with Start-Stop Pulses Generated by
Mechanical Switches ............................................................................... 60
Phase Measurement .................................................................................... 61
Measuring Complex Pulse Trains ............................................................. 63
Comparison with Other Ways of Making Time Interval
Measurements on Narrow Pulses or Fast Rise Signals .......... 67
4
Precision Time Interval Measurements Using
an Electronic Counter
Preface
A time interval measurement is a measurement of the elapsed time
between some designated START phenomena and a later STOP
phenomena. This is in contrast to real-time observations (time of day)
used in our day-to-day living to schedule meetings or transportation,
in astronomical observations and for celestial navigation among other
things. One might make a time interval measurement with a mechanical stopwatch as when timing a track meet or other sporting event or
in making time and motion studies. With increased speed of the timed
object as when timing automobiles or airplanes the timed interval
becomes shorter and shorter until the human factor involved in
determining when to start and when to stop the measuring device, a
stopwatch or clock for instance, begins to introduce significant error.
Mechanical, optical, or electrical transducers or a combination of all
were developed to reduce this error. Finally with advances in many
scientific fields, mechanical and electrical time measurements were
required which were beyond the resolution of a mechanical stopwatch. This led to the development of a time interval measuring
electronic counter, in essence an electronic stopwatch. A time
interval counter can measure electrical delays, pulse widths, and
other time related electrical phenomena required in the development
and maintenance of communications, navigation, television, and other
present day systems. Increased measurement capability has helped
bring on more and more sophistication in all of these fields until now
modern electronic time interval counters are used to measure electrical events spaced as close as 0.1 nanosecond (the time required for
light to travel 3 centimeters) on a “one-shot” basis. Time interval
averaging on repetitive events gives still greater resolution than this.
5
Time Interval Measurement Using an
Electronic Counter
Introduction
Time Interval is an important measurement frequently made with
electronic counters. In this role, the counter makes an elapsed time
measurement between two electrical pulses, Figure 1, just as a stopwatch is used to time physical events.
Gate Closes
Gate Opens
Start
Stop
Gate
Open
Clock
Accumulated
Clock Pulses
Accumulated
Count
Figure 1. In a
time interval
measurement,
clock pulses are
accumulated for
the duration the
main gate is
open. The gate is
opened by one
event, START
and closed by the
other, STOP.
Minimum time measurement is much less (to a nanosecond and
below) than possible with a stopwatch. Also resolution and accuracy
are much greater than attainable with a stopwatch.
What Can Be Measured
Some typical time measurements that might be made are:
Characterization of Active Components
Propagation delay of integrated circuits
Radar Ranging
Nuclear and Ballistic Time of Flight
Pulse Measurements
Width
Rise Time
Repetition Rate (Period) of pulse train
Spacing on complex pulse trains such as used by airborne
identification and navigation systems
Cable Measurements
Propagation Time
Cable Length
Phase
Delay Line Measurements
Time interval measurements can also be made on any physical
phenomena that can be translated into appropriate electrical signals.
Transducers such as photo electric cells, magnetic pickups, strain
gauges, micro-switches, bridge wire systems, or thermistors can be
used to translate physical events into the electrical start and stop
signals required for a time interval measurement.
6
How Measurement is Made
The START pulse, received at channel A of the counter in Figure 2A,
opens the GATE to start the measurement, the STOP pulse occurring
later in time and received at channel B closes the gate to end the
measurement. Elapsed time between start and stop is measured by
counting the Time Base “clock” frequency while the gate is open.
The resolution of a conventional time interval counter (HP 5328A,
HP 5345A, etc.) is determined by its “clock” frequency. A clock frequency of 1 MHz gives 1 µsec resolution, 100 MHz gives 10 ns resolution,
500 MHz gives 2 ns resolution and so on. Clearly, the elements within
the time interval counter (input amplifier, main gate, DCAs) must
operate at speeds consistent with the clock frequency; otherwise the
instrument’s resolution would be meaningless. Present state-of-the-art
limits resolution to about 2 nsec, although special techniques can
improve on this.
Electronic Totalizer
Channel A
Other Controls Same
in "B" Channel
3 2 5. 7
Start
Input
Input Amp/Trigger
Reset
µsec
Gate
Com
Channel B
Stop
Input
Sep
On
Chan B
Trig Level
Off
Input Amp/Trigger
ac 50Ω Atten Slope +
×1
dc or
Pos Level
1Meg ×10 Neg –
×20
a. Block Diagram for Basic Time
Interval Counter
1. START and STOP trigger circuits
open or close a gate which in its
elementary form is merely an ONOFF switch connected between the
clock pulse source and the
electronic pulse counter.
2. The counted clock frequency is
typically 0.1 Hz to 100 MHz in
decade steps or 500 MHz each
derived from a crystal oscillator.
Period = 0.1µsec (10 MHz)
Time Base
"Clock"
b. A modern universal counter having
2 ns Time Interval resolution for one
shot measurements. It also does True
Time Interval averaging for higher
resolution.
c. HP 5328A Opt. 40 General
Purpose Counter
Figure 2. Measuring Time Interval with an Electronic Counter
7
Resolution
One-shot Measurements
Most general purpose counters will make a “one-shot” (the time between a single pair of start and stop pulses) time interval measurement
with resolution to 100 nanoseconds — i.e., the counter counts a 10 MHz
clock. The HP 5328A offers either 100 ns or 10 ns resolution depending
on the configuration. The HP 5345A will resolve a one-shot time interval
measurement to 2 nanoseconds. By way of reference 2 nanoseconds is
the time it takes light to travel six tenths of a meter.
For conventional counters, direct readout is achieved by using clock
frequencies related by powers of 10 — i.e., 1 MHz, 10 MHz, 100 MHz,
etc., (period of 1 µs, 100 ns, 10 ns, respectively) and a correctly placed
decimal point and annunciator. Single shot resolution of conventional
counters is limited to 10 ns as the next step up, 1 ns resolution, requires
a 1 GHz direct count decade which at present is not economically
feasible. Counters which have arithmetic capability are not limited in
this way as the measurement can be made with any convenient clock
period then translated to engineering units before being displayed. The
HP 5345A, a reciprocal counter, counts in 2s of nanoseconds then does a
multiply by 2 before displaying a time interval measurement. Some
sophisticated modern counters like the HP 5370A operate on a digital
interpolation scheme which allows single shot resolving capability of
20 ps. However, with resolution this high other factors like noise in the
input amplifiers or on the input signal become limiting factors. Perhaps
a realistic way to look at resolution would be to say it is the probable
repeatability from measurement to measurement for a given set of
circumstances. Since the two noise components limiting resolution are
statistical in nature the resolution must be described in statistical terms.
For example, 30 ps rms would be a typical description of resolution
when using the HP 5370A.
TI Averaging
Time interval averaging can be used to get resolution to the picosecond
(10–12 sec) region on a repetitive signal. Averaging operates on the
assumption that the factors limiting resolution are random in nature and
will tend to average towards zero. A counter needs synchronizers in gate
circuits and a noise modulated clock to achieve TRUE TIME INTERVAL
AVERAGING with accuracy and repeatability independent of the input
signal repetition rate. The HP 5345A and HP 5328A with Option 040
Universal Module both have this true averaging capability.
8
Minimum Interval, Dead Time and Pulse Width
Three important specifications are sometimes overlooked when considering time interval measurements.
1. The minimum time interval or minimum range specification is the
minimum time between start and stop pulses which the counter will
recognize. For single shot measurements in a conventional counter
this time must always be one or more clock periods. However, if
interpolation is used this time can be reduced to in the region of
20 ps. A more typical specification is 100 ns corresponding to the
period of a 10 MHz clock. Another technique for reducing the minimum time interval is to use averaging with synchronizers. This allows
intervals of less than one clock period to be measured but a repetitive
signal is required.
2. The minimum dead time is the time from a stop pulse to the acceptance of the next start pulse. Typical dead time specifications are
10 ns for the HP 5345A, 150 ns for the HP 5328A. Dead time determines the maximum upper repetition rate of an acceptable signal.
3. The minimum pulse width is the shortest pulse the counter will
recognize as a start or a stop pulse and is largely determined by the
bandwidth of the input amplifiers. The typical minimum pulse width
for a 50 MHz counter is 10 ns or the period of half a cycle.
Some measurement errors may result if these specifications are not
considered. For example, if a rise time is being measured which is less
than the minimum time interval specification the first stop to be recognized will be on the next pulse giving a measurement result corresponding to the pulse period instead of the desired rise time.
9
Start and Stop Signal Input Channels
General
High resolution is meaningless if measurements on a stable signal are
not repeatable as only the digits that consistently repeat represent
accurate information. Since the input amplifier-trigger circuits of the
counter are the interface from the signal of interest to the counter they
are the most critical circuit elements in accurate time interval measurements. Their performance directly influences measurement accuracy.
Lack of attention to these circuits as related to the measurement is the
prime source of measurement error and the major reason a counter’s
potential accuracy is often not achieved.
The input amplifier and trigger circuits, one for the start channel and one
for the stop channel, establish the voltage level at which an input signal
will trigger the counter. Noise, drift, ac-dc coupling, and other factors
relating to these circuits all influence the measurement. Since these
circuits are so important it is worthwhile looking in some detail at the
operation of one of these input channels.
Desirable Characteristics
Several requirements must be met by each input if a time interval
counter is to make useful measurements:
1. The input circuits must be able to accept a signal which might be a
sine wave, square wave, pulse, or a complex waveform of varying
amplitude and generate from that signal one and only one output
pulse of constant amplitude, rise time and width for each cycle of the
input.
2. The circuit will need controls to let the operator choose the exact
voltage point on the input waveform at which he wants to START and
STOP his measurement. This is necessary to achieve flexibility of
measurement.
3. The input should have a means of externally setting and/or measuring
the trigger point voltage to facilitate setting up a measurement.
4. The input needs good stability with time and temperature and low
internal noise so that once set, triggering will occur at the same
voltage level regardless of input signal amplitude, wave shape, or
duty cycle.
5. The input should be dc coupled so the trigger voltage point will not
change with repetition rate or duty cycle of the input signal yet be
capable of ac coupling for measurements on signals with a dc offset.
10
6. Input circuit protection is necessary so inadvertently applied high
amplitude signals, regardless of duration, will not damage input
circuit components.
7. High input impedance (high input resistance and low input capacitance) is desirable for bridging measurements (connecting directly
across an input signal) with minimum input waveform distortion yet
switchable to 50 ohms to provide a good termination and thus prevent
reflections when doing fast pulse work in a 50 ohm environment.
8. Provision to connect the start and stop inputs together is desirable to
simplify time interval measurements on a signal appearing on a single
cable. This is necessary for measuring pulse width for instance.
9. Matched input amplifiers are a necessity for meaningful time interval
measurements on fast rise time or high frequency signals. If one
channel has significantly less bandwidth than the other, propagation
delay and rise time will be vastly different for the two. This introduces large errors in measurements involving high frequency signals.
Controls Associated with Time Interval Measurements
Slope, level and attenuator controls which determine the trigger point on
the input waveform give the operator flexibility in setting up a measurement. Understanding the function of each control is important or triggering may not occur at the expected voltage point on the input signal.
Input Circuit Operation as it Affects the User
All electronic counters have an input sensitivity specification, i.e.,
100 mV rms for sine waves (282 mV peak-to-peak), which indicates the
minimum voltage necessary to operate the counter. This specification
normally applies over the full environmental range as well as takes into
account aging effects, therefore when operated at a moderate room
temperature, sensitivity may be significantly better than the specification. Sensitivity may change with aging, with ambient temperature or
other environment changes; however, for a well designed circuit these
effects are held to a minimum. Also sensitivity may depend on frequency.
For frequency measurement, selecting the trigger point is not too critical,
the sole requirement being that the counter trigger once (and only once)
for each cycle of the input signal. Accurate time interval measurement
however, places a much more exacting requirement on the input circuits
as they must trigger precisely at the selected trigger voltage set up by the
input controls.
11
Time interval is a two dimensional problem. The dimensionality of the
time interval measurements is illustrated by the simple example of
Figure 3, measuring signal rise time.
0.9 V
V
0.1 V
Start
Stop
tR
Figure 3.
Measuring the rise
time tR by
adjusting the
trigger levels to the
10% and 90%
points of the input
amplitude.
The time interval meter must generate a START signal at the 10 or 20%
amplitude point of the input signal and generate a STOP signal at the
80 or 90% point. Clearly this is different from the frequency or period
measuring case when the input triggers at the same point on the waveform from cycle-to-cycle of the input. Inherent in the time interval
measurement, therefore, is the dual dimensionality, amplitude and time.
It is this dimensionality that places much more stringent requirements on
the input amplifier-triggers than those necessary for the measurement of
frequency or period.
Any deviation from the selected trigger point because of circuit drift
degrades the accuracy of the measurement as the counter time base
“clock” will not be counted for the proper interval. Obviously, if the clock
used for timing the electrical events is not started or stopped at the right
moment, the measurement will be incorrect just as measurements made
with a stopwatch if it is not started or stopped at the right moment.
12
Input Signal Conditioning Controls and
Trigger Circuit Operation
Signal Conditioning Controls Set the Trigger Point
Figure 4 shows the effects of the SLOPE, POLARITY, LEVEL, and
ATTENUATOR controls in establishing the trigger voltage point on the
input signal.
1. The SLOPE control determines whether the trigger point will be
on a rising or a falling voltage as in Figure 4a.
2. The POLARITY control determines whether the trigger point is
positive or negative with respect to zero volts as in Figure 4b.
3. The LEVEL control adjusts the trigger point of the circuit up or
down in voltage and usually has a range of from one to three volts
peak for a counter with 100 mV rms (282 mV peak-to-peak) sensitivity as in Figure 4c. The polarity and level functions are often combined using a zero center variable control having a range of –3 volts
to 0 to +3 volts.
Most counters also have a PRESET switch position at one end of the
level control range to set up the most sensitive trigger condition for
ac coupled symmetrical input signals. Functioning of these controls
as they relate to setting a trigger point are discussed in detail later.
4. The INPUT ATTENUATOR reduces high amplitude input signals
up to 100 volts or more so these fall within the dynamic range of the
amplifier/trigger circuits which are limited to a few volts rms
maximum as in Figure 4d.
Other Input Controls
1. SEPARATE COMMON Switch
A SEPARATE common switch ties the START and STOP inputs
together without having to resort to external cables or hardware. On
counters with 50Ω inputs this is done using appropriate matching
networks so the input looks like 50Ω for either the SEPARATE or
COMMON mode of operation. Depending on the circuit configuration this may or may not result in a 2:1 loss in voltage sensitivity.
13
2. 50 OHM-HIGH IMPEDANCE Switch
Some modern counters have a panel switch to select a high input
impedance (1 meg, 35 pF) for bridging applications or 50 ohm input
impedance to provide a good termination (low VSWR) for a 50 ohm
transmission line. If the counter has a 50Ω position the whole input
circuit up to the gate of Q1 (Figure 6) is designed as a 50 ohm strip
line. Also, the overload protection resistor R1 is shorted out so the
operator must be more careful when measuring high amplitude
input signals (usually 5V rms maximum) or the input circuit can be
damaged.
3. dc-ac Coupling
All general purpose time interval counters have a dc coupled
amplifier trigger circuit so as to maintain a consistent trigger point
on input signals down to zero frequency. AC coupling, when needed,
is achieved by connecting a capacitor, C1, in series with the input
connector either with a switch or through a second input connector.
AC coupling is necessary when measuring a signal with a large dc
offset; however, the trigger point changes with both the input
frequency and duty cycle when using ac coupling.
4. CHECK
While not strictly related to time interval measurements, the SELF
CHECK function checks the multiplier, divider, and gate circuits of a
counter for correct operation and should be done before using a
counter. The self check function does not give any indication of
crystal oscillator accuracy.
14
+
+
0
0
Volts
A. Slope
–
B. Polarity
TRIGGER ON
POSITIVE SLOPE
• Rising voltage
• Independent of polarity
–
TRIGGER ON
NEGATIVE SLOPE
• Falling voltage
• Independent of polarity
+
0
0
–
TRIGGERING ON
PLUS POLARITY
• Triggers at some voltage
above zero
TRIGGERING ON
NEGATIVE POLARITY
• Trigger at some voltage
below zero
+ Overload
+3
C. Level
0
–3
– Overload
D. Attenuator
• Triggering can be set to occur at any
level within the dynamic range of the
input circuit
+100
–10
×1
×10
+1
0
–1
×100
Input Volts
+10
Voltage to input amplifier
–100
Figure 4. The three parameters under operator control which define the trigger point on an input signal.
15
Trigger Operation
The input amplifier trigger circuit accepts the input signal which may
vary in amplitude, frequency, and wave shape. It puts out one pulse of
constant amplitude and width as required by the internal counter
circuits each time the input signal crosses the selected trigger voltage
point.
1. HYSTERESIS LIMITS define input sensitivity
The input signal must cross two voltage thresholds to activate the
trigger circuit. The sensitivity of the electronic counter is determined by the voltage difference between these two thresholds,
called hysteresis limits, which define the hysteresis window of the
trigger circuit. The hysteresis limits correspond to voltage levels on
the input signal, one of which will trigger the circuit Figure 5a, at
(m) and the other voltage level which will reset the trigger circuit at
(n). A plot, Figure 5b, of the transfer function of the trigger output
Hysteresis window defined
by upper hysteresis limit VU
and lower hysteresis limit VL.
Positive
Slope
+0.3
Input Volts
+0.2
m
+0.1
VU
VC
0
–0.1
n
–0.2
VL
Input Signal
Triggering on positive slope
Triggering on negative slope
–0.3
V2
V1
Output Voltage
Trigger
Reset Trigger Circuit
Output Pulse from Trigger Circuit
a. Triggering on a positive slope. Signal must cross both
hysteresis limits to activate the trigger circuit.
Output Volts
Lower Hysteresis Unit
Hysteresis
Window
Upper Hysteresis Unit
V2
V1
–0.2 –0.1 0 +0.1 +0.2
b. A plot of the transfer function of the input trigger circuit of an
electronic counter resembles the BH curve of a magnetic core.
Figure 5. Hysteresis limits and transfer function of a trigger circuit.
16
J1
Trigger
Light
+5
AC
Parasitic Capacitance
CR1
C1
1:1
J2
R1
DC
S1
CR2
10:1
C2
R6
C4
R7
C5
50Ω
R3
1 Meg
R2
1 Meg
dc
Amp
50Ω
Pulse
Stretch
C3
Q1
–5
Attenuator
+
S1
To Counter
S3
Slope
+3
Q2
Trigger
Level &
Polarity
Actually
Determines
Both Polarity
And Level
R4
dc
Amp
Marker
Gen
+
VC
–
S2
Preset
+
B1 X
–
– B2 +
VC'
For Positive Slope Hysteresis
For Negative Slope Comp
–
R5
J3
Trigger
Volts
J4
Marker
Output
Figure 6. Symmetrical Input Amplifier Trigger Circuit for Time Interval Counter.
voltage resembles the familiar BH curve (flux density vs. magnetizing force) or hysteresis curve of a magnetic core. Even though these
hysteresis limits define the trigger and reset voltage points on the
input signal they do not exist as nodes (component junctions)
anywhere in the circuit so the trigger voltage point cannot be
measured directly. There is, however, a voltage VC approximately
half way between the hysteresis limits that does exist as a physical
circuit node which can be measured with a dc voltmeter to determine the trigger point. This will be discussed later. Once the circuit
has triggered, it cannot be retriggered until the input signal crosses
the opposite hysteresis limit then once more crosses the level of
interest.
It is important to note that if the signal crosses only one limit then
returns to its original level the circuit will not trigger. The signal
must pass all the way through the hysteresis window to cause either
triggering or reset of the input trigger circuit. Hysteresis limits are
defined by peak voltages, however most counter sensitivities are
specified in rms volts; therefore, the specified rms sensitivity must
be multiplied by 2.82 to convert to peak-to-peak volts to get an
indication of input sensitivity for other than sine wave inputs.
17
This triggering action might be compared to a mouse trap. With the
trap, nothing happens until the trigger is depressed below a certain
point at which time the trap is sprung. Operation of the trap once
tripped is independent of how fast or how slow the trigger was
depressed. Once sprung, further movement of the trigger has no
effect until the trap is reset. Triggering the trap corresponds to
crossing the upper hysteresis limit (m) of Figure 5a, resetting the
trap corresponds to crossing the lower limit (n) for this example.
2. TRIGGER CONTROLS as they relate to the input circuit
The diagram, Figure 6 of a typical input amplifier for one channel of
a solid state time interval counter shows the circuit elements which
are directly influenced by the setting of panel controls. A look at the
circuits associated with each control helps understand the correct
setting procedure needed to make valid time interval measurements.
a. Input Attenuator
The frequency compensated input attenuator, R6R7C1C5, reduces
an input level up to 100 volts or more by a factor of 100:1, 20:1,
10:1, or 2:1 (sometimes labeled x100, x20, x10, x2) to a level that
can be safely applied to the input amplifier circuit. One usually
thinks of an attenuator as a device that reduces the input signal
to the linear range of the input amplifier. With respect to the
signal, another way to look at attenuator operation is that it
multiplies the hysteresis window of the counter by the attenuation factor. For example, the counter with a 25 mV rms sensitivity
(hysteresis limits 25 × 2.82 = 70.5 millivolts apart) would have
250 mV rms sensitivity (hysteresis limits 25 × 2.82 × 10 = 705
millivolts apart) on the X10 attenuator setting. Even though large
signals applied to a sensitive range may not damage the counter
the overload may cause miscounting.
b. Overload Protection
Diodes CR1 and CR2 in conjunction with R1 provide overload
protection to prevent damage to Q1 in case of accidental overload. R1 is large enough to prevent damage with an input signal
as high as 115 volts rms at power line frequency on the most
sensitive attenuator range of most counters. A capacitor, C2,
across this resistor prevents sensitivity roll off at high frequencies. Important to the operator is the fact that at high frequencies, C2 effectively shorts out the protection resistor R1 so
maximum voltage is limited to a few volts rms, Figure 7, rather
than 100 volts or more as at low frequencies.
18
Volts Peak
150
5
60 Hz
Figure 7. Overload
voltage as a
function of
frequency.
1000 Hz
Freq
Also important to the operator is the fact that the protective
diodes CR1 and CR2 can change the input characteristics of the
counter. So long as the input signal is below ±5 volts peak the
diodes CR1 and CR2 for the circuit in Figure 5 are back biased so
have no effect. If the peak input signal goes beyond these limits
however, the input resistance of the counter drops from
1 megohm down to a value perhaps as low as a few hundred
ohms dependent on the value of R1. This places a heavy nonlinear load on the signal source which may drastically alter its
waveshape. For normal operation the input signal must be kept
below this overload level even though the input circuit may not
be damaged because double counting or other erratic counting
may occur due to the shape of the altered input signal. When
working with a transducer such as a tachometer generator which
has an output proportional to rotational speed, the simple
external limiter shown in Figure 8 is effective in preventing
counter overload for a signal that varies over wide amplitude
limits. When using this circuit, the source always sees a minimum
load of 22K at the input to the limiter so ringing and other
distortion is not a problem. When working with low frequency
sources (below 50 kHz) such as tachometer and flowmeter
pickups C1, in the range of 100 to 500 pF, keeps high frequency
noise from causing false triggering. The input signal is symmetrically clipped as amplitude increases so the trigger point of the
counter must be set between ±0.5 volts.
Input
From
Transducer
R1
22K
C1
500pF
Output
To Counter High
Impedance Input
CR1
CR2
Diodes 1N914 Silicon Diodes or 1901-0040
Figure 8. Simple
clipper circuit to
prevent counter
input overload.
19
c. dc or ac Coupling
For time interval measurements dc coupling is very important, as
triggering is generally required at a specific voltage point on the
input waveform. With ac coupling the location of the trigger point
varies with respect to zero volts dc anytime the pulse width,
repetition rate, rise time, or waveshape (any change that effects
average dc level) are changed. The shift in the measuring point is
shown in Figure 9 and Figure 10. Zero volts is defined in each
case by positioning the waveform such that the average voltage
above zero equals the average voltage below zero for a repetitive
signal. This also implies that for any symmetrical waveshape —
sinewave, triangle wave, square wave, etc. — zero volts will be at
the center of the waveform.
Use of ac coupling causes only simple translation of the waveform along the voltage axis as in Figures 9d, 10b, or 10d if the RC
time constant of the coupling network is long compared to the
period of the waveform of interest. When coupling circuit time
constants are of the same order as the period of the input signal,
the waveform is distorted as well as translated as in Figure 9e.
With yet shorter time constants (R1 decreased in value) the
circuit changes from a coupling to a differentiating network with
resultant signal transmission and waveform distortion as in
Figures 9f and 9g. In each instance the counter input sees quite a
different input signal although the generator output signal
remains unchanged.
These waveforms point up rather dramatically some of the
problems associated with making time interval measurements
using ac coupled amplifier/trigger circuits. If a counter had been
set to trigger on a positive slope at + zero volts dc for instance, it
would have triggered at the peak of the waveform Figure 9c if dc
coupled. If ac coupled it would have triggered near the middle of
the input signal, Figure 9d or near the top of some cycles and
near the bottom of others, Figure 9e, depending on circuit
constants. This shift is through no malfunction of the counter
input circuits but rather is due to the characteristics of the ac
coupling network. (These same factors cause the vertical shift of
an oscilloscope display if the duty cycle of the input signal is
changed in the ac coupled mode of operation.) AC coupling may
be unavoidable under some circumstances, as when the signal
has a large dc component, so it is up to the operator to recognize
the attendant problems and determine what signal the time
interval input is really seeing or the actual trigger point may be
far removed from the desired trigger point.
20
ac
Coupling
dc
+1V
Output
Input
C1
0.033 µf
0V
R1
–1V
b. Coupling Network
a. Input Signal
+1V
0V
–1V
c. dc Coupling Output Signal
d. ac Coupling R1 = 1 Meg
e. ac Coupling R1 = 47K
Time constant of the coupling network is
long compared to the period of the input
signal.
Time constant of the coupling network is
about the same as the period of the input
signal.
Note: The output waveform is preserved
but the dc level has shifted.
Note: The output waveform is distorted
and the dc level has shifted.
+1V
0V
–1V
c. dc Coupling Output Signal
(Same as "c" above.)
f. ac Coupling R1 = 10K
g. ac Coupling R1 = 2200 ohms
Time constant of the coupling network is
shorter than 5 milliseconds.
Time constant of the coupling network is
much shorter than the shortest period of
the input signal.
Note: The output waveform differentiated
only with respect to the low frequency
component of the input signal.
Note: The output waveform is completely
differentiated, i.e., the output voltage returns
to zero after each input transition.
Figure 9. dc and ac Coupling for a complex pulse train.
21
+1V
0V
–1V
a. dc Coupling
b. ac Coupling C1 = 0.033 µF, R1 = 47K
Time constant of the coupling network is long
compared to the period of the input signal.
Note: Average voltage of the input signal is near
zero so dc shift is small.
+1V
0V
–1V
c. dc Coupling
d. ac Coupling C1 = 0.033 µF, R1 = 47K
Same coupling network as for waveform above.
AC COUPLING CIRCUIT IS THE SAME AS FOR FIGURE 9.
Note: Average voltage of the input signal is approximately
–1 volt so dc shift is large when switching to ac coupling.
Figure 10. dc or ac Coupling illustrating change in Zero Level with Pulse Width.
d. Slope Control
The slope control, S3 Figure 6, determines if the circuit is triggered by a signal with a positive (+) slope (going from one
voltage to a more positive voltage regardless of absolute polarity)
to generate an output pulse at the upper hysteresis limit (VU) of
Figure 5a, or by a signal with a negative (–) slope which causes an
output pulse to be generated at the lower hysteresis limit (VL).
e. Level Control
Moves hysteresis window voltage-wise. The level control, R4
Figure 6 moves the center of the hysteresis window VC to any
positive or negative voltage within the dynamic range of the input
22
circuit without appreciably changing the window as in Figure 11.
Best sensitivity for an ac coupled, sine wave input signal is with
the hysteresis limits positioned symmetrically with respect to
0 volts dc since the smallest amplitude signal can now cross both
limits. Many counters have a PRESET position at one end of the
range of the LEVEL control to set up this condition. (Triggering
will occur at either the upper or lower hysteresis limit depending
on the setting on the SLOPE control.)
The voltage VC which defines the approximate center of the
hysteresis limits comes from the arm of the TRIGGER LEVEL
control, R4, Figure 6, so can be measured with a dc voltmeter.
This voltage is often brought to a panel connector, J3, for ease of
measurement. A resistor, R5, of several thousand ohms may be
included to prevent circuit damage if J3 is accidentally shorted;
therefore, a high impedance voltmeter should be used.
0.5
Peak Volts
0.4
+ 0.3
0.2
–
0.1
0
0.1
0.2
0.3
VU
VC
VL
VU
VC
VL
VU
VC
Hysteresis
Window
VL
0.4
0.5
Figure 11. Polarity
and Level Controls
f. Triggering at a Particular Voltage
To actually trigger at a particular voltage, either the upper
hysteresis limit, VU, or lower hysteresis limit VL, (once again
depending on slope) must be positioned at the desired voltage
level using the LEVEL control. This is not easy since VU or VL
cannot be measured with a voltmeter as mentioned earlier.
Instead, one must measure the hysteresis window peak-to-peak
voltage, Figure 12, then add 1/2 this value to VC when triggering on
a positive slope or subtract when triggering on a negative slope to
determine the actual trigger point.
VU – VL
for positive slope
2
V – VL
+ U
for negative slope
2
V Trigger = VC +
V Trigger = VC
Where: VC can be measured with a dc voltmeter.
The hyteresis window, VU–VL, can be determined
using procedures outlined in the following section.
23
Volts
+
–
0.2
0.1
0
0.1
0.2
Trigger Point
Positive Slope
VU
VC
VL
Hysteresis Window
Hysteresis
Limits
Figure 12. To
trigger at a
particular voltage
(zero volts dc in
this example) with
a positive slope.
Upper and lower
hysteresis limits
are positioned as
shown above.
Determining the Hysteresis Window and Triggering
at Zero Volts
The distance between the hysteresis limits VU–VL (hysteresis window)
which defines input sensitivity can be determined using one of these
methods:
1. Methods of Measuring of Hysteresis Window and Determining VC
a. The first method, Figure 13a, measures the hysteresis window by
counting a low distortion 10 kHz to 100 kHz sine wave input to
the counter. Reduce the input amplitude, readjust the trigger
LEVEL control, then repeat these steps to determine the minimum amplitude signal that will just trigger the counter. The
hysteresis limits are then spaced by the peak-to-peak sine wave
voltage. This is 2.82 × rms input voltage measured with an ac
voltmeter. A calibrated dc coupled oscilloscope and a sine wave
or a square wave generator could also be used in a similar
manner. In this case, the × 2.82 factor is not needed since an
oscilloscope already displays peak-to peak volts.
b. The second method, Figure 13b, is the inverse of the first method
in that the hysteresis limit of interest is moved through a zero
volt input signal to establish triggering at zero volts. For triggering on a positive slope, the counter input is first grounded
insuring zero volts input, then the trigger LEVEL control is
turned to its most positive extreme after which it is decreased
slowly until the input circuit just triggers. Triggering occurs when
the upper hysteresis limit coincides with the input voltage which
is zero because of the grounded input terminal. Voltage
V − VL
VC′ , U
and negative in this case, can be measured at J3
2
with a dc voltmeter. This voltage which is one-half the hysteresis
window is added to other VC settings to get the actual trigger
voltage for any settings within the linear range of the LEVEL
control. The same can be done when triggering on a negative
slope except the trigger LEVEL control is first turned to its most
negative extreme then slowly advanced until the circuit just
triggers. The measured difference voltage is subtracted from
24
other VC readings to get the actual trigger voltage.
VU − VL
, Figure 13c, requires a
2
square wave generator with a variable output amplitude that
swings between some minus voltage and zero or between zero
and some plus voltage. Zero volts must be accurately at 0 as any
residual offset will give incorrect results. For triggering on the
positive slope, turn the LEVEL control on the counter to its most
positive extreme. Connect a 10 kHz square wave that goes from
–1 volt to 0 to the input of the counter. Slowly decrease the
trigger LEVEL until the counter just begins to trigger. This
happens when the upper hysteresis limit coincides with upper
excursion (0 volts) of the input square wave. Triggering is at zero
volts and V'C can be measured as before.
A similar procedure can be used to establish triggering at zero
c. A third method of determining
Hysteresis Window
Volts
+
VU
0
VC
be measured
{ Can
with a dc voltmeter
Hysteresis
Limits
VL
–
a. Using sine wave or square wave to determine spacing
between hysteresis limits.
Triggering ceases if input signal does not touch or
cross both upper and lower hyteresis limits.
VU
Input Grounded
+
VC
Hysteresis Window
VL
V'U
0
V'C =
V'L
–
V'U – V'L
2
}
For positive slope
Circuit triggers when
the upper hysteresis
limit crosses zero volts
b. With input grounded, hysteresis limits are moved slowly
downward with TRIGGER LEVEL control until input
circuit triggers once. V'C can be measured.
VU
+
0
–
VL
}
No Triggering
Hysteresis Window
V'U
Begins triggering
V'C
for positive slope
V'L
}
c. With square wave input 0 to –1 volt triggering
begins when upper hysteresis limits coincide with top
of square wave at zero volts.
Figure 13.
Determining the
hysteresis window,
i.e., spacing
between the
hysteresis limits,
which define the
sensitivity of an
electronic counter
and setting trigger
levels at zero
volts.
25
volts with a negative slope. In this case, the square wave output
is from +1V to 0 volts and the LEVEL control is initially offset to
its negative extreme.
2. Establishing Triggering at Zero Volts on a Sine Wave
d. Establishing triggering at zero volts on a sine wave signal. A
method particularly convenient on sinewave signals is determining if triggering does indeed occur at zero volts using the following procedure: A time interval measurement is set up using any
convenient trigger point. The start channel input amplitude is
then changed by some factor, 2:1 for instance, and observing if
the counter time interval reading changes or not. This is shown in
Figure 14. (It is important that neither the source impedance or
counter input impedance change when the amplitude change is
made. A change of either would change the phase of the input
signal as well as its amplitude.)
If the start channel is triggering at (a) of Figure 13 the counter
will change because the trigger point will shift to (b) with an
accompanying time change when the input amplitude is changed.
Readjust the start LEVEL slightly and repeat the amplitude
change. Continue until the counter reading is the same for either
amplitude. The start channel is now triggering at zero volts (c)
and the time of triggering is independent of input signal amplitude. Repeat the same procedure for the stop channel to get it
triggering at zero volts also.
Time Change
Start Channel at Initial Amplitude
VOLTS
Amplitude Reduced by 2:1
a
0.1
0.0
c
TIME
26
b
VU
NOTE:
2:1 Amplitude Difference
Trigger at Zero Volts
Time of Triggering
Independent of Amplitude
Figure 14. Setting
START or STOP
Channels to trigger
at zero volts dc on
a sine wave.
Hysteresis Compensation
1. What is Hysteresis Compensation?
A conventional counter triggers on the upper hysteresis limit at (a)
Figure 15 when set to trigger on a positive slope and at (b) the lower
hysteresis limit when set to trigger on a negative slope.
Triggering will always occur regardless of the setting of the SLOPE
switch if the input signal is large enough to cross both hysteresis
limits; however, the trigger point on the input waveform shifts to a
different voltage whenever the SLOPE switch is changed.
The voltage VC can be measured to give an indication of the trigger
level but it is offset by about half the hysteresis window so does not
define the exact trigger voltage.
Hysteresis Window
a
VU
Volts
+
VC
0
–
b
VL
Time
Figure 15.
Triggering on
positive or
negative slope.
Hysteresis compensation keeps the trigger point at approximately
the same voltage even though the trigger SLOPE is changed from
positive to negative by introducing a voltage, B1 or B2 of Figure 6,
between the level control, R4, and the trigger circuit. Figure 16 shows
the upper hysteresis limit VU shifted down to V'U by this built-in
voltage source which introduces an offset or “buck out” voltage
between the trigger LEVEL control and the trigger circuit to move VC
to V'C. This voltage corresponds to the battery B1 or B2 in Figure 6.
The magnitude of this offset voltage is one-half the hysteresis
window. Note that the upper hysteresis limit (trigger voltage for a
positive slope) is now opposite VC. A dc voltmeter, internal to the
counter or external can thus be used to indicate the trigger voltage
by measuring the voltage at the arm of the trigger LEVEL control on
a counter that has this hysteresis compensation feature. When
triggering on a negative slope, operation of the compensation circuit
is similar except V'C is shifted upward so V'L coincides with VC
instead of downward as for a positive slope.
27
Without
Hysteresis
Compensation
VU
+
Volts
With
Hysteresis
Compensation
V'U
0
–
V'C
VL
V'L
Hysteresis Window
Figure 16.
Hysteresis
compensation for
triggering on a
positive slope.
2. Limitations
A counter with hysteresis compensation may quit counting when the
slope switch is changed as the input signal may no longer cross both
hysteresis limits. This places no limitation on a measurement a
counter having hysteresis compensation can make however, as this
would also happen in trying to a get a conventional counter to
trigger at the same voltage point. The only difference is that the
operator would use the manual LEVEL control to shift the hysteresis
limits rather than have them shift automatically when the SLOPE
switch is changed. While giving a better indication of trigger level
than an uncompensated trigger, this technique has limitations:
a. The measured trigger voltage VC, see Figure 6, established a
reference voltage for one-half of a balanced amplifier, and is
related to the signal voltage into the other half.
Any gain change or drift between the two halves of the amplifier
with time, temperature, or other environmental factors will
destroy the initially set fixed relationship between VC and the
actual trigger voltage.
b. If the input amplifier does not have a flat frequency response, the
hysteresis window Figure 17 changes with frequency (corresponds to change in rise time for TI measurements) so no single
internally generated “buck out voltage” will be correct over the
frequency range of the counter. This results in a discrepancy
between the measured trigger voltage and the actual trigger
voltage point.
28
VU
Input Volts
+0.4
Hysteresis Window
+0.2
0.0
–0.2
Input Frequency
–0.4
0
VL
Max
Figure 17.
Hysteresis
window gets
wider if input
sensitivity
decreases with
increasing
frequency.
3. HP 5328A Option 040 and HP 5326A/B, 5327A/B have Hysteresis
Compensation
The HP 5328A Option 040 and the HP 5326A/B and HP 5327A/B
counters have this hysteresis compensation feature in the time
interval mode of operation. An HP 5328A with Option 040 (time
interval) and Option 020 or 021 (digital voltmeter) has switch positions labeled READ A and READ B to select and display either Input
A (start) or Input B (stop) trigger voltage to 1 millivolt. One millivolt
resolution is greater than justified in terms of absolute accuracy of
the trigger voltage for reasons mentioned earlier; however, this high
resolution is useful as it is possible to:
a. Return very closely to a previously selected trigger voltage.
b. Accurately move the trigger point by some small amount since
the DVM gives 1 millivolt resolution on ∆V readings. This is
helpful when determining rise times.
The HP 5326B or HP 5327B which has a built-in DVM also has a
READ LEVEL A and READ LEVEL B position to read trigger voltage
in the time interval mode.
Polarity Control
This control determines if the center of the hysteresis window moves to
a positive voltage or to a negative voltage from zero when the level
control is changed. Most counters have the level control connected
between a +V and a –V supply. This puts 0 volts in the center of the
control range so a single control functions both as a POLARITY and a
LEVEL control as in Figure 6. In this case the control element may have
a nonlinear taper to give greater settability around zero volts. The
voltage from the arm of the level control to the trigger circuit is often
brought to an external connector where it can be measured with a dc
voltmeter. This voltage defines the approximate center of the hysteresis
window, VC.
29
Input Attenuator for Measuring Higher
Amplitude Signals
A frequency compensated 2:1, 10:1, 20:1, or 100:1 attenuator between
the input terminal and the input amplifier permits measurement of high
amplitude signals which might otherwise overload or damage the input
circuit. This attenuator has the effect of increasing the hysteresis
window by the attenuation factor, for instance: a counter with 100 mV
rms sensitivity would have a hysteresis window of 282 millivolts (peakto-peak value of 100 mV rms). The x10 position on the attenuation
raises sensitivity to 1V rms and the hysteresis window becomes 2.82
volts so signals below this amplitude can no longer be counted. Similar
reasoning applies for other attenuation factors.
The SLOPE, POLARITY, LEVEL, and ATTENUATOR controls allow the
operator to start or stop a measurement anywhere on an electrical
input signal except the most negative part of the signal when triggering
on a positive slope or the most positive part of a waveform when
triggering on a negative slope. (At the peak, one or the other of the
hysteresis limits is no longer crossed by the signal.) Operation of these
controls is similar to that on the sweep circuit of a modern oscilloscope. On the oscilloscope these controls determine where, on an input
waveform, the sweep begins. On the counter they determine where on
the signal the measurement begins and ends.
Trigger Lights
When measuring time interval the counter displays a reading only if it
gets a Start and a Stop pulse. During setup on an unknown signal it is
not always obvious if both input channels are triggering or not if the
counter is not gating. To make initial setup easier, trigger lights, one for
each channel, are often provided. A neon lamp or LED is used to
indicate channel activity so the operator can tell by looking at the light
if the channel in question is triggering regardless of whether the other is
triggering or not. The trigger light drive circuit includes a pulse
stretcher to insure that the light stays on long enough to be seen even
though the actual input pulse may be too narrow.
1. Two-State
Two general types of trigger light presentations are used: The two
state display used on the HP 5308A and on the HP 5326/5327 series
counters has lights that are OFF when the circuit is not triggering
but BLINK when the circuit is triggering. As the input repetition rate
increases above about 50 Hz the trigger lights appear to stay on
continuously.
30
2. Three-State
The trigger lights of the three state display used on the HP 5328A
may be OFF, BLINKING, or ON. A trigger light is OFF if the input is
below the trigger level (due to too small a signal, a dc component on
the signal or the trigger level control incorrectly set) and ON continuously (but no triggering) when the input is above the trigger level.
The light BLINKS each time the input triggers for rates up to about
10 Hz and blinks at about a 10 Hz rate for inputs of 10 Hz to 100 MHz.
This not only gives the operator an indication of triggering but also
some indication of the problem if the counter is not triggering.
Markers
Many electronic counters generate electrical signals for use as markers
when an input channel is triggered.
Some types of markers are:
1. DOT MARKERS
When a channel is triggered it puts out a short duration electrical
marker pulse (100 ns wide) that can be used to intensity modulate
the trace of an oscilloscope displaying the input signal. The marker
shows up on an oscilloscope display as a bright dot on the waveform.
Dot marker pulses, Figure 18a, are useful to indicate the trigger point
on a slow rise waveform since marker width and circuit delays are
both small compared to the risetime of the input waveform; however, as one gets into fast risetime pulses these effects are no longer
insignificant. This coupled with the CRT phosphor rise and decay
time makes intensity markers of little use in the nanosecond region
as the markers begin to look more like comets than dots so the
actual trigger point is no longer well defined.
Dot marker outputs are useful on sine wave signals from 100 Hz to
100 kHz. At higher frequency, the marker width becomes an appreciable part of the period of the input signal so the marker no longer
defines a specific point on the input waveform. Also the delay
through the intensity modulation (Z axis) amplifier is not usually
known so dot markers are of little use in the sub-microsecond
region.
At low frequencies, dot markers are not useful unless the dot width
is increased as the trace is intensified for such a small portion of
one cycle that it is difficult to see the marker. Separate connectors
are supplied for the START marker and STOP marker outputs on
most time interval counters.
31
Start Positive Slope
Triggers at +0.7 Volts
Peak Volts
+2
+1
+1
0
0
Stop Negative Slope
–0.5 Volts
–1
–2
–1
TI
TI Measured
a. DOT Markers. Both the stop and the start
markers have been connected to the Z axis
(intensity modulation) input of an oscilloscope.
b. GATE marker for the same
measurement as in (a).
Figure 18. Time
Interval Markers
as they appear on
an oscilloscope.
2. GATE MARKERS
A gate marker, Figure 18b, generates a dc voltage for the duration of
the counter measurement (gate OPEN to gate CLOSE). This can be
used to intensify an oscilloscope display of the input signal from the
receipt of a START signal by the counter until the receipt of a STOP
signal.
High impedance gate markers work well to very low frequencies or
to demonstrate basic triggering ideas on time interval measurements
but are not useful for fast pulse or short delay measurements as they
have the same drawbacks as dot markers. Figure 19 shows markers
as they appear on an oscilloscope.
+2V
0V
–2V
0V
Start (A Channel) marker triggering on
positive slope at +1.25 volts
Stop (B Channel) marker triggering on
negative slope at –1.25 volts
0V
Combined markers
0V
Gate output for above conditions
0.5ms
HP 5326A/B — HP 5327A/B Marker and Gate Outputs
NOTE:
32
Dot markers come from the trigger circuit so they appear
on every input cycle. A gate marker appears only when the
gate is open so it does not appear unless the gate opens.
Figure 19. Actual
Dot and Gate
Marker Outputs
The HP 5345A and HP 5328A Option 040 also have 50Ω gate outputs.
These have fast risetime. They are useful over the full frequency
range of the counter as this gate output can be displayed on the
second channel of a wide band oscilloscope connected to the signal
or can be mixed with the input signal before display. These gate
markers may be offset from the actual gate time by 10 ns to 100 ns.
For detailed information consult the data sheet for a specific
counter. Figure 20 and Figure 21 are illustrations of markers.
3. SQUARE WAVE MARKERS
The HP 5328A has 100 mV into 50 ohm marker outputs which are
inverted replicas of the Channel A and Channel B Schmitt trigger
outputs as in Figure 21b. These can be displayed along with the input
signal on a dual channel oscilloscope. These markers are fast enough
that they are useful to 100 MHz in a 50Ω system. In general there is
some delay between the marker output and actual triggering so the
counter data sheet should be consulted for detailed information. The
HP 5328A Option 040 High Performance Universal Module has a
Channel A marker as described above as well as a gate type marker
which is high during a TI A B measurement. Both are available from
front panel connectors.
Input Signal
Gate Output 1V P-P
Marker Output
1 ms
Figure 20.
HP 5328A Gate
Output 1 peak-topeak into 50 ohm
load.
33
(B)
(A)
+1V
Input Signal
0V
–1V
Triggering at 0 volts
Triggering at –1.25 volts
Triggering at +1.25 volts
}
Marker Output
1µs
a. HP 5328A Standard Unit Marker Output into 50 Ohm Load. Positive Slope - Triggers at
Trailing Edge of Marker at (A). Negative Slope - Triggers at Leading Edge of Marker (B).
+1V
0V
Input Signal
–1V
Triggering, Positive Slope, 0 volts
Triggering, Negative Slope, 0 volts
Triggering, Positive Slope, +1.25 volts
Triggering, Positive Slope, –1.25 volts
1µs
b. HP 5328A Option 040 Marker Output into 50 Ohm Load. Triggering always corresponds to
Leading Edge of Marker.
Delay ! Control
Besides the conventional input controls, the HP 5328A Option 040 and
the HP 5304A each have a DELAY control associated with the start
input. This control is useful in making a measurement on a complex or
noisy signal. Once a measurement begins (following a START pulse)
STOP pulses are locked out until expiration of a pre-determined,
adjustable, analog delay. The measurement terminates with the first
STOP pulse following delay run down. This feature is useful in relay
testing for example as the delay is used to get away from errors due to
contact bounce. In Figure 22a, a measurement started at 1 would
normally end at 2 on the first bounce. Adjusting the delay time to
something greater than 3 but less than 4 lets the operator measure the
time from the first contact closure to the first contact opening after the
coil is de-energized. Figure 22b illustrates another application where a
measurement is required between specific pulses on a complex pulse
34
Figure 21.
Square wave
marker output
of the HP 5328A
Counter.
train such as that used in ATC (Air Traffic Control) systems for example. Here the DELAY control is used to select specific STOP locations
within the pulse train. The SAMPLE RATE control is also helpful as this
gives control of the counter “dead time” and so gives some control of
the time between pulse trains.
The HP 5345A can also be used for this kind of measurement; however,
in this case the DELAY signal is generated by an external generator such
as the HP 8007A Pulse Generator. Even greater measurement flexibility
is possible with this combination as the pulse generator gives the
operator control of both dead time and delay with respect to an externally supplied reference (sync) pulse. Pulse width defines the lock out
interval and pulse delay defines the position of the start pulse on the
pulse train or other waveform.
Start
Contact Bounce
Stop
{
Contacts Closed
Contacts Open
Delay Control
12
3
4
Delay
Start
Stop
Measured Interval
a. Delay Control used to lock out spurious signals due to contact
bounce when measuring relay operate time.
Pulse Train
Time Between
Pulse Trains
Pulse Train
Delay Control
Delay
Measured Interval
b. Delay Control used to confine measurement to a
specific pair of pulses of a complex pulse train.
Figure 22. Using
the DELAY Control
in Time Interval
Measurements.
35
Time Interval Averaging
On a repetitive signal, time interval averaging increases the resolution
with which a time interval measurement can be made. Also, depending
on the design of the averaging circuits, this technique may extend the
minimum measurable interval to less than the period of the counter
clock.
Reduce +1 OR –1 Count Error by N on
Repetitive Signals
The basis of time interval averaging is the statistical reduction of the
random +1 or –1 count error inherent in digital measurements. As more
and more intervals are averaged, the measurement will tend toward the
true value of the unknown time interval but only if the ±1 count error is
random. The word “random” is significant. For time interval averaging
to work the time interval must (1) be repetitive; and, (2) have a repetition frequency which is asynchronous to the instrument’s clock.
Under these conditions the resolution of the measurement is increased
by the factor:
±1 count
N
Where N = the number of independent time intervals averaged
N defines the improvement in resolution with TI averaging.
When doing time interval averaging the number of digits actually
displayed by the counter increases directly as the number of intervals
averaged, i.e., 10 averages display one additional digit, 10000 averages
display four additional digits, etc. This can be confusing to the operator
since the improvement in resolution of the measurement with averaging
is only as the N i.e., by 3 or by 100 for the example above. The
displayed digits beyond this are random numbers therefore completely
useless. The HP 5345A Electronic Counter has a display position switch
which can be set to eliminate these useless digits to reduce operator
confusion. Modern microprocessor controlled instruments like the
HP 5370A and HP 5315A automatically truncate unwanted digits.
Time Interval Averaging is Useful When
• +1 count or –1 count error from a single time interval measurement
significantly degrades the accuracy or resolution of a time interval
measurement; and,
• The input signal has superimposed noise or jitters.
36
For Example:
If the width of a repetitive pulse is approximately 1 µs, the +1 count or
–1 count error in a pulse width measurement using conventional oneshot techniques is 100 ns, 10 ns, or 2 ns (the period of the counter’s
clock). This error is a large part of the time interval; however, averaging
104 time intervals can produce 1 ns or better resolution. True time
interval averaging is achieved only when the signal repetition rate is not
coherent with the counter clock as the time relationship between the
signal and the counter clock must be such as to sweep through the full
range of the 0 to –1 or 0 to +1 count ambiguity in a random manner to
satisfy the statistical requirement of averaging.
If the clock and the input become coherent the system behaves as a
sampling system so no improvement whatsoever is had by averaging.
The HP 5345A and the HP 5328A Option 040 both achieve true time
interval averaging by using a patented noise modulated clock for all TI
averaging measurements. This frees the operator from repetition rate
considerations.
With averaging, resolution of a time interval measurement is limited
only by the noise inherent in the instrument. A typical figure of
50 picoseconds resolution can be obtained with good low noise design.
Synchronizers Needed for True TI Averaging
Synchronizer circuits are necessary in the counter start-stop channels
when doing time interval averaging, these circuits insure that the
counter gate does not receive partial pulses as this would bias the
displayed answer away from the true value in an unpredictable manner.
The synchronizers operate as in Figure 23. The top waveshape shows a
repetitive time interval which is asynchronous to the square wave clock.
When these signals are applied to the main gate, an output similar to the
third waveform results. Note that much of this output results in transitions of shorter duration than the clock pulses. Decade counter assemblies designed to count at the clock frequency dislike accepting pulses
of shorter duration than the clock. The counts accumulated in the DCAs
will therefore approximate those shown in the fourth trace — the exact
number of counts is indeterminate since the number of short duration
pulses actually counted by the DCAs cannot be known. Since the time
interval to be measured is slightly greater than the clock period, the
fourth waveshape shows that the average answer will be in error,
having been biased, usually low, because of the DCAs requirement of
having a full clock pulse to be counted.
37
This problem is alleviated by the synchronizers which are designed to
detect leading edges of the clock pulses that occur while the gate is
open. The waveshape applied to the DCAs, when synchronizers are
used, is shown by the fifth waveform. The leading edges are detected
and reconstructed, such that the pulses applied to the DCAs are of the
same duration as the clock.
Synchronizers are a necessary part of time interval averaging; without
them the averaged answer is biased even though the reading appears
to settle down to a stable number. In addition, with synchronizers
involved, the counter can be designed to make time interval measurements of much less than the period of the clock. This technique is only
as good as the synchronizers, however, high-speed synchronizers can
enable intervals as small as 100 picoseconds to be measured, even
though the clock period might be 100 nsec for example.
Start
Start
Stop
Start
Stop
Start
Stop
Stop
Input
Clock
Gate Out
Counts Accumulated
in DCAs
With Synchronizers
Counts Accumulated
in DCAs
Extending Time Interval Measurements to Zero Time
This technique is used with the HP 5328A, HP 5308A, HP 5326A/B, and
HP 5327A/B Counters to extend minimum TI average measurements
down to the nanosecond or sub-nanosecond region. The main disadvantages of the synchronizer system used with these counters is that the
time between a stop pulse and the next pulse must always be longer
than the clock period, and multiple stop pulses can lead to incorrect
readings.
The HP 5345A uses a different synchronizer approach so it will not
make time interval average measurements below 10 nanoseconds. It
can make a measurement of 10 ns wide pulses spaced as close as 10 ns
(50 MHz rate).
The 10 ns minimum time interval is not a serious limitation as it can
always be circumvented by adding an additional 10 ns delay (approxi-
38
Figure 23.
Synchronizer
operation with
time interval
averaging.
mately 200 centimeters of RG-58/U cable) in stop (B) channel input. The
added delay can be measured as in Figure 24 after which it can be
subtracted from all subsequent measurements using the same interconnecting cables. The HP 5363B probe box already has this delay capability which can be adjusted to 10.0 ns in a way that includes all of the
system differential delays as well. Application Note 162-1, Time Interval
Averaging, Hewlett-Packard Co., discusses TI averaging in detail.
Disadvantages of Time Interval Averaging
• Requires repetitive pulses.
• Is not useful for statistical measurements such as rms jitter or
histograms as the averaging process destroys the very information
which is sought.
• Takes a long time to make the measurement for low repetition rate
time intervals.
HP 5345A
A
B
Figure 24.
Extending
Time Interval
measurements
to Zero time by
adding additional
delay in the stop
channel.
Pulse Gen
10 ns delay = approx.
200 centimeters of
RG-58/U 50Ω co-ax cable
39
Time Interval Error Evaluation
Total counter error in a time interval measurement is made up of
several parts all added together in a way to get the largest number thus
defining maximum measurement error. These are:
•
•
•
•
±1 count
± trigger error (including trigger settability)
± time base error
± systematic error (when measuring short intervals)
+1 Count Error or –1 Count Error
As for any measurement made with an electronic counter a +1 count or
a –1 count ambiguity can exist in the least significant digit of a time
interval measurement. This comes about because the digitized internal
counted clock frequency and the input Start/Stop signals are not
coherent in the usual measuring situation. The counter has no way of
interpolating fractional clock pulses if a Start or Stop falls between
clock pulses so the measurement could be off by as much as +1 or –1
clock period. For a 500 MHz clock, ±1 count represents ±2ns; for a
100 MHz clock ±10 ns, etc. One exception is the HP 5370A counter
which uses a digital interpolation scheme to give theoretical resolution
of ±20 ps.
The ±1 count error becomes more significant on short time interval
measurements where the total number of displayed digits is only 2, 3, or
4 — as when making rise time or narrow pulse measurements — since
this fixed error becomes a more significant part of a short interval. For
one-shot measurements this 1 count error determines the ultimate
resolution of the counter.
The 1 count error can be reduced by N by doing time interval averaging (N = number of intervals averaged) if the input signal is repetitive.
Time interval averaging depends on the random occurrence of the +1
count or –1 count error so will not work if the internal clock of the
counter and the input signal are coherent. The HP 5345A and the
HP 5328A Option 040 both have patented noise modulated clocks so the
operator need not worry about coherence.
± Trigger Error
Trigger error occurs when the counter does not trigger at the expected
voltage level on the input signal. This may be due to:
Noise or distortion.
• On the input signal.
• Added to the signal after it enters the counter.
40
Drifting of the trigger voltage point of either channel with tem
perature change, line voltage change, or component aging.
Energy effect on fast rise signals.
1. Noise on Input Signal
Triggering is set to occur at +1 volt on the input signal as shown at
Figure 25a.
Low frequency noise on the input signal can cause triggering to
occur too early (Figure 25b) or too late (Figure 25c). High frequency
noise can cause early triggering only.
Noise can occur on either the Start or Stop pulse or both, consequently the measurement may be either too long or too short by
twice the error shown in the example.
a.
1
0
Volts
Noise
1
b.
0
1
c.
0
Noise
∆t
Time
Figure 25. Time
Error introduced
by noise on the
input signal.
2. Distortion on Input Signal
Figure 26 shows how harmonically related noise as well as
nonharmonically related noise moves the trigger point in time.
Also shown is the effect of noise on a signal.
3. Increased Signal Amplitude Reduces Time Error
Increasing signal amplitude reduces the time error ∆t if the high
amplitude signal does not have any more noise than the lower
amplitude signal. In practical applications this is often the case as
noise is frequently due to ground loops or to other spurious signals
41
Hysteresis
Window
+1
0
–1
a.
+1
0
–1
b.
+1
0
–1
c.
+1
0
–1
d.
0
0.5
1
1.5
Time in Millisec
2
2.5
NOTES:
1000 Hz signal.
Circuit set to trigger on a positive slope at +1 Volt. (Hysteresis window is 2 volts peak-topeak centered on 0 volts dc.) Circles denote actual trigger points.
a. Triggering on a low distortion sine wave signal. Trigger points, identified by circles,
occur at the same relative point time wise on each cycle.
b. Sine wave with harmonic distortion. Trigger points are displaced in time (but not voltage)
from waveform “a” however they are consistent from cycle to cycle. Only one output
pulse per cycle occurs unless the distortion voltage becomes large enough to cross
both hysteresis limits. Depending on the measurement, harmonic distortion on the input
signal may or may not introduce an error.
c. Sine wave with nonharmonically rated distortion. Trigger points can occur early or late
depending on the distortion and are not consistent in time from cycle to cycle.
Nonharmonically related distortion of the input signal usually introduces errors into a
measurement.
d. Sine wave with noise. Since noise is random, the trigger point may occur anywhere with
respect to the sine wave signal. Note particularly that improper triggering is caused by
the voltage of the noise signal crossing the trigger level. In the case of a narrow spike
on a low frequency sine wave, distortion may seem insignificant when measured with a
distortion analyzer that looks at average voltage or power. Only a peak reading voltmeter
or oscilloscope will give the necessary information when looking for this kind of error.*
*When using an oscilloscope to examine a signal for noise pulses, the bandwidth of the
oscilloscope must be compatible with the resolution of the counter being used for the
measurement. For example, when using an HP 5328A Counter one must examine the input
signal for noise pulses as narrow as 5 nanoseconds as the counter will resolve these
narrow pulses even if the input signal frequency is as low as a hundred Hertz or less. The
HP 5345A will resolve 2 ns so a 500 MHz oscilloscope is needed when looking for noise
pulses.
42
Figure 26. Effect
of Harmonic
Distortion or
non-harmonic
distortion and
noise on the time
a measurement
begins or ends.
whose amplitude is relatively constant regardless of the level of the
desired signal. In general, the higher the amplitude of the input signal
on any given attenuator setting the smaller the error due to noise.
Amplitude cannot be increased indefinitely, however, as ultimately
the input overload protection circuits of the counter come into play.
The protection circuits may completely distort the measurement and
give erroneous readings under heavy overload.
Volts
With sine wave inputs in particular, increased amplitude is helpful as
the slope of the signal through the hysteresis window increases as in
Figure 27. Triggering should occur near the center of the sine wave if
possible as slope is steeper here.
+1
b
a
Noise
0
t' t
Figure 27.
Increased signal
amplitude reduces
time error.
Time
Triggering at zero volts on an ac coupled symmetrical signal, Figure
14, makes the trigger point timewise independent of the amplitude of
the signal. Triggering near either peak of a sine wave should be
avoided as the low slope both before and after the peak gives rise to
large trigger error. By the same token if the input signal has infinitely
fast rise time (theoretical only) it would make no difference where
on the waveform triggering occurs as there is zero time difference
between the voltage levels in Figure 28.
V2
0
V1
∆t = 0 between V1 and V2
Figure 28.
Trigger time is
essentially the
same regardless of
trigger voltage for
an input signal
with infinitely fast
rise time.
Often a time interval measurement is made from a point on one
signal to a corresponding point on another signal. For this measurement the location of the trigger points is not important so long as
Start and Stop are at corresponding points on the input signal. When
possible, the trigger point should be located where the rate of change
of voltage is greatest as the time error is then at a minimum for a
given amplitude of noise voltage as indicated in Figure 29.
43
3
b
Volts
2
a
Figure 29. The
same amplitude
noise pulse gives a
smaller error (E)
on a fast rise
signal (a) than on
a slow rise signal
(b).
1
0
E
E1
Time
4. Calculation of Trigger Error Due to Noise or Distortion
Two expressions are used to evaluate the trigger error in a time
interval measurement, one is for sine wave inputs and the other for
pulse inputs. The general error expression is derived as follows. See
Figure 30.
Noise Spike
et
+ Slope Triggering
Hysteresis
Window =
2 × 2 × rms sens
et = Trigger Error
Vn
peak
TRIGGER ERROR in time, e t =
Signal
Slope =
dv s
dt
±2Vn peak
Signal SLOPE
Vn = peak noise amplitude and is the sum of
the noise present on the signal before it
enters the counter plus any noise
contributed by the counter’s input channel.
t = time
et = trigger error in time
then
dv s
V peak
= n
dt
et
Vn peak
et =
for one channel
dv s dt
A factor of ±2 is included to account for both the start channel and
stop channel errors and a ± sign takes into account the fact that the
gate may open or close early or late.
44
Figure 30. Trigger
error derivation
5. Trigger Error Due to Input Amplifier Noise
The effective rms noise contributed to the total noise by the
counter’s input is not usually specified and in many cases has been
negligible. However, as sophisticated applications for counters
increase, the noise contribution should be specified. For example, in
the HP 5315A this noise component is specified as 80 µv rms over the
bandwidth of the counter.
6. Trigger Error for Sinewave Input Signals
Signal = v s peak sine 2 π ft = v s
dv s
= 2 πf v s peak cos 2 πft
dt
Slope is max at center of waveform where cos 2πft = 1
dv s
dt
= 2πf v s peak
max
Substituting in the general error equation developed earlier
2 V n peak
2 πf v s peak
for 40 dB signal to noise ratio (amplitude, not power)
TRIGGER ERROR e t = ±
V n peak
= 0.01
v s peak
(
)
±2 0.01
V peak
=
TRIGGER ERROR = ±2 × n
= ± 0.003
2 πf
2 πf
v s peak
f
where T = period of signal
e
Fractional time error = ∆T = t = ± 0.003 = 0.3%
T
T
f
T
For noise 40 dB down (peak voltage-wise) from the signal. This is the
expression most often seen in both the period and time interval
accuracy specifications. This derivation places no requirement on
input signal amplitude; however, it does assume triggering at a
maximum slope, therefore, the peak input signal must exceed the
actual hysteresis window by about 30% or more for sine wave signals
to satisfy this requirement. Trigger error can be greater by a factor of
10 when triggering at or near either peak of a sinewave signal.
Reducing the distortion on the input signal reduces this error somewhat but not as much as expected because noise introduced internally by the input circuits becomes significant when input distortion
is down by 50 dB or more.
45
7. Trigger Error Expression for Pulses


V n peak 
 V n peak
TRIGGER ERROR IN TIME, e t = ± 
+

Signal Slope 
 Signal Slope
 Start Channel Stop Channel 
For a counter with 100 mV rms sensitivity the hysteresis window =
100 mV × 2.82 = 282 mV
For a 40 dB (amplitude) signal to noise ratio the peak-to-peak noise
for a 100 mV rms signal is 2.82 mV and peak noise is 1.41 mV.
Therefore:
Trigger error, et, for one channel =
( 1.41 mV)
= 0.00141 Volt
Signal Slope
Signal Slope
TOTAL TIME INTERVAL ERROR ≈ ±
0.00141
0.00141
±
Signal Slope of Signal Slope of
start channel
stop channel
where signal slope is in volts per unit time
This derivation does depend on the input sensitivity of the counter
(100 mV rms used for this example) so the expression will be
different for counters with other input sensitivities. Start channel
slope and noise and stop channel slope and noise may not be the
same so must be evaluated individually for each channel.
8. Combining the Effect of Input Signal and Counter Contributed Noise
The rms noise associated with a single trigger is
X 2 + e 2n
Input slew rate at trigger point
where X = effective rms noise of the counters input channel
(= 80 µV for the HP 5314A and HP 5315A)
en = rms noise voltage of the input signal measured over a bandwidth equal to the counter’s bandwidth.
Since for time interval measurements trigger error occurs at both
the start and stop signals and the noise adds on an rms basis, the
rms trigger error for time interval measurements is
1.4 X 2 + e n2
Input slew rate at trigger point
46
s rms
In most cases, the dominant noise component is contributed by the
input signal. If the input signal is especially “quiet” the input amplifiers become a consideration. The fast slew rates of modern logic
greatly reduce the effect of noise errors and in many practical cases
these can be neglected.
9. Energy Effect
In addition to the previously discussed factors which contribute to
the trigger point time error, the so called “energy effect” is also a
factor to consider when measuring intervals of a microsecond or
less. This effect although always present is small enough so it is
generally overshadowed by noise and other factors when measuring
longer intervals. The energy effect comes about because the input
voltage must remain beyond the selected trigger voltage point long
enough to transfer a finite amount of change, represented by the
shaded area of Figure 31 into the trigger circuit to activate it. The
amount of change necessary to initiate triggering is a function of
both the input amplifier-trigger sensitivity and the circuit bandwidth
so may be different for each model of counter. Due to this effect, the
counter always triggers at a higher voltage (or lower depending on
selected slope) than selected so in all cases the measurement begins
or ends later in time than expected.
Input Volts
Circuit Triggers
Selected Trigger
Voltage Point
∆T1
∆T2
0
Time
Input Signals
Figure 31.
Energy effect
introduces time
error ∆T which
can be significant
on short TI
measurement if
signal does not
extend well
beyond the
selected trigger
point.
The time error due to “energy effect” can be minimized by selecting
trigger points well within the maximum excursion of the signal. For
the HP 5363B Time Interval Probes the specification requires the
signal to go beyond the selected trigger voltage point by at least
100 millivolts or the trigger voltage selected must be below the peak
by 8% or more of the peak-to-peak input voltage.
The energy effect error is negligible for long time intervals and when
triggering the start and stop channel at the same voltage point on
input signals with identical risetime and amplitude as when making
phase measurements. It does, however, require careful consideration
47
when making risetime measurements on fast pulses as this calls for
setting one trigger voltage point near one extreme and the other
near the opposite extreme of the input signal.
This discussion points up the need for identical start and stop
channels when making sophisticated TI measurements on fast
signals. All circuits introduce some delay, but if start channel and
stop channel delays are equal, the accuracy of the measurement is
unaffected even though displaced in time. Identical amplifiers are
most apt to satisfy this requirement. Amplifiers with different
bandwidths will introduce error as the internal time constants are
longer, hence propagation delay is greater for low frequency
amplifiers.
± Time Base Error
Crystal oscillator (clock) accuracy is seldom of consequence in practical time interval measurements. Most electronic counters have quartz
oscillators with accuracy of a part in 106 (1 part per million) or better.
As a result, the effect of oscillator stability does not affect a time
interval measurement unless the display has 5 or 6 valid digits of
information. While it is possible to measure long intervals with high
resolution, most practical measurements today are rise time of fast
signals, propagation time through high-speed logic or on narrow pulses.
Resolving a 5 µsec interval to 1 nanosecond entails only 4 digits of
information — i.e., 5000 nanoseconds — so an oscillator as poor as
1 part in a million introduces only 1/200 as much error as ±1 count for
this measurement. For shorter intervals the oscillator error is proportionally less.
Time interval averaging increases the number of valid digits, but here
again usually not to the extent that crystal accuracy is important. Short
term stability may become important when doing time interval averaging on narrow pulses. The short term stability specification is statistical
in nature so is worse for short averaging times. Consider for example a
short term accuracy specification of 1 × 10–9 for a 1 second averaging
time. This would be 1 × 10–3 for a 1 sec averaging time (5 nanosec for
the example above) and would be greater for an oscillator with poorer
short term stability.
Aging rate is generally of consequence only in an application where
the counter is used to make phase or time measurements to compare
high precision frequency standards. In this application a counter is used
to measure the time variations between once per second time ticks
from the two different frequency standards. Each time interval measurement may be as long as one second (1 × 106 µsec) on frequencies
that are stable 5 parts in 1012 or better so a stable crystal oscillator is
needed.
48
± Systematic Error
These fixed errors introduced by interconnecting cables, trigger level
settings and similar parameters introduce errors which are essentially
constant for any given measurement configuration. Since these errors
are relatively constant they are of importance only when making
absolute measurements of width, delay, etc., and do not enter in when
looking for changes in the measured parameter as an experiment
progresses. Particularly on short intervals, the difference in delay due to
propagation time on both the interconnecting cables as well as through
the start and stop channel circuits may no longer be insignificant. This
differential delay must either be adjusted out or taken into consideration when making absolute measurements. This error is usually small
so that it need not be considered when making measurements longer
than 10 microseconds or so. Signal propagation time is about 0.052
nanosecond/centimeter (19.2 cm/ns) for polyethylene insulated RG 58
A/U 50 ohm coax cable, therefore cable lengths must be matched for the
start and stop channels or the measurement may be in error. In most
measurement situations, the use of equal length cables cancels out this
error; however, cable delay varies with the dielectric constant of the
insulation so cables of the same kind should be used. Where this is not
practical, the electrical lengths of the cables can be measured by the
counter to determine a correction factor to apply to the counter reading.
Application Note 174-10, Measuring Electrical Length (Delay) of
Cables, discusses this measurement in detail.
Just as with other measuring instruments looking at fast rise input
signals, one must also consider the fact that the loading introduced by
connecting the timer interval counter may actually change the shape of
the input signal. The resistive component of loading is usually high
enough so as not to be troublesome, rather it is the input capacity of the
counter and of the interconnecting cables that accounts for the loading.
The HP 5363B Time Interval Probes box reduces this loading as far as
feasible for a general purpose instrument.
49
Time Intervals by Digital Interpolation
Digital Interpolation
Another technique for improving resolution is by interpolation which is
a method of directly resolving the ±1 count inherent in digital counting
schemes. A principle advantage of interpolation over averaging is that
the technique does not depend on a repetitive signal for high resolution.
In the HP 5370A the interpolation scheme provides 20 ps single shot
resolution. If this is compared to a conventional scheme where cycles
of the internal time base are counted, it is the equivalent resolution that
would be achieved by using a 50 GHz clock. Such clock frequencies are,
of course, well beyond the present state of the art.
Another advantage of interpolation is that the high single shot resolution allows the collection of meaningful statistical information. Important characteristics concerning time intervals like max., min., standard
deviation may be collected in addition to the mean. If fast data collection facilities are available full histographical information may be
obtained. In many time interval measurement situations, these statistical properties are of prime importance. For example, in a digital
communication system where jitter is an important consideration, the
standard deviation gives an excellent indication of this.
Phase-Startable Phase-Lockable
Oscillators (PSPLO)
Precision digital interpolation schemes are based on phase-startablephase lockable oscillators (See HP Journal, August 1978). These
oscillators have two essential characteristics
1) When triggered oscillation will start, virtually instantaneously, and
with a fixed phase relationship to the external trigger pulse.
2) Once oscillating the stability of the frequency is preserved by phase
locking to a stable source such as a quartz oscillator.
The Dual Vernier Method
A precision dual vernier method of interpolation similar to that used in
the HP 5370A is shown in Figure 32. This system uses phase startable
phase lockable oscillators in both the start and stop channels. These
oscillators have a slightly longer period of oscillation than the reference
oscillator such that once started, they will reach coincidence with the
reference oscillator some number of cycles later dependent on the
initial phase. If the reference oscillator period is T0 then the start and
stop oscillators have periods equal to T0 + T0/N where N is the interpolation factor. From the diagram, it can be seen that the required time
interval T may be evaluated as follows:
T = T1 – T2 + T3
= T0 (N0 + (1 + 1/N) (N1 – N2))
50
where N1 = Number of start oscillator pulses to coincidence
N2 = Number of Stop oscillator pulses to coincidence
N0 = Number of reference oscillator pulses between start and
stop is incidence.
In the HP 5370A the reference oscillator has a period of 5 ns and the
start and stop oscillators have periods equal to

1  ns
.
(5 ns + 5 ns/256) = 5  1 +
256 
This allows time to be resolved to 5/256 ns ≅ 20 ps.
A practical limitation on the interpolation factor N is imposed by the
inherent noise in the counter’s input and other circuitry. If the typical
noise jitter is say 35 ps rms, there is little advantage to be gained by
having an interpolation factor capable of giving resolutions substantially better than this. Furthermore, the coincidence detection between
the interpolators and the reference clock would need to be done with a
correspondingly high time discrimination. This also imposes a practical
limitation. The dual oscillator scheme described above will operate
regardless of whether the start or stop oscillator commenced first. In
the HP 5370A if the stop signal arrives first, the time interval is given a
negative sign.
Digital Vernier Interpolation in the HP 5370A
Reference Oscillator
T0 =
5 ns
T1
Start
N1
T0 +
T0/256=
5.02 ns
0
Start
Coincidence
1
Stop
Coincidence
3
2
T2
T
N2
Stop
5.02 ns
0
1
2
T3
N0
5 ns
0
1
2
TI = 5 [ 1.004 ( N1 – N2 ) + N0 ] ns
Figure 32. Digital
Interpolation
Scheme using
phase-startable
phase-lockable
oscillators.
51
The Generation of Precise Time Intervals
Phase-Startable Phase-Lockable Oscillators
Phase-startable phase-lockable oscillators (PSPLO) as described in the
foregoing section may also be used for the generation of very precise
digital delays as in the HP 5359A Time Synthesizer. Here the output from
the PSPLO is counted digitally in a preset counter to generate a very
precise delay. An important benefit derived from using PSPLOs is that
the phase with which the delay commences being synthesized has a
constant relationship to an external trigger signal. By contrast, a more
conventional scheme would involve interrupting an internal clock train
at some random phase resulting in a time interval being generated which
is subject to + or – one period variation of the clock.
A simplified block diagram of the HP 5359A and the timing relationship
between pulses is shown in Figure 33.
Ext
Trigger
Auto
Sync
Out
Preset
Digital
Sync Delay
Digital
Delay #1
0–160 ms
in 10 ns Steps
Analog
Delay #1
0–10 ns
S Q
R
Startable
Oscillator
(Lock to Xtal)
Auto
Calibrate
Digital
Delay #2
0–160 ms
in 10 ns Steps
Microprocessor
Controller
Output
Analog
Delay #2
0–10 ns
Power
Supply
HP-IB
Interface
Events Count
Ext Timing
Compensation
Delay (Fixed)
Ext Pulse
Sync Out
TD
Delay 0–160 mS
TW
Width
5 ns–160mS
Output
35 ns Nominal
Edge 1
35 ns Nominal
Edge 2
52
Figure 33. Block
Diagram for the
HP 5359A Time
Synthesizer
showing timing
relationships
between output
pulses.
Here the digital delay #1 determines the time from the sync output pulse
to the output (TD) and digital delay #2, which causes the SR flip flop to
be reset, determines the width of the output pulse. Both digital delays
which operate with 10 ns resolution are followed by analog delays
which provide resolution from 0 to 10 ns. An interesting feature of the
delay scheme is the auto calibrate facility which calibrates the delays
including the analog portion back to a reference crystal oscillator. Since
the instrument contains a microprocessor, any variations can be
automatically compensated for. This scheme allows generation of time
intervals with sub-nanosecond repeatability even at large delays of up
to 150 ms.
The instrument is useful as a calibration source in Radar and other
navigation systems. For example, the sync output could be used to
simulate the send pulse, and the output which is precisely controllable
in time could be used to simulate the return pulse corresponding to a
certain range.
The instrument also has many general purpose laboratory applications
in circuit design where it may be used as an extremely controllable
delay line for the investigation of circuit tolerances and race conditions.
53
Time Interval Measurement Using the
HP 5363B Time Interval Probes
Solve Troublesome TI Measurement Problems
The HP 5363B Time Interval Probes box, Figure 34, is a signal conditioner which solves the most troublesome problems encountered in
making time interval measurements using an electronic counter:*
Figure 34.
HP 5363B Time
Interval Probes
1 . By reducing circuit loading.
High impedance, low capacitance, active probes let the operator
transport the measurement point to the circuit under test. This
keeps circuit loading to a minimum as the usual interconnecting
cables are no longer required for the input signal. This is particularly
important when working with pulses on other than 50 ohm systems.
2. By accurately defining the trigger voltage point.
A simple calibration procedure after SLOPE selection accurately
sets triggering at zero volts and calibrates out the residual offset
voltages using digital techniques. Once calibration is completed, the
trigger point voltages are accurately determined by digital
thumbwheel switch settings.
3. By calibrating out differential delay between the start and stop
channels.
A second calibration step lets the operator calibrate out all of the
differential delays in the system including those in the time interval
counter start and stop channels as well as the interconnecting
cables. An additional control position on the HP 5363B inserts a
fixed 10 ns delay in the stop channel so a counter such as the
*
54
Active Probes improve precision of Time Interval measurements,
HP Journal, October 1975, Volume 27, Number 2, pp. 11-16.
HP 5345A which requires a minimum of 10 ns between a start pulse
and the following stop pulse can be used for measurements down to
zero nanoseconds. Figure 35 illustrates this procedure.
Calibrate
Connectors
Input Probes
"A" Probe
0
Internally
generated
fast rise
calibrate
signal.
Routed to
both Start
and Stop
channels.
+0.5
0
–0.5
Start
Stop
Output to time interval
counter into 50Ω load
VREF
Stop
Time Zero
–0+
"B" Probe
VREF
Stop
10 ns
Pull to
add
10.00 nsec
+0.5
0
–0.5
Subtract 10 ns
from measurement
+0.2
0
–0.2
≈
Volts
a. Block Diagram
≈7 µsec
≈49 µsec
b. Internally generated time calibration signal appearing of calibrate
jacks on the front panel. Incorrect setting of the SLOPE switches
of the counter will result in a counter display of approximately
7 µsec, the pulse width, or 49 µsec, the space between pulses.
Figure 35.
Calibrating TIME
ZERO of the
HP 5363B to
compensate for the
differential delay
between the Start
and Stop Panels of
the Measuring
System.
4. By offering greater dynamic range than counter inputs.
The start and stop channels of the HP 5363B are designed for all
input signals including those with large voltage vs time excursions
so will accept signals over the range of –10 volts to +10 volts without
sacrificing sensitivity. This is particularly important when triggering
from many of the currently used IC logic levels. Logic swings are too
large for the X1 range of most counters but sensitivity is sacrificed
on the X10 or X20 range.
5. By offering remote programming for system applications.
Option 011 to the HP 5363B provides full remote programming
capability including probe selection, SLOPE, POLARITY, LEVEL,
and LEVEL CALIBRATION by a controller on the HP Interface Bus.
The TIME ZERO control is not bus programmable; however, the
differential delay can be measured and software adjusted in a
system.
55
Level Calibration
1. Depressing the LEVEL CALIBRATION switch first sets both the
START LEVEL and STOP LEVEL voltages (normally determined by
thumbwheel switches) to zero and grounds the tips of both PROBE A
and PROBE B which are inserted in the calibration connectors.
Calibration can also be done by grounding the probe tips through
50 ohms or less in systems applications, as it may not be convenient
to use the calibration jacks.
2. The HP 5363B employs two DACs (digital to analog converters) for
trigger point setting and self calibration.
a. The main DAC, controlled by a thumbwheel switch, supplies
trigger level voltages settable from +9.99 to –9.99 V in 10 mV
steps.
b. The offset DAC, used to calibrate out residual offset voltages of
the two probes, is initially offset from the main DAC by +75 mV,
and covers the range from +75 mV to –75 mV in 1 millivolt steps.
c. When calibrating for a positive slope, the LEVEL calibrate switch
loads zeros into the main DAC and zeros (represent +75 mV) into
the offset DAC.
Internal calibration circuits then scan the offset DAC voltage
output which determines VC from +75 millivolts downward in
1 millivolt steps until the input channel just switches triggering
at zero volts as in Figure 36. This voltage VC is stored digitally and
added to the digitally selected trigger voltage from the
thumbwheel controlled DAC to give a true trigger point voltage
with respect to zero volts dc for subsequent measurements.
d. When calibrating for a negative slope, the process is similar
except the offset DAC starts from –75 mV and scans upward.
56
Input
Grounded
for Level
Calibration
Probe
Output
Trigger
Level
Output
Ref
Voltage
Millivolts
+75
Offset
DAC
Voltage
a.
0
Operate
Hysteresis Window
+75 mV
0
Level
Cal
–75
Offset D to A converter scans VC down
from +75 in 1 millivolt steps.
VC
VC
Trigger at
Zero volts on
Positive Slope
VC
from
Plus V Thumbwheel Switch
b.
Hysteresis window scans down until upper limit just crosses
zero volts (ground) to trigger the circuit.
Figure 36. Level Calibration of one channel of one probe to trigger on a positive slope.
3. The TRIGGER LEVEL OUTPUT voltage appearing at a rear BNC
connector (Figure 37) is the trigger voltage set on a thumbwheel
switch plus the calibration voltage supplied to the start or stop
channel by the offset DAC. Therefore, this level could be off as
much as ±75 millivolts (the range of the offset DAC) from the actual
trigger voltage. This voltage can be displayed on an oscilloscope
along with the input signal to get an approximate idea of the trigger
point location as an aid to initial set up. In practice the offset will be
much less for an HP 5363B operating in a moderate environment.
(Nominal offset voltage is –10 mV to –30 mV at room temperature.)
The offset DAC voltage is present at the TRIGGER LEVEL OUTPUT
on the rear panel after calibration when the associated thumbwheel
LEVEL switch is set 0.00 and can be measured with a dc voltmeter.
4. LEVEL CALIBRATION must be done if the HP 5363B has been
turned off or at any time the SLOPE switch or PROBE selector
settings are changed or trigger voltage errors as great as ±75 mV can
occur.
5. An easy way to determine spacing between the hysteresis limits is
to go through the normal calibration procedure using a plus slope
( ) and measuring the voltage output from the TRIGGER LEVEL
OUTPUT connector, then repeat the same procedure with the
channel set to trigger on a minus ( ) slope. The difference between
these two voltages is the spacing between the hysteresis limits to
within a few millivolts.
57
Time Zero Calibration
Figure 35a (page 54) shows how the TIME ZERO calibration is made
using an internally generated fast rise signal, Figure 35b. This calibration
can also be done by using an external signal with fast transition time
(>100 V/µsec slew rate) and setting both channels to the same trigger
point (slope and level) preferably near the center of the input transition.
Range
+9.99 to –9.99V
in 10 mV steps
Programmable
(except delay
adjust) via
HP Interface Bus
{
{
Status Lights
Indicate probe selection
Select
Slope
Self Calibration
• Amplitude accuracy
• Delay error elimination
Low Capacitance Active Probes
Take HP or Tektronix® R.F. Oscilloscope
Probe Accessories
Main output to
counter, +0.5 to –0.5V
50Ω. Use counter
in “PRESET”
{
{
{
Trigger level outputs
for Oscilloscope
Figure 37. HP 5363B Controls and Connectors
58
Select Probe A,
Probe B, or both
Programming Capability
via HP Interface Bus
To Make a TI Measurement Using the HP 5363B
First determine that the signal is within the range of the HP 5363B.
1. Maximum input voltage to the HP 5363B probes without damage is
±30 volts dc + peak ac! Minimum pulse width is 5 ns and maximum
repetition rate is 10 MHz. Other considerations are discussed in
Figure 38. The trigger point is adjustable over a ±9.999 volts range in
10 mV steps either by thumbwheel switches or via the HP interface
bus using Option 011.
Trigger points should be below the peak
pulse amplitude by 8% of peak-to-peak
amplitude or 100 mv minimum to reduce
energy effects.
a. Accurate triggering precludes
triggering at the extremes of a
Start
waveform due to noise on slow
rise pulses and “energy effect”
on fast rise pulses.
9.99 Vmax to –9.99 Vmin linear range
Stop
0
5 nsec
min
Trigger Level
8% or 100 mV*
Trigger Level
8% or 100 mV*
5 nsec
min
* Within specified accuracy range,
whichever is greater
b. Signal must remain beyond the
trigger point for a minimum of
5 ns. After triggering the signal
must swing beyond the trigger
point in the opposite direction
for a minimum of 5 ns. At maximum rate the HP 5363B will trigger on every cycle of a 100 MHz
symmetrical square wave or sine
wave of sufficient amplitude.
Figure 38.
Input Signal
requirements for
the HP 5363B.
CAUTION
KEEP PROTECTIVE GROUNDING CAPS ON PROBES WHEN
NOT IN USE OR WHEN SHIPPING TO PREVENT DAMAGE FROM
STATIC CHARGES.
2. Set Stop and Start PROBE selectors to A or B, SLOPE or ,
polarity + or – and LEVEL of each channel as required for the
measurement of interest. Perform LEVEL calibration as discussed
earlier in D.(d).
3. Connect the probe or probes to the input signal and make the
measurement.
NOTE
Perform LEVEL calibration any time the PROBE selector or SLOPE
switches are changed as the offset calibration for the new condition
is not made unless this is done. Calibration may not be necessary
when measuring high amplitude input signals as maximum trigger
voltage error without voltage calibration < ±100 millivolts.
4. Make your measurement with confidence!
59
Some Applications of Time Interval Measurement
Simple Timing System with Start-Stop
Pulses Generated by Mechanical Switches
The simplest time interval measuring system consists of a time interval
counter, a voltage source, two mechanical switches, and two resistors
connected as shown in Figure 39.
Even with this simple circuit of Figure 39 precautions must be taken to
get valid measurements.
TI Counter
K1
Start
1.5 Volts
– +
Start
K2
R1
10K
Coax Cable
Stop
Stop
R2
10K
Figure 39. Simple
Time Interval
measuring System
1. The SEPARATE-COMMON switch on the counter must be set to
SEPARATE or both inputs will be tied together internally.
2. The voltage at the counter inputs goes from 0 to +1.5; therefore the
LEVEL switches must be out of the PRESET position and set to
trigger at about 0.7 volts. The system may work in the PRESET
position; however, triggering could be erratic as the operator is
depending on noise and overshoot caused by switch operation to
generate a signal that will cross the lower hysteresis limit.
Alternately, the inputs may be ac coupled; however in this case, it is
vital that resistors R1 and R2 be used. If not, the counter will trigger
only once the first time a key is closed but will not trigger on subsequent closings.
60
This happens because the coupling capacitor (either in the counter
or in the external circuit) is initially discharged so it charges and
triggers the counter the first time the key is closed. With the key
open this charge leaks off very slowly as the only discharge path is
via the leakage resistance of the capacitor so discharge may take
from minutes to hours depending on the quality of the capacitor.
Until this charge is dissipated, the circuit cannot be retriggered. The
10k resistor R1 or R2 insures the coupling capacitor discharges in a
small fraction of a second once the key is opened, consequently the
input circuit will trigger each time the key is operated.
This system works very nicely using microswitches, telegraph
keys, or other mechanical switches to START and STOP the count.
Start and Stop must be from separate switches and the counter
SEPARATE/COMMON switch must be in SEPARATE for correct
operation.
3. This system will generally not make a pulse width measurement
with the counter set in COMMON (both Start and Stop input connected together internally) and the time interval controls set to Start
on a voltage rise (positive slope when the switch is closed) and stop
on a voltage fall (negative slope when the switch is opened). It does
not work because mechanical switches exhibit contact bounce so
the counter displays the time of the first bounce rather than the
times the switch was closed. This is discussed under Delay
Control and illustrated by Figure 22a.
For counters without a built-in delay circuit in the start channel, one
can be assembled from a single SN74121 or similar integrated
circuit, a 5-volt dc supply, one resistor, and one capacitor. The “on”
time of this delay circuit, is made greater than the contact bounce
time of the start channel by selecting appropriate values of R and C.
A 1-µfd capacitor and 10k to 20k resistor gives an “on” time of 10 to
20 milliseconds which is adequate for most mechanical switches or
relays. A delay is not needed in the STOP channel as “off” time is
controlled by the SAMPLE RATE control on the counter.
This same delay circuit is also needed to count the number of
operations of a mechanical switch using a counter in a TOTALIZE
mode. Without it, the counter counts and displays each contact
bounce as well as each desired operation.
Phase Measurement
Phase measurement as made with an electronic counter is an important
special case of a time interval measurement. First the period a-b is
measured, then the Time Interval a-c. From this information phase is
calculated using the formula:
Phase in degrees =
t0
× 360
T
61
The general setup as in Figure 40 is the same as for any other time
interval measurement; however, several factors must be considered to
insure the greatest possible accuracy in phase measurement.
Period of Ref
T
Reference Phase
Unknown Phase
a
c
t0
b
where:
T = Period of reference phase in µsec
t0 = Time difference in µsec between the positive
going zero axis crossings of the reference
phase (START) channel and the unknown
phase (STOP) channel.
Counter stops counting
Counter starts counting
1 . For any stable phase measurement, the frequency of the reference
phase and the unknown phase must be identical.
2. Although a phase measurement could be made between any corresponding points on the two signals this method works only if both
signals have the same constant amplitude otherwise the measurement will not yield absolute phase.
3. Set both the START and STOP channels to trigger at zero volts using
the procedure of Figure 14 to get absolute phase. Triggering at zero
volts also has the added advantage that the phase measurement is
independent of input signal amplitude. When triggering at any point
other than zero, amplitude variations to either input channel show up
as changes in the phase as measured by the counter. The HP 5363A
Time Interval Probes are ideal for phase measurement as an internal
calibrate procedure plus setting both the START and STOP
thumbwheel switches to zero assures triggering at precisely zero
volts dc.
4. Measure the time (T), Figure 40, by making a period average measurement. A period averaging measurement reduces the trigger error
below what it would be if the time interval mode were used since
period measurements use the same trigger circuit and consequently
the identical trigger voltage point for both Start and Stop.
5. Use the highest input amplitude available provided it does not
exceed the internal protective clamp voltages of the counter. A
higher amplitude signal increases the slope of the signal at the trigger
point (zero volts) and thus reduces trigger error. If clamp voltages
are exceeded, however, the counter input impedance changes
drastically above the overload point and this distorts the input signal
which in turn invalidates the desired time vs. voltage information
needed for the measurement.
62
Figure 40. Phase
measurement with
a time interval
counter.
6. Error evaluation for a phase measurement: The counter accuracy
specifications for period average applies to the period T, Figure 41,
and the time interval accuracy specification applies to (t0). Phase
measurement error is ± the sum of these two errors. For most phase
measurements, the crystal oscillator accuracy is not important as it
is far outweighed by the other factors.
7. Improving resolution and accuracy. The time interval averaging
mode can be used to increase accuracy and resolution of a phase
measurement for counters that have this mode of operation.
8. Direct Reading Phase Measurements. Counters which can count an
external clock while making a TI measurement will display phase
directly in degrees if an external clock frequency,
fclock = 360 × fref, is counted instead of the internal clock oscillator.
(fclock = 3600 × fref gives 0.1° resolution, etc. The only limiting
requirement on the clock frequency, fclock, is that it must always be
below the maximum frequency the counter will accept at the time
base input.)
9. Application Note 174-9, Making Automatic Phase Measurements (to
1 MHz) With The 5345A Electronic Counter, has details on manual
and on automatic phase measurement using the HP Interface Bus
with bus compatible instruments.
Measuring Complex Pulse Trains
Complex pulse modulation patterns are used by airborne identification
and navigation systems. Requirements and specifications on these
systems are controlled by the Federal Aviation Agency in the United
States.
1. ATC (Air Traffic Control) System
The ATC (Air Traffic Control) system is used in conjunction with the
ground radar system at an airport to automatically keep track of
aircraft in the area.
Interrogation Pulse
The ground station sends out an interrogating RF pulse train as
shown in Figure 41. This signal is demodulated and analyzed by the
on board aircraft transceiver which automatically sends back a
coded reply.
63
P1
P3
P2
2 ± 0.15 µsec
0.8 ± 0.1 µsec
Mode Spacing
Mode
Mode Spacing
A
B
C
D
8 ± 0.2 µsec
17 ± 0.2 µsec
21 ± 0.2 µsec
25 ± 0.2 µsec
Fall time of pulse from
90% to 10% between 0.05
and 0.2 µsec.
Figure 41.
Three-pulse ATC
ground station
interrogation.
Reply Pulse
When the airborne transponder judges an interrogation to be valid, it
replies with a coded signal consisting of from 2 to 15 pulses. The two
framing pulses identify an aircraft on the ground radar. With increased air traffic all commercial aircraft and all aircraft flying in
congested traffic areas are now required to include automatically
encoded altitude information in their reply. Pulse timing for the
airborne system is shown in Figure 42.
Some of the tolerances are in the order of 0.5% or better so an
electronic counter is an ideal instrument for making these
measurements.
An electronic counter can be used to make time measurements on
the demodulated ATC pulse train. For the interrogation pulse train,
Figure 41, the selection of the start and stop points can be made in
terms of amplitude using the LEVEL control of the Stop Channel.
The reply pulse train, Figure 42, is much more complex so a counter
such as the HP 5345A (with ET gate control in the time interval
mode) is necessary to select the appropriate pair of pulses for the
measurement.
64
500 µsec
Min
24.65 µsec
20.3 µsec
2.9 µsec
1.45 µsec
0.5 µsec
First
Framing
First
Framing
Width
Rise Time
Fall Time
Spacing
Last
0.45 ± 0.10 µsec
Framing
0.05 to 0.10 µsec
0.05 to 0.2 µsec
1.45 µsec between each pulse accuracy ±0.10 µsec with
respect to the first framing pulse. One pulse to any
other pulse in the group except first framing pulse is
±0.15 µsec
Ident
Pulse
Figure 42.
Available ATC
reply Pulse codes.
The EXT gate of the counter can be opened at any time following the
last pulse of the previous pulse train. The measurement will then
begin with the next pulse to occur — in this case, the first framing
pulse of the pulse train of interest at (a) in Figure 43.
First
Framing
Pulse
Pulse of
Interest
Demodulation
Pulsed
Signal
Exit Gate
to Counter
a
b
0
–1
TI Measured
Time
Figure 43.
Making time
measurements
on an ATC
Transponder
replay pulse
with the HP5345A
Counter.
A pulse generator with a variable pulse width (HP 8007A or similar)
triggered by the GATE OUT of the HP 5345A is used to select the
stop pulse. The counter stops the measurement on the input pulse
following the end of the pulse from the pulse generator (–1V to 0V
transition). Pulse width setting is not critical as spacing between
pulses of the input signal is about one microsecond. A circuit as in
Figure 44 can be used to synchronize the measuring system with
respect to the first pulse in the complex pulse train shown.
65
Approx
500 µsec Min
Approx
25 µsec Max
20,300 µsec
HP 5345A TI
Pulse Train
Gate Out
+1
0
Retriggerable
One Shot mV
200 µsec
Ext Gate
HP 8007A Pulse Gen.
Ext Trigger
Out
0
–1
Pulse width adjustable
from 1 to 25 µsec
Figure 44.
Synchronizing a
complex pulse
train with respect
to first pulse.
2. DME Pulses
The DME, distance measuring equipment, used on commercial
aircraft sends out a pair of pulses which trigger a ground based
station that sends back a reply.
DME channels are spaced at 1 MHz intervals up around 1 GHz so
strict adherence to transmitted pulse shape is necessary to keep
adjacent channel energy to a minimum. All of these parameters can
be measured on a demodulated DME pulse with a time interval
counter and the HP 5363B Time Interval Probes which have digitally
set trigger levels.
Rise Time
2.5 µsec
± 0.5 µsec
Decay Time
2.5 µsec ± 0.5 µsec
Duration
3.5 µsec ± 0.5 µsec
66
0.9A
A
0.1A
0.5A
These examples cite just a few of the applications of time interval
measurement to the signals emitted by radar identification, location,
and navigational equipment.
Mode
Spacing
Airborne
Ground Station
Interrogation
Reply
12 µsec.
12 µsec.
Mode X ± 0.5 µsec
± 0.5 µsec
36 µsec.
30 µsec.
Mode Y ± 0.5 µsec
± 0.5 µsec
Figure 45.
DME Pulse
characteristic
Comparison with Other Ways of Making Time Interval
Measurements on Narrow Pulses or Fast Rise Signals
The oscilloscope was one of the first instruments used for pulse and rise
time measurements, and it has been used in that role ever since. Both
oscilloscopes and time interval counters have specific advantages in
making time interval measurements, and in many cases, the techniques
are complementary. In general, time interval counters give greater
resolution (0.00000001% on the HP 5370A) on pulse to pulse measurements and can be more readily automated. The oscilloscope gives
excellent visual feedback and recently, the introduction of ∆time
oscilloscopes has made a very real contribution to improving oscilloscope resolution for time interval measurements. For example, the
HP 1743A is capable of giving a resolution of 0.002%. ±1 count. The
visual feedback characteristic of the oscilloscopes is important when
dealing with noisy or complex waveforms as it allows a skilled operator
to concentrate only on the points of interest. If rise times are fast, a time
interval counter will not be affected by noise, but in measurements with
slow rise time, caution should be exercised.
Interpolating counters give very high single shot resolution which would
not be possible on a ∆time scope which requires a repetitive signal for
operation. In addition, this high single shot resolution of the interpolating allows the collection of statistics.
From the user’s viewpoint, the characteristics of the measurement and
output required will determine the most acceptable measurement
solution. The following summary may help in the selection process.
Time Interval Counter:
• Highest resolution in pulse to pulse measurements.
• Easy Automation — IEEE 488 Interfacing.
• Statistics.
• Noise effects negligible on fast rise time pulses.
• Single shot capability.
• Different trigger voltages on stop and start channels.
• Measurements above 100 MHz.
Oscilloscope:
• Excellent visual feedback
• Noisy complex waveforms may be visually filtered.
• Excellent for slow rise time measurements on noisy signals.
• Less prone to false triggering due to noise.
67
H
For more information about HewlettPackard test and measurement products,
applications, services and for a current
sales office listing, visit our web site,
http://www.hp.com/go/tmdir.
You can also contact one of the following
centers and ask for a test and measurement sales representative.
United States:
Hewlett-Packard Company
Test and Measurement Call Center
P.O. Box 4026
Englewood, CO 80155-4026
1 800 452 4844
Canada:
Hewlett-Packard Canada Ltd.
5150 Spectrum Way
Mississauga, Ontario
L4W 5G1
(905) 206-4725
Europe:
Hewlett-Packard
European Marketing Centre
P.O. Box 999
1180 AZ Amstelveen
The Netherlands
(21 20) 547 9900
Japan:
Hewlett-Packard Japan Ltd.
Measurement Assistance Center
9-1, Takakura-Cho, Hachioji-Shi,
Tokyo 192, Japan
Tel: (81-426) 56-7832
Fax: (81-426) 56-7840
Latin America:
Hewlett-Packard
Latin American Region Headquarters
5200 Blue Lagoon Drive
9th Floor
Miami, Florida 33126
U.S.A.
(305) 267 4245/4220
Australia/New Zealand:
Hewlett-Packard Australia Ltd.
31-41 Joseph Street
Blackburn, Victoria 3130
Australia
1 800 629 485
Asia Pacific:
Hewlett-Packard Asia Pacific Ltd.
17-21/F Shell Tower, Time Square,
1 Matheson Street, Causeway Bay,
Hong Kong
Tel: (852) 2599 7777
Fax: (852) 2506 9285
Data Subject to Change
Printed in U.S.A. June 1997
68
Hewlett-Packard Company
Copyright © 1997
5965-7663E
">