Philips PM 6670 User manual

Timer-Counters
PM 6670. . .72
Operating Manual
PHiUPS
« ' tufwasB itm , ac m
A- -A
PHILIPS

important
As the instrument is an electrical apparatus, it may be operated only by trained personnel. Maintenance and
repairs may also be carried out only by qualified personnel.
Please note
In correspondence concerning this instrument, please quote the type number and serial number as given on
the type plate.
Contents
1. Technical specification 4
2. Installation instructions 9
3. Operating instructions 10
Frontpanel controls 10
Rear panel controls 11
Theory of measurements 12
Practical measurements 18
4. Performance check 21
Bedienungsanleitung 25
Table des matieres 47
©Philips Export B.V. — Eindhoven — The Netherlands — 1984.
Printed in Sweden

1. Technical Specification
Measuring modes
Frequency
Range: 0,1 Hz. . .120MHz (PM 6670, 6671).
0.1 Hz. . .1GHz(PM 6672).
Mode: Input-signal synchronized, high
resolution computing measuring method
(reciprocal).
Signal mode: ON. SINGLE BURST, MUL
TIPLE BURST FREQUENCY AVERAGE.
In the FREQUENCY AVERAGE mode
(rear panel selectable), the counter
measures the average of a multiple of
frequency samples. Samples are taken
with external gate control (s^soons) and
totalized during the selected measuring
time (10ms. . .96s) to allow multiple burst
frequency measurements or to sample
frequency sweep profiles.
LSD displayed: 10"'Hz. . .10^Hz(PM 6670,
6671). 10-'Hz. . .10<Hz(PM 6672), de
pending on measuring time and input
frequency. At least 7 digits displayed
per second of measuring time.
Resolution: LSD*.
Inaccuracy (rel.error):
. resolution . trigger* error
: rel
FREQ
base error.
measuring time
. time
Period average
Range: 100ns. . .100s
LSD displayed: 10"'^. . .10"®s, depending
on measuring time and period duration.
At least 7 digits displayed per second of
measuring time.
Resolution: LSD*.
Inaccuracy (rel.error):
. resolution trigger* error
: rel
PERIQD
base error.
measuring time
Time Interval A to B ; single
Range: 100ns. . .10®s.
LSD displayed: 10''s. . .10°s.
Resolution: LSD*.
Inaccuracy (rei.error):
± resolution ± trigger* error
± rel
TIME INT
base error.
. time
Pulse width A
Pulse width measurements are similar to
single time interval measurements. Both
start and stop triggering take place in
channel A, with common trigger level
setting and automatic trailing edge trigger
slope inversion with respect to selected
leading edge trigger slope.
All other specifications are identical to
single time interval. High resolution, pulse-
width measurements on narrow pulses
can be made in the time interval average
mode with manual selection of common
source and stop slope polarity.
Time Interval A to B; average
Range: 0ns. . .100s.
LSD displayed: 10''^s . . .10"®s, depending on
measuring time and time interval.
Resolution:
10-'s
VTJ
Inaccuracy (rei.error):
4 ns ± resolution trigger* error
TIME INT VTJx TIME INT
± rel. time base error.
Number of Intervals averaged (N):
Measuring time x interval repetition rate.
Minimum dead time from stop to start:
300 ns.
Note: Input signal must be repetitive and
asynchronous with respect to the time
base.
' or, 1 LSD unit, whichever is greather.
. time
Count A
Totalize range; 1. .10", with indication of
M-pulses and G-pulses beyond the 10®
display range.
Pulse pair resolution: 80 ns.
Mode:
Manual: Start-stop by DISPL. HOLD push
button. Sequential start-stop periods are
accumulated or individually totalized after
reset.
External: Totalizing interval controlled via
input B, selectable: count A gated during
pulse duration on B or count A gated
between start and stop pulse on B.
Inaccuracy: Pulse repetition rate A x
trigger* error B.
Phase A—B
Phase is the result of a simultaneous time
interval average and period average
measurement. The maximum phase range
is therefore limited, due to the 300 ns
dead time between stop and start of the
next time interval, and consequently
frequency dependent. For phase measure
ments between 180° and 360° on high
frequency signals, it is recommended to
measure the complementary phase
0°. . . 180° by changing the input leads
(phase B to A).
Phase range: 0.1°. . .360°x
|1-(300nsx FREQ)1 ,
Example 0.1°. . . 359.995° at 50 Hz
0.1°. . 357.3° at 25 kHz
0.1°. . . 180° at 1.666 MHz
Frequency range: 0.03Hz. . . 1.6 MHz
LSD displayed: 10"® . 10'^ degrees, de
pending on measuring time and phase
difference.
„ . .. 10-'sx FREQ X 360°
Resolution: -=
VTT
Inaccuracy (rei.error):
^ 4 ns X 360° x FREQ ^ resolution
PHASE ~ PHASE
trigger* error x FREQ x 360°
Vrpx PHASE
Number of Input cycles averaged (N):
Measuring time x FREQ A.
Minimum Input signal: 100 mVfmg.
Phase jitter tolerance: The tolerated phase
jitter peak-value ?£ phase difference, i. e.
phase jitter around 0° should not cause
consecutive measurement results of slight
ly above 0° and slightly below 360?, since
the toltal measurement result is the statisti
cal average of all individual results.
Ratio
Range:
FREQA 0. . .10MHz
(all models)
FREQ B 0. . .10MHz
LSD displayed: 10"'. . . 10°
Resolution: LSD*.
Inaccuracy (rei.error):
, resolution ^ .
RATIO trigger* error B..
Note: Frequency ratio measurements
FREQ A (or C)*_ 0.01 Hz. . .120 MHz (orl GHy)
FREQD 50kHz. . .10MHz
can also be made in the frequency mode,
by making use of the external reference
input D. However, this arrangement does
not give correct decimal points.
** Conly on PM 6672.
RPM
The RPM (revolutions per minute) mode
is similar to the frequency mode. The
measured frequency is multiplied x 60,
before being displayed.
Range: 0.6 RPM. . .10°RPM (with one pulse
per revolution).
LSD displayed: 10"' RPM . . . 10° RPM, de
pending on measuring time and RPM.
Resolution: LSD*
Inaccuracy (rei.error):
resolution trigger* error
± rel
RPM
base error.
measuring time
. time
' see definitions.

Input and output
specifications
Inputs A and B
Frequency range:
Decoupled: 0. .120MHz
AC coupled: 50 Hz . . .120 MHz
Rise time: Approx. 4 ns.
Sensitivity:
Decoupled: lOmVprps sine wave or
30mVpp(0. . .75MHz).
20mVrms sine wave or
60mVpp(75. . .120MHz).
AC coupled: lOmVrms- -lOOrnVrms sine
wave(50 Hz. . .75MHz).
20 mVcnig. . . 100 mVrrps sine
wave(75. . .120MHz).
Attenuation: x 1 /x 10 fixed.
For frequency related measurements, tfie
fixed attenuator can be used in combina
tion with the continuously variable attenu
ator x1. . . xlO (AC coupled position).
Noise immunity/tiysteresis band:
Decoupled: approx. 20mVpp/200mVpp.
AC coupled: approx. 20mVpp. . .2Vpp.
Dynamic input voitage range:
Decoupled: 30mVpp. . .5Vpp/
300 mVpp. . . 50 Vpp.
AC coupled: lOmV^ms - -SVrms/
100 mVrms ■ ■ 20 Vrms-
Trigger levei:
Decoupled: -2.5V. . .-1-2.5V/
-25 V. . .-1-25 V.
AC coupled: fixed OV; level control acts
as continuously variable attenuator, which
is more suitable for frequency related
measurements.
Trigger level output, not available on PM 6670:
Set trigger voltages -2.5 V. . .-1-2.5 V, avail
able on 1 mm jacks at the front for moni
toring of set trigger level.
Trigger indicators, not available on PM 6670:
Tri-state LED trigger lights to indicate
trigger status:
On: trigger level is too low.
Blinking: triggering occurs, input signal
crosses hysteresis band.
Off: trigger level is too high.
Coupling: DC/AC
Impedance: Approx. 1 MOhm//35pF, inde
pendent of sep/com switch position.
Ct^annel input: Separate A and B or
common A.
Noise filter: Switchable 50 kHz Low pass
filter in channel A. Noise suppression
approx. 40dB at 1 MHz.
Maximum voltage wittiout damage:
DC; 300 V.
AC; 260 Vrms al -440 Hz declining to
12Vrms a' -1 MHz (in ATT x 1 position),
260Vrms (in ATT x 10 position).
input C, PM 6672 only
Frequency range: 70MHz. . .1GHz.
Operating input voitage range:
15mVrms. .12Vr,ms( 70 MHz. . .800 MHz).
2o mVpms. . . 12 Vrms (800 MHz. . . 1 GHz).
Impedance: 50 Ohm nominal; VSWR <2.
Coupling: AC.
AM tolerance: 98%, minimum signal must
exceed minimum operating input voltage.
Maximum voitage without damage: 12 Vrms;
overload protection with PIN diodes
Ext. reference and
Ratio input (channel D),
not available on PM 6670
Frequency range: 1 kHz. . .10MHz.
Sensitivity: 500mVrms
Impedance: Approx. 2 kOhm
Coupling: Ac
Max. voltage without damage: 25 Vrms
Note: As external reference frequency,
only 10 MHz will give correct decimal point
and unit indication. With the optional
frequency multiplier PM 9697 references
of 1 and 5 MHz can also be accepted.
Internal standard output
(channel D),
not available on PM 6670
Crystal frequency: 10 MHz.
Output level: LS-TTL compatible.
Output impedance: Approx. 400 Ohm.
Coupling: DC.
Overload protection: Short-circuit proof.
Ext. arming/Freq-avg/
Reset (channel E),
not available on PM 6670
A 3-position rear panel switch gives
choice of external control over:
ARMING: In this position, the counter is
prevented from starting a new measure
ment when input E is high. A high-to-low
going pulse arms the counter to start a
new measurement.
Note: Arming not applicable in COUNT A,
manual mode.
FREQUENCY AVERAGE: Frequency
measurements (max. 100MHz) and period
measurements are interrupted when
input E is high. The measurement is con
tinued again when input E is low. Each
individual frequency sample must contain
at least 20 pulses (FRED mode) or 2
pulses (PERIOD mode).
The effective measurement time (defining
resolution and accuracy) is the sum of
external gate times that occurs during the
selected measurement time.
EXT. RESET-START: Electrical reset,
equivalent to the front panel RESET push
button. (See HOLD and RESET). Counter
is reset when input E goes high. A new
measurement can be made after input
E has returned low.
input ieveis:
High: '^ 2\/.
Low: =£0.5V.
Input impedance: Approx. 2kOhm.
Max. input voltage without damage: ±25 V
Minimum puise duration:
Arming and frequency avg: 500 ns.
External reset: 200>js.
Gate monitor output (rear),
not available on PM 6670
The gate status monitor output permits
observation on an oscilloscope of the
measured time interval and the trigger
hold-off time (PM 6671 only).
Output level:
Main gate open: approx. 0.4 V.
Hold-off active: approx. 1.2 V.
Main gate closed: approx. 2.5 V.
Output impedance: Approx. 1.5 kOhm.
Delay: Internal delay between actual trig
gering and gate monitor output is approx.
150 ns.
Overload protection: Short circuit proof.
Auxiliary functions
Measuring time
The measuring time is "continuously" vari
able (33 steps/decade): 10ms. . .96s. with
clear setpoints at 10ms, 100ms, Is, 10s
and 96s. Selected measuring time is dis
played, without any delay, when depres
sing the measuring time control.
The actual measuring time equals the
selected measuring time plus the time
needed to synchronize the measurement
with an integer number of cycles of the
input signal (Reciprocal measurements
are synchronized with multiples of 10 in
put cycles).
In the FREQUENCY AVERAGE mode,
the measuring time can be externally con
trolled to make burst frequency average
measurements.
Hold-off PM 6671 only
With trigger hold-off activated, the counter
ignores re-triggering (channel A) or stop
triggering (channel B) during the set
hold-off time. The hold-off time can be
digitally measured by pressing CHECK.
Applicable in all time modes.

Range: 200>js. . . 200 ms, in period, time
interval arid pulse width mode, the set
hold-off time is visible with CHECK de
pressed.
On/off indication: LED indicates when
hold-off is activated.
Monitor: The selected hold-off duration can
be made visible via the gate monitor
output
On Stand By
In "ST BY" position, power is available to
maintain an ovenized crystal oscillator
heated and to recharge the optional bat
tery pack.
Check
10 MHz internal reference connected to
logic circuitry. Self-test of most measuring
functions can be selected. By using this
mode, the COUNT function provides a
stop-watch facility
Display hold
Depressing "DISP HOLD" button sets
display time to infinite and freezes the last
measurement result A new measurement
can be initiated using reset.
In the COUNT mode, the "DISP HOLD"
control is used to start and stop manual
totalizing.
Reset
Manual via pushbutton or electrical via
input E.
General
Display
Read out: 8 digits, 7,6 mm high-efficiency
LED's. Microprocessor control of display
format, decimal point and unit indication:
Hz, kHz, MHz, GHz, ns, ps, ms and s.
Display lime: Continuously variable
80ms. . .96s plus DISP HOLD.
Gate lamp: Indicates that main-gate is
opened and measurement takes place.
ST BT: Stand-by indication with LED when
instrument is not switched ON.
REMOTE:** Indicates when control over
counter is taken by the installed BUS
interface option (lEC 625 — IEEE 488).
Low-battery:** Indication by blinking display
some 15 min. before recharging is needed.
Power requirements
In addition to the normal line voltage
supply, the PM 6671 and PM 6672 can
also be powered from an optional battery
pack or external DC voltage.
Line: 115/230V ±15%; 45. . .440Hz;<25 VA,
Internal battery unit:** PM 9693.
External DC Cource:**
Voltage: 11 8V. 28V; 4.5. 8W depending
on version and options installed.
Connector: Battery jack fitting DIN 45323.
Line interference: Below VDE 0871 (B) and
MIL STD461
Safety: According to I EC 348 and
CSA 556 B.
** not available on PM 6670.
Dimensions and weight
Width: 210 mm (8.25 in)
Height: 89 mm (3.8 in).
Depth: 280 mm (11.0 in).
Weight:
Net: approx. 2.5 kg.
Shipping- approx 3.6 kg.
Environmentai conditions
Temperature:
Rated range of use: — 5°C. . . + 50°C
Storage and transport: — 40°C. . . + 70°C,
Humidity:
Operating: 10 . 90?/, RH, no condensation.
Storage: 5 , 95%RH.
Altitude/Barometric pressure
Operating. 5000m(15000ft)-53.3kN/m7
Storage: 15000 m (50000 ft)- 15.2 kN/m^.
Vibration test: According to I EC 68 Fc.
Bump test: According to lEC 68 Eb
Handling test: According to lEC 68 Ec.
Transport test: According to NLN-L88
For TIME INTERVAL AVG: LSD
2.5 X 10"'S
Definitions
LSD displayed
Unit value of Least Significant Digit,
displayed.
For FREQUENCY, PERIOD AVERAGE,
RPM and PHASE:
2.5
LSD
X
=
measuring time
FREQ or PERIOD or RPM or 360°
10'Hz
For RATIO: LSD =
_ 2.5 X prescaling factor (P) x RATIO
measuring time x FREQ A or C
(P) = 1 Channel A, all models
(P) = 256 Channel C, PM 6672.
For SINGLE TIME INTERVAL and
PULSE WIDTH:
LSD = 100ns (for times <10s).
LSD = (for times ^i^lOs).
10®
meas. time x time int. rep. rate in Hz
All calculated LSD s shall be rounded to
nearest decade (e g 5 ns will be 10ns and
0.4 Hz will be 0.1 Hz) and cannot exceed
the 8th digit
Resolution
For multiple event measurements:
FREQUENCY. PERIOD AVERAGE, RPM
and RATIO, the resolution is the smallest
increment between two measuring results
being most often 1 LSD unit. Due to arith
metic truncation, the resolution can be
2 LSD units if:
LSD X measuring time
FREO oTp^RlOD w'^M or'RAT 10
but can then be reduced to 1 LSD unit,
by doubling the measuring time.
For single event measurements:
PULSE WIDTH and SINGLE TIME INTER
VAL, the measuring resolution is 100 ns
(one clock pulse) The counter can accu
mulate up to 10'® clock pulses, of which
only the 8 most significant digits are shown
For statistical measurements:
TIME INTERVAL AVERAGE and PHASE,
the measuring resolution is the smallest
increment between two measuring results,
with confidence level of 95%.
Trigger error
Trigger error is the absolute measurement
error due to noise on the input signal
causing a too early or too late triggering
For any waveform (FREO, PERIOD, RATIO,
RPM and PULSE WIDTH):
peak-to-peak noise voltage
signal slope (V/s)
For sinewave (FREO, PERIOD, RATIO, RPM):
1
FREO X TTx S/N ratio
Example: For S/N ratio of 100 (40 dB)
and 1 second measuring time, the trigger
. .3x10"®
error is.
For separate source TIME INTERVAL:
peak noise voltage (input A) ^
signal slope A (V/s)
peak noise voltage (input B)
signal slope B (V/s)
For PHASE:
1
FREOx27r
1
X
1
S/N ratio (input A)
S/N ratio (input B)
Note: S/N ratios calculated with peak-to-
peak signal and noise values.

Accessories
Supplied with the Instrument:
• Line power cord
• Fuse, 1.6A fast-blow
• Front cover
• l\/lanual
To be ordered separately:
PM 9678: TCXO, 1 XlO-'/montti,
Included in version /02.
PM 9679: Proportionally oven controlled
oscillator 1 xiO'Vmonth.
Included in version /03.
PM 9690: Proportionally oven controlled
oscillator 1.5XlO-724h.
Included in version /04.
PM 9691: Proportionally oven controlled
oscillator 5 xi0''°/24h.
Included in version /05.
PM 9693: Battery unit.
PM 9694: BCD output and display offset unit.
PM 9695: Analog recorder output (DAG).
PM9696: IEC625/IEEE488 BUS interface.
PM 9483/50: lEEE-to-lEC adapter.
PM 9487/10: IEEE cable, 1 m.
PM 9487/20: IEEE cable, 2m.
PM 9487/40: IEEE cable, 4 m.
PM 9697: External reference frequency
multiplier.
PM 8923: 120MHz, 1Mohm probe set,
1:1 and 1:10.
PM8943: 650 MHz, 50ohm/1 Mohm FET
probe set, 1:1-10-100.
PM 9639: 1.5GHz, 500ohm probe set 1:10.
PM 9581: 50ohm feed-through termination,
1W.
PM 9585: 50ohm feed-through termination,
3W.
PM 9074: Coaxial cable. 50ohm, BNC to
BNC, 1m.
PM 9588: Set of 15 coaxial cables, 50ohm,
BNC to BNC.
5 cables (20.7 cm), 4 cables (40.5 cm),
3 cables (60.3 cm), 3 cables (198.6 cm).
PM 9669/01: 19" rack mount adapter to fit
one instrument.
PM 9669/02: 19" rack mount adapter to fit
two instruments,
PM 9672: Carrying case.
NOTE: The timebase oscillators, PM 9678,
-79, -90 and -91, can also be ordered sepa
rately for later upgrading of the counters.
The counters can not simultaneously be
equipped with more than one of the follovi/-
ing options: PM 9693, PM 9694, PM 9695
and PM 9696. The multiplier PM 9697 can
only be instal led simultaneously with the
.'01 osci llator. In 702, . ,705 versions the
osci llator must be removed before a
PM 9697 can be plugged in.

2. Installation Instructions
General information
This counter has been designed and tested in
accordance with lEC Publication, Safety Require
ments For Electronic Measuring Appartus For Class 1
Instruments, and has been supplied in a safe
condition. The present manual contains information
and warnings that shall be followed by the user, to
ensure safe operation and to retain the counter in a
safe condition.
Before connecting the counter to the line (mains),
visually check the cabinet, controls, connectors, etc,
to ascertain whether any damage has occurred in
transit. If any defects are apparent, do not connect
the counter to the line. All components on the primary
side of the line transformer are CSA approved and
should only be replaced with original parts.
In the event of obvious damage, missing parts or
if the safety of the counter is suspected, a claim
should be made to the carrier immediately. A
PHILIPS sales or Service organisation should also
be notified in order to facilitate the repair of the
counter.
External battery operation
(PM 6671 and PM 6672)
For field applications, PM 6671 and PM 6672 can
be operated from an external 11 .8...28Vd(s supply,
connected to the EXT BATT socket.
Connecting the counter to both the line and an
external battery at the same time, gives a power
back-up facility that maintains heating of the oven
oscillator and recharges the optional internal battery
pack PM 9693 when fitted.
Fuses
The counter is protected by a thermal fuse, located
in the line transformer and a secondary fuse, 1.6A
fast-blow (not PM 6670) on PCB U1. Remove the
line plug before fitting a fuse. Ensure that only fuses
of the specified type are used. If the counter is set
for operation on 11 5V line voltage, but is connected
to a 220V supply, the thermal fuse will blow im
mediately to protect the counter.
Grounding
The counter is connected to ground via a three-core
line cable, which must be plugged into a socket outlet
with a protective ground contact. No other method
of safety grounding is permitted for this counter.
When the counter is brought from a cold to a warm
environment, condensation may cause a hazardous
condition. Therefore, ensure that the grounding
requirements are strictly met.
Any interruption of the protective ground, inside
or outside the counter is dangerous. Line extension
cables must always have a protective ground
conductor.
Opening of the cabinet
The counter shall be disconnected from all voltage
sources before it is opened. If adjustment or main
tenance of the counter with the covers removed is
inevitable, it shall be carried out only by a qualified
person, who is aware of the hazard involved. Bear in
mind that capacitors inside the counter may still
retain their charge, even if the counter is discon
nected from all voltage sources.
Opening of the cabinet or removing of parts, except
those to which access can be gained by hand, is
likely to expose live parts and accessible terminals
that can be dangerous to life.
Line voltage setting
Before connecting the counter to the line, ensure that
it is set to the local line voltage. On delivery, the
counter is set to either 11 5V or 220V, as indicated
on the line voltage selector on the rear panel. If the
voltage setting is incorrect, set the line voltage
selector in accordance with the local voltage, before
connecting the counter to the line.
Type
Thermal fuse
1.6A fast-blow fuse 5x20mm
Service code
number
4822 252 20007
4822 253 20022
Operating position
The counter can be operated in any desired posi
tion. A fold-down tilting handle can be rotated and
locked in several fixed positions by first depressing
the knob in the centre of each hinge.
Front cover
The front panel controls and connectors can be
protected by a plastic snap-on front cover.
Cleaning the counter covers
The top and bottom covers, lacquered with suede
coating, need special treatment if the surface gets
soiled. The 3M Company has developed a "Doodle
bug Pad" (Catalog l^o. 8440) which when soaked
in water, ethanol or common household cleaning
agent, will penetrate holes and pores to restore its
former lustre.
Abrasive cleaning pads will result in surface
scratches. Although the Nextel suede coating is
ethanol resistant, it is susceptible to methyla
ted spirit, which could damage the surface due
to one of the denaturing substances present.

10
3. Operating Instructions
PM 6671 high resolution timer/counter 120MHz
PHILIPS
TRIGGER LEVEL
SENSITIVITY
O MHi
O GHi
- Z 5V
OO gate MIN
POWER CHECK RESET DISPL
HOLO
OOOO-
I ON
M ST BY
MEASURING TIME
PULSE
WIDTH A
COUNT A
MANUAL
TIME INT L-
SINCLE A-B
AVERAGE A-B
' 96s
PUSH TO READ
PERIOD A
— RPM A
REMOTE ATT SLOPE SOhHi COM SLOPE ATT
kIO FILTER v<a A "L. *10
PHASE A-B
' IN DEGREES
RATIO A/B
MAX 260V_,
KOKI 51
FRONT PANEL
POWER
Supplies power to the counter in the ON, depressed position.
In the ST BY, released position, the counter is switched off, but
power is still available for an optional oven oscillator and a
rechargeable battery. A decimal point marked STBYindicates
the stand-by mode.
This is a secondary power switch. Even in the stand-by
mode the counter contains live conductors and parts.
The line cord (mains lead) must be removed to
disconnect power from the counter.
CHECK
When depressed, the internal 10MHz standard signal is
connected to the logic circuits. In conjunction with the function
selector rotary switch, CHECK enables a self-test of most
measuring functions.
RESET
When depressed, resets the counter and blanks the display.
On release, RESET initiates a new measurement.
DISPL HOLD
When DISPL HOLD is depressed, the display time is set to
infinity. A new measurement can be initiated with the RESET
push-button.
MEASURING TIME
PUSH TO READ
The measuring time can be set between 10ms and 96s for
optimum resolution and measuring speed.
When pushed, the set measuring time is displayed.
HOLD OFF, PM6671 only
During the set hold-off time, the counter ignores all input events
that should have ended the measuring cycle. To display the set
hold-off time, select TIME INT SINGLE A-B and press CHECK.
Function selector rotary switch
RPM A
Sets the counter to perform a revolution per minute
measurement on the signal connected to input A, provided
that the transducer sends one pulse per revolution.
FREQA
FREQ A or C, PM6672 only
Sets the counter to perform frequency measurements on the
signal connected to input A or input C. In the frequency
average mode, rear panel selectable on PM6671...72,
the counter measures the average of several frequency
samples, controlled by an external gate signal via
Input E.
RATIO A/B
Sets the counter to measure the ratio between signals
connected to input A and B up to 10MHz. To obtain full
frequency range on PM6671 ...72, a ratio measurement can
be done in the FREQ A or C mode by using input D set to
EXT STD IN. However, this method does not give correct
setting of the decimal point.
PHASE A-B
Sets the counter to measure the phase (in degrees) between
signals connected to input Aand B. Max frequency is 1.6MHz.
The sensitivity knob should be pulled and set fully clock-wise.
The ATTxlO push-button should be released.
COUNT A MANUAL
Sets the counterto totalize events (pulses or periods) on
input A during the time interval between releasing and
depressing the DISPL HOLD push-button. An event
is defined as a positive-going slope.
The result can be accumulated with another count sequence
or reset with the RESET button.
COUNT A _| 1_ Gated by B
The counter will totalize events on input A, between the leading
and trailing edge of the input B signal.
COUNT A n n Start and Stop by B
The counter will totalize events on input A, between the start
and stop event on input B.
PULSE WIDTH A
Sets the counter to measure the pulse-width of the signal
connected to input A.
TIME INT SINGLE A-B
Sets the counterto measure the time interval between pulses
on input A and B.

11
o
ARMirUC
FREQ AVERAGE
EXT RESET
INCL OPTION
THERM FUSED MAINS TRANSF
GATE
MONITOR
©
©
EXT BATT
11.8- 28
©
©
o-
©
TIME INT AVERAGE A-B
Sets the counter to measure the time interval average of a
repetitive signal that is not synchronized with the used
standard frequency.
PERIOD A
Sets the counter to measure the average period time of the
signal connected to input A.
Input amplifier
ATTxIO
When depressed, the sensitivity is attenuated by a factor 10
for input A or B. In DC mode the equivalent trigger level is
multiplied by 10.
SLOPE ©_
When depressed, the triggering occurs on the trailing edge
instead of the normal leading edge.
50kHz FILTER
A low-pass filter to improve triggering of noisy signals with
frequencies below 50kHz. Applies to input A only.
COM via A
Connects channel A and 8 internally. In this mode input B is
disconnected.
TRIGGER LEVEL
When the rotary knob is depressed, the DC mode is selected
and the trigger level is adjustable between - 2.5Vand -f-2.5V.
Tri-state trigger indicators on PM6671 ...72 light to indicate
trigger status:
On = trigger level is too low
Blinking = triggering occurs, i.e. input signal
crosses hysteresis band
Off = trigger level is too high
The trigger level output on the front panel of PM6671...72
makes it possible to measure the set trigger level.
When ATTxlO is depressed, the trigger level is adjustable
between - 25\/ and -I- 25V, but the output level is still
-2.5V...-I-2.5V.
SENSITIVITY
When the rotary knob is pulled, the AC mode is selected and
the sensitivity is adjustable between approx 10...100mVR|\/]s
IGHz, PM 6672 only
When depressed, input C is connected. The sensitivity is
adjusted automatically.
LED indicators
REMOTE
Indicates that the counter is in the remote-controlled mode
via the optional Bus Interface PM9696.
GATE
Indicates that a measurement is in progress.
UNIT INDICATOR
A multi-purpose 4-LED unit indicator.
For FREQ. read: Hz, kHz, MHz, GHz.
For PERIOD, TIME and PULSE WIDTH read: ns, ps, ms, s.
For COUNT read: ps/kHz = 10® pulses
ms/MHz = 10® pulses
s/GHz = 10® pulses
REAR PANEL
EXT STD IN/INT STD OUT
Input D
Two-position switch selection of using BNC connector D
either as an output for the internal 10MHz standard signal
or as an input for an external standard signal.
ARMING/FREQ AVERAGE/EXT RESET
Input E
Three position slide switch for selecting the functions of
input E.
GATE MONITOR
Enables observation of the measured time interval
and the trigger hold-off time on an oscilloscope.
EXT BATT
Input from an external DC source 11.8...28V.

12
Theory of
Measurements
Introduction
The microcomputer-based PM 6670...72 provide a
wide range of frequency and time measuring func-
tinons, including period average, time interval average
as well as three count modes. Resolution down to
lOps is possible with time interval average. In ad
dition, these timer/counters offer pulse width
measurements, phase delay directly in degrees and
tachometer readings i RPM.
PM 6670...72 feature automatic truncation of digits
and automatic calculation of displayed LSD ( Least
Significant Digit). Only significant digits are displayed
and no overflow can occur, except for RPM measure
ments.
A block diagram is shown in Fig. 3.1. To identical
channels are used for accurate time interval measure
ments.
£XT R£F FRtQ ^
COUNT RCC.
{INPUT CYCLES)
IEEE -psaaus
BCD OUTPUT
ANALOG OUTPUT
Fig. 3.1. Block diagram
Input triggering functions
As the input signal can have very different wave
forms, it is necessary to pulse shape the signal.
Ohterwise, the counting circuits could not handle the
signal. The input circuits consist of:
• AC/DC-coupling selector;
• an input step attenuator (xlO), to attenuate exces
sive input signals to fit the± 2.5V trigger level
off-set range;
• a switchable low-pass filter for noise rejection;
• a differential amplifier that allows trigger level
setting;
• trigger circuits with variable hysteresis band;
The functional difference between AC- and DC-
coupled input is illustrated in Fig. 3.2.
Input A, AC - coupled
Input A, DC'COupled
TCT!
A ^NSITIVITV ( HYS TERLS / 5)
(XlO)
Fig. 3.2. Block diagram of the input circuits.
In AC-coupled mode, which is the normal case for
frequency measurements, the trigger level/sensi
tivity potentiometer controls the Schmitt-trigger
sensitivity. In combination with the xlO attenuator,
this enables a variable trigger hysteresis band, thus
a variable input sensitivity (variable noise immunity).
This is essential for correct triggering for frequency
measurements of noisy signals.
In DC-coupled mode, which is the normal mode for
time measurements, the trigger level/sensitivity
potentiometer controls the differential amplifier. This
results in a continuous variable trigger level setting.
The Schmitt-trigger sensitivity is not influenced, but
is set to minimum. It is of utmost importance to
have as narrow hysteresis band as possible for time
measurements.

13
Schmitt-trlgger function
The Schmitt-trigger function is illustrated in Fig. 3.3.
Tr,gg.
poin t
r Input Sign
at
Tnsger
I Hysteresis Dana
1 (trigger unndoui)
te vet
t f - \ / IS et^uQl to noise
i ' V \ 1 mmun i1
Recoverg
point
Ou tpu t
from
tr,gg^r
C! r.cui t
Fig. 3.3. Visualization of the trigger function.
The hysteresis band (trigger window) is centered
around the trigger level and the width of the hyst
eresis band at the input, is the same as the effective
input sensitivity of Vpp.
The ideal hysteresis band is 50-60% of the signal's
peak-to-peak value. Since input signals can have
any amplitude within the specification, a continuously
variable input attenuator is prefered.
Signals, that are superimposed on a DC voltage, need
to be separated via a capacitor, i.e. AG-coupling. The
advantages of AG-coupling are:
• No DG drift.
• Good protection against DG overload.
However, AG-coupling gives a drop in sensitivity for
very low frequencies.
Time Interval measurements
For time interval measurements, too wide a hys
teresis band, i.e. too low a sensitivity, means that
different signal slopes at the start and stop trigger
point, cause different delays between the trigger level
crossing and the trigger point, see Fig. 3.5.
Frequency measurements
Timer/counters are used for both frequency and time
interval measurements. However, frequency and time
interval measurements have contradictory requi
rements in respect of correct triggering.
For frequency measurements, too narrow a hyste-
resisband, i.e. too high a sensitivity, means that the
counter is too sensitive to noise; see Fig.3.4. The
hysteresis band is equal to noise immunity.
Pulse duration
I Measured ft
r
\pulse duration I I
, Measured time (
r n
Fig. 3.5. Too wide a hysteresis might cause incorrect
time interval measurements.
WRONG
Three counts per cycle
CORRECT
One count per cycle
Fig. 3.4. Do not use a higher sensitivity than needed
for correct triggering.
The highest possible sensitivity with low noise is
ideal. However, a calibrated input attenuator is needed
to enlarge the trigger level setting range. A separate
xlO step attenuator, which expands the trigger level
range to -25V... + 25V is available on PM 6670...72.
A continuously variable setting of the trigger level
is necessary for setting the trigger level at any
required point of the input signal.
If the duty factor of the input signal changes, the
average DG-component will also change. In case of
AG-coupling, the trigger level follows the average
DG-component. This is not acceptable if the time
interval measurements with accurate trigger level
settings are to be made. Hence, DG-coupling is
necessary. Two identical inputs and slope selection
are also necessary.

14
The low-pass filter
The built-in 50kHz low-pass filter is used for improved
triggering on noisy LF-signals. The filter characteristic
is shown in Fig. 3.6. It is also possible to use this
filter for signals with frequencies above 50kHz, but
at a reduced sensitivity.
SO -
A t ienuQ i ion
in d B
AO -
30-
20-
JO-
ErcCfucncy
in MHz
1 1 1
0.02 0.05 0.1 0.2
1 1
0.5
\
2
Fig. 3.6. The low—pass filter reduces noise and inter
ference.
Input C
PM 6672 has a special RF input, called input C.
Note that not more than 1 2Vrivis's allowed at input C
and that the input sensitivity is adjusted automatically.
MEASURING MODES
Frequency A, Period A and RPM
PM 6670...72 perform a frequency and period
measurement as given in the definitions;
Frequency
Period =
Number of cycles
Time
Time
Number of cycles
The counter:
1 .Measures the effective measuring time.
2.Counts the number of input cycles during the
measuring time.
3.Computes the number of cycles per second
(frequency) or time units per cycle (period).
The measurement is synchronized with the input
signal. This is called the input synchronized or
reciprocal method.
In the input synchronized mode, both the opening
and closure of the main gate is synchronized with
the input signal, so that only completed input cycles
are counted. This means that a±1 input cycle error
is avoided. During the gate time, the counter also
totalizes the number of 100ns crystal clock cycles;
see Fig. 3.7.
fOOns
X- TAL
CLOCK
TRiGGCR
OUTPUT
"MEASURE'
S yNCHRON/EC
GATE
SIGNAL
CO UN TED
INPU T
CYCLES
TIME COUNTS
[Nurnber of
/OOns
clockpu (ses )
COUNTED INPUT CYCLES
rRcguENcv
TIM£ COUNTS ' 10-''s
■ COUNTCD INPUT CYCLES
Fig. 3.7. input synchronized mode.
The resolution in the input synchronized mode, is
caused by truncation of the clock pulses, which
results in a ± 1 clock pulse error (100ns). The
resolution of the measurement thus only depends
on the measuring time, e.g. the resolution for Is
measuring time is 10"' (lOOns/ls) independent of
frequency.
In conventional counters, the gate time is synchron
ized with the clock signal. The first and last input
cycle, can therefore be truncated, causing a ±1 cycle
error. This results in a good resolution for high
frequency measurements, but a poor resolution for
low frequency measurements (±1/frequency for Is
measuring time).
A RPM measurement is in fact a normal frequency mea
surement, but the microcomputer multiplies the fre
quency for 60 before displaying the result.
Measuring time
The measuring time can be varied in 33 steps per
decade between 10ms and 96s. Set measuring time
can be displayed by pushing the measuring time
rotary knob. This resets the counter and a new
measurement will start. The counter continues to
totalize input cycles until the set measuring time
has elapsed and the synchronization conditions are
met. Hence, the effective measuring time (also called
gate time) is longer than the set measuring time.

15
The number of cycles (N) is:
effective measuring time
N =
period duration
rounded to the nearest higher multiple of 10 for
frequency measurements or higher integer for period
measurements.
When the measuring time has elapsed, the micro
computer calculates the measuring result with a
10-digit resolution. However, the number of digits
displayed, is limited only to the significant digits,
depending on the measuring resolution. This
measuring resolution is defined by the input frequency
and the measuring time.
The number of digits is selected in such a way
that the measuring resolution is equal to 0.2...2 units
of the least-significant digit (LSD), where:
LSD = 2.5 X Frequency
Measuring timex 10^Hz
rounded to the nearest decade.
2.5 X Period
Measuring time x 10'Hz
Ratio Measurements
The counter measures the frequency ratio between
signals connected to input A and input B. The
frequency range is 0...10MHz.
To obtain full input frequency range on PM 6671 ...72,
a ratio measurement can be done in the FREQ A or C
mode. Connect the signal with the highest frequency
to input A (0.1H2...120MHz) and the other signal
to input D set to EXT STD IN. The frequency range
at input D is 50kHz... 10MHz. However, this arrange
ment does not give correct setting of the decimal
point. The display will indicate a frequency. To get
the correct ratio, divide the displayed value by 10' Hz,
e.g. when the display shows 215.513MHz, it
represents a ratio of 21.5513.
A ratio measurement is useful, for instance, when
calibrating oscillator with an awkward frequency. For
example, say that the fequency should be
4.3625872MHz. This is difficult to read on the
display. By connecting such a reference signal to
input B and measuring the ratio instead, the oscillator
is correctly calibrated when the display shows
1.0000000, which is much easier to read.
Count measurements
There are three different count modes:
Manual
The counter totalizes events
at input A, during the time
interval between releasing
and depressing the DISPL
HOLD pushbutton. An event
is defined as a positive-
going slope.
Gated by B
_l 1_
Start and stop by B
The counter totalizes events
at input A, between the
leading and trailing edge of
the input B signal.
The counter totalizes events
at input A, between the start
and stop event at input B.
Time Interval single measurements
In the time interval single mode, the time (i.e. number
of 100ns clock pulses) is measured between a start
event at channel A and a stop event at channel B.
The start and stop triggering can be individually set
with respect to coupling, trigger level, slope and
attenuation (x1 orxlO).
To perform single source measurements, such as
rise-time and pulse width, only channel A has to
be connected. The input B connector is disconnected.
However, channel B is internally connected by
means of the COM via A pushbutton. In this case,
the coupling and attenuation of channel B are dis
connected (identical as for channel A). The trigger
level and slope in channel B can still be set independ
ently of the channel A setting. The resolution of the
measurement is ± 1 clock pulse ( ± 100ns).
— Trigger level A
Channel A.
Trigger level B 4- C honnel B
f
Gale signal
I I
. JUllllUUUl fUUllUULJ''"''''
The measured time interval -
'number of clock pulses totalized
Fig. 3.8. Time interval single mode.
Time interval average measurements
By using the time interval average technique, which
means multiple measurements of a repetitive signal,
the measuring accurancy and resolution are greatly
improved. Compared to single time interval measure
ments, the basic 100ns resolution is improved by a
factor of I/iTN, where N is the number of time
intervals being averaged.
N =
Measuring time
Pulse repetition time
When using time interval average, the number of
leading edges of the clock pulses occurring in each
individual "time window" are totalized. Figure 3.9.
illustrates a rise-time measurement.

16
rmjL
/OOns
clock cycles ! cou
Measuring
Trigger let/ei 3
Trigger level A
Mam Gate
Fig. 3.9. Time interval average mode.
For a signal as illustrated in Fig. 3.9, approx 10000
time intervals are being averaged during a measuring
time of 10ms. Let's say that 6000 are measured
as 200ns (2 clock cycles) and 4000 as 100ns. The
statistical average is calculated in the microcomputer.
In this case the average is 1 60ns. The resolution
is 100ns/ VI0000 = Ins. Note that the input signal
must be repetitive and asynchronous with respect
to the time base and that the minimum dead time
from stop to start is 300ns.
Pulse width measurements
These measurementsare similarto singletime interval
measurements. Both start and stop triggering takes
place in channel A, with common trigger level setting
and automatic trailing edge trigger slope inversion
with respect to selected leading edge trigger slope.
Phase delay measurements
The timer/counters PM 6670...72 can measure the
phase delay between two signals connected to input A
and B. The measurement is performed as a simultan
eous measuring of time interval A-B and period.The
phase delay is calculated as:
Phase delay =
Time Interval A-B
Period
x360°
The measurement is made as an average measure
ment to improve accuracy and resolution; see
Fig. 3.10.
/7\\ /
\\ A A \\
\y/1 w/i
' 1 •'?/ 1 1
^ 1 1
n n
_j
—j
PHASC DELAY = " •
n^T
In order to get a high accuracy phase delay time
interval measurement, the setting of the trigger level
is very important. The trigger levels should be
identical for both channels and as close to zero
as possible. This is normally achieved with AC
coupling and max sensitivity setting. Unequal
settings of the trigger levels will result in inaccurate
time interval measurements; see Fig. 3.11.
Trigger level B
— L Trigger level A
Fig. 3.11. Unequal settings of the trigger levels will
result In inaccurate time interval measurements.
Very large differences in slew rate between the
signals, can result in a systematic phase error,
which can be up to 3...5.° This is caused by the
hysteresis band (typically ± 10mV). Although the
trigger level is set to OmV, the actual trigger point
will be + 10mV. With variations in slew rate, the time
before crossing the -FlOmV limit will vary; see
Fig. 3.12.
-WmV
Fig. 3.10. The phase delay measurement.
Fig. 3.12. Varitation in slew rate can result in phase
error.
It is therefore important to keep the signals at
about equal amplitude ( sine and triangular waves).
Since a normal time interval average measurement
is made, we also have the normal restriction
concerning minimum dead time between stop and
start of the time interval, i.e. 300ns. the dead
time also determines the maximum signal frequency,
which is 1.6MFIz.
Duty cycle measurements
A duty cycle measurement can be made by means of
a phase measurement. Connect the signal to Input A,
push COM via A and SLOPE for Input B. However,
the displayed read-out will be in degrees. Divide the
result by 3.6 to get 0...100% or by 360°to get 0...1
(duty factor). If the counter is equipped with a
PM 9696 IEC-625/IEEE-488 bus interface, this
calculation could easily be done by the controller.

17
SPECIAL FUNCTIONS
Arming
Arming enables the counter to be prevented from
starting on unwanted signals. The external ARMING
input (input E on the rear panel) allows an additional
trigger condition. When input E is set to high logic
level ( > 2V), the counter is prevented from starting
a new measurement. However, the counter makes all
preparations for a measurement. When input E returns
to low (< 0.5V), the measurement will start with a
minimum of delay. The delay is approx O.Sps.
Note that arming cannot be used in the COUNT A
manual mode.
-m
/mp u { E —n '
^ le s
IMlt-
Burst to be
^ 10
« - »
iiil/il
lliilii
Fig. 3.13. Burst requirements for muitipie burst
frequency average.
A monostable flip-flop or a pulse generator with
external triggering (e.g. PM 5716) could be used,
see Fig. 3.14. It is also possible to measure a single
burst by means of the input E control.
External reset
Electrical reset, provides an equivalent function to
the front panel reset pushbutton. The counter is
reset when input E is set to high logic level (>2V).
A new measurement can be made when input E
has returned to low ( < 0.5V).
Single burst frequency measurements
The input synchronized counter is in general suitable
for burst frequency measurements. The frequency
measurement does not start until the burst has
arrived, because the opening of the main gate is
controlled by the input signal. However, there are
some restrictions:
• The set measuring time must be shorter than
the burst duration.
• The burst must contain at least 20 cycles.
• The minimum measuring time is 10ms.
Multiple burst frequency average
PM 6671 and PM 6672 are equipped with an exter
nal gate function, permitting the counter to make
burst measurements down to 500ns. The external
gate control signal should be connected to input E
(Set to FREQ AVERAGE) for controlling the multiple
burst frequency average function. The measure
ment is interrupted when input E is higher
than 2V. The external gate time can be down to 500ns.
The actual measuring time is the sum of all
individual gate openings made during the set
measuring time.
Note that the burst must contain at least 10 cycles
during the time input E is low and 10 cycles after
input E has returned high, as shown in Fig. 3.13.
Fig. 3.14. Example of a measurement set-up for
multiple burst frequency average.
The difference in propagation delay in the two inter
nal channels in the counter-on-a-chip is approx 1 5ns.
In FREQ AVERAGE mode with very short external
gate times, this delay will cause a measureable error
as the 15ns will be repeated for each external gate
pulse. However, it is possible to compensate for
this error. Measure a stable, continuous wave of
approx the same frequency in the normal mode
without external gate signal with an external gate
signal having the same number of samples and the
same gate duration as will finally be used (measured
value = F2). To compensate for the error obtained
in the frequency average mode, multiply the reading
with the factor K = F1/F2.
The total relative error for a multiple frequency
average measurement is:
Rel error = -(- ^ 100ns ± trigger errorgg ± trigger error
" 'EG~ tEGVN
± rel. time-base error
where tgQ = external gate duration
N = number of burst samples.

18
Hold-Off
PM 6671 is equipped with trigger hold-off, which
avoids false stop triggering on spurious or unwanted
signals. A typical example is the pulses from relay
contact bounce, as illustrated in Fig. 3.15.
Spurious signal posses
through trigger window
Trigger hold-off time
Correct measurement
Afeosurement obtained
without trigger hold-off
Practical
Measurements
Preliminary settings
1.Before connecting the counter to the line, check
that the line voltage selector is correctly set.
2.Set the slide switches on the rear panel to INT SID
and EXT RESET (or ARMING).
3.Release all pushbuttons, as necessary.
4.Press POWER ON.
5.Set HOLD OFF to 0 (PM 6671).
6.Set MEASURING TIME to approx 0.1 s.
7.Set the SENSITIVITY controls fully clockwise
and pulled (AC mode).
Fig. 3.15. Hoid-off avoids false triggering.
Gate monitor
PM 6671 and PM 6672 are equipped with a gate
monitor output. By connecting an oscilloscope to
this output, it is possible to look at the gate open
signal. Figure 3.16 is an illustration of the gate
monitor function. Note that the gate open signal is
longer than the set measuring time due to the
synchronization time.
GATC MONITOR
Set meos. Iirr\e
25 ms
Input A signol
Penod A selected
Disployed value 9rns
GAT € MON! TOR
with Hold- Off
set to JOrnS
Measured Signal
Displayed value Jdrns
n
I I
I I
I 1
r T
I i
I t
Frequency measurements
1 .Make the preliminary settings.
2.Select FREQ.
3.Connect the signal to be measured to Input A
if the frequency is between 10Hz...120MHz.
4.Decrease the sensitivity until no triggering at all
occurs. Push ATT xlO, as necessary.
5.Increase the sensitivity again, until the GATE LED
is flashing and a stable reading is obtained.
6.For improved triggering on noisy LF-signals, use
the low-pass filter.
7.Set the measuring time to give optimum resolution
and measurement speed.
8.PM 6672 only:
If the frequency of the signal to be measured is
higher than 1 20MHz, input C must be used. The
frequency range for input C is 70MHz...1 GHz.
Press pushbutton C (1GHz) to select input C.
The sensitivity is automatically adjusted for input C,
thus facilitating perfect triggering under most
conditions. Maximum allowed voltage at input C
is 1 2Vrivis.
Fig. 3.16. An illustration of the gate monitor function,
with and without hoid-off.
Period measurements
1 .Make the preliminary settings.
2.Select PERIOD A.
3.Connect the signal to be measured to Input A.
4.Decrease the sensitivity until no triggering at all
occurs. Push ATT x 10, as necessary.
5.Increase the sensitivity again, until the GATE LED
flashes and a stable reading is obtained.
6.For improved triggering on noisy LF-signals, use
the low-pass filter.
7.Set the measuring time to give optimum reso
lution and measurement speed.

19
Count measurements
There are three different count modes:
Manual
Gated by B
_J L_
Start - stop by B
The counter totalizes events
during the time interval
between realising and
depressing the DISPL HOLD
pushbutton. An event is
defined as a positive-going
slope.
The counter totalizes events
at input A, between the
leading and trailing edge
(SLOPE released) of the
input B signal.
The counter totalizes events
at input A, between the
start and stop event at
input B.
1 .Make the preliminary settings.
2.Select the appropriate COUNT mode.
3.Connect the signal to be counted to input A.
4.Depending on the shape of the input signal, use
either AC mode or push the trigger level control
for input A (DC mode) and set the trigger level
approx to half the amplitude of the signal to be
counted.
5.Connect the control signal (if used) to input B.
6.Set the trigger level/sensitivity control for input B
to suit the control signal (if used).
7.At 10® counted events, the display is full. When
further events are totalized, truncation of the LSD
starts and the unit indicator now indicates
exponents. If the ps/kHz indicator glows read 10®
pulses, for ms/MHz read 10® pulses and for
s/GHz read 10® pulses.
Hold-Off (PM 6671 only)
To avoid false stop triggering on spurious or unwanted
signals, the hold-off function could be used. To display
the set hold-off time, select TIME INT SINGLE A-B
and push CHECK.
Ratio measurements
1 .Make the preliminary settings.
2.Select RATIO A/B.
3.Connect the signals to input A and input B. The
frequency range is 0...1 OMHz.
4.Decrease the sensitivity until no triggering at all
occurs.
5.Increase the sensitivity again, until a stable
reading is obtained.
6.Set the measuring time to give optimum resolution
and measurement speed.
To obtain full frequency range on PM 6671...72,
a ratio measurement can be made in the FREQ A or
C mode:
1 .Make the preliminary settings.
2.Select FREQ A or C.
3.Connect the signal with the highest frequency to
input A or input C (PM6672).
4.Connect the other signal to input D and set the
switch to EXT STD IN. Note that the frequency
range at input D is 50kHz...10MHz.
5.Set the sensitivity for channel A.
6.Set the measuring time to give optimum reso
lution and measurement speed.
7.Divide the displayed value by 10^ (10MHz) to
obtain the correct ratio.
Time Interval measurements
There are two modes for measuring time intervals.
Single: The counter measures the time interval
between a start event at input A and a
stop event at input B.
Average: The counter measures the time interval
average of a repetitive signal that is not
synchronized with the used time-base.
I.Make the preliminary settings.
2.Select TIME INT single A-B or average A-B.
3.Connect the signals to input A and input B.
4.When measuring time intervals on a single channel,
connect the signal to input A and push COM via A.
5.Push the trigger level control (DC mode) and select
correct trigger level, e.g. 50% of the signal
amplitude for measuring pulse width or 10% and
90% for measuring rise-time.
6.Select correct trigger slopes.
7.Time interval average only:
Set the measuring time to give optimum resolution
and measurement speed.
Phase delay measurements
1 .Make the preliminary settings.
2.Select PHASE A-B.
3.Connect the signals to input A and input B.
4.The input voltages must exceed lOOmVRMS
5.If any of the input voltages exceeds . Push
ATT x10 for the appropriate channel.
6.Do not change the sensitivity setting.
7.Try to obtain approx the same signal amplitude
to both inputs. Use ATT xlO or an external
attenuator (preferable resistive) if necessary.
8.Do not use oscilloscope probes as they introduce
extra phase delays.
9.Invert the trigger slope for channel B when measur
ing phase angles around 0° , thus 180° will be
added. Unstable readings caused by phase-jitter are
now avoided.
lO.Set the measuring time to give optimum reso
lution and measurement speed.

20
Pulse width measurements
1 .Make the preliminary settings.
2.Select PULSE WIDTH A.
3.Connect the signal to input A.
4.Push the trigger level control (DC mode) and set
the trigger level to 50% of the signal amplitude.
5.Select SLOPE, i.e. released for positive pulse
width and pushed for negative pulse width.
6.Pulse width measurements can also be performed
in the time interval mode, but require more control
settings. However, pulse width below Ips should
be made in the time interval average A-B mode.
Note;
• RPM will give
9.9.9.9.9.9.9.9.
• CHECK cannot be
overflow condition, i.e.
used for checking phase
delay, as the internal clock frequency is higher
than 1.6MHz.
Checking of COUNT A gated by B or start/stop
by B is not possible.
PULSE WIDTH will display 0, TIME INTERVAL
Single 100ns and TIME INTERVAL Average 50ns.
FREQ C (PM 6672) will display 256, i.e. the
prescaling factor.
Check mode
1 .Make the preliminary settings.
2.Press CHECK. The internal 10MHz time-base signal
is now connected to the logic circuits and all
input controls are disconnected.
CHECK enables a self-test of most measuring
functions. The resolution is given by the set measuring
time (not applicable for pulse width and time interval
single mode).
Overflow condition
An attempt to divide by zero (in ratio A/B mode)
or effective measuring times longer than 99s will
result in an overflow condition.
The display will show: 9.9.9.9.9.9.9.9.

21
4. Performance Check
Required Test Equipment
• Voltmeter, e.g. Philips PM 2517
• Frequency counter, e.g. Philips PM 6650
• Oscilloscope, e.g. Philips PM 321 5
• Sampling oscilloscope, e.g. Philips PM 3400
• Pulse generator, e.g. Philips PM 5771
• Function generator, e.g. Philips PM 5131
• HF signal generator, e.g. Wavetek 2002A
• Probe, lOMohm, 120MHz
• FET probe, e.g. PM 9354
• T-piece, BNC-type
• Temination, 50ohm, BNC-type
Initial set-up
All pushbuttons should be in a realeased position.
Set MEASURING TIME to 10ms (fully anticlockwise)
and HOLD OFF (PM 6671) to 0 (off). Pull both
TRIGGER LEVEL/SENSITIVITY potentiometers (i.e. AC
mode) and turn them fully clockwise. Select FREDA.
The slide switches on the rear panel should be set
to INT STD and EXT RESET (not applicable to
PM6670). Set the line voltage slide switch on the rear
panel to the local line voltage. Connect the counter
to the line and press POWER ON.
Check of the oscillator
The /01 version of PM 6670...72 can be checked
by connecting the 10 MHz OUT from a counter, e.g. a
PM 6650 or a PM 6673 equipped with at least a
calibrated TCXO, to connector D on the rear panel via
a 10 Mohm probe. The frequency should be 10 MHz
± 10Hz.For/02.../05 versions, please referto Section,
Optional Oscillators in the service manual.
Check of the sensitivity and
frequency response
• Connect a signal generator and an oscilloscope via
a T-piece to input A.
J O C)
I g i O g ; o ig j
• Check the sensitivity at the following frequencies:
50Hz...75MHz : 25mVpp
75...120MHz : 50mVpp
Check of Input C (PM 6672 only)
• Connect a RF signal generator and a sampling
oscilloscope via a T-piece to input C. Use a FET
probe on the oscilloscope.
• Set the amplitude to 1 5mVRMS ( — 24 dB) and
vary the frequency between 70 MHz and 800 MHz.
The counter should measure correctly.
• Set the generator to 1GHz with 25mVRivis
(-1 9dB) amplitude. Check that the counter displays
a correct value.
• Push CHECK.
• Adjust the measuring time and check the display
read-out as follows.
Measuring time
1 0ms
100ms
Is
1.5s
' Release CHECK.
Display read-out
0.25600GHz
256.000MHz
256.0000MHz
256.00000MHz
Check of TRIGGER LEVEL
(PM 6671 and PM 6672)
• Push both TRIGGER LEVEL potentiometers.
• Connect a DMM to the trigger level output
connector for input A on the front panel.
• Vary the trigger level between both end positions.
• Check that the voltage can be adjusted between
-2.5 and -F 2.5V with a tolerance of 0.1 5V.
• Repeat the procedure for input B.
• Pull both TRIGGER LEVEL potentiometers and set
them fully clockwise.
• Use short 50ohm coaxial cables. If the signal
generator has a 50ohm output impedance, a
50ohm termination has to be used at the oscillo
scope. The termination should not be used with
a sampling oscilloscope (above approx 25MHz).
Check of COUNT A MANUAL
•Push CHECK.
• Select COUNT A MANUAL.
• Push DISPL HOLD and then RESET. Check that
the display read-out is 0 and the GATE LED is off.
• Release DISPL HOLD. The counter should start
counting and the GATE LED start blinking.

22
Push DISPL HOLD. The counting should stop.
Release DISPL HOLD. The counting should start from
the accumulated value.
Count 10® pulses (takes approx IDs) and press
DISPL HOLD, the decimal point to the right of the
6:th decade should light. The ms/MHz LED should
also light, indicating that the result should be multi
plied by 10®
Push RESET and release DISPL HOLD.
Check of COUNT A Gated by B
Note: The instructions for checking Count, Puise Width
and Time Interval are written in such a manner that
the checks should be made in sequence.
• Select COUNT A Gated by B -J 1—
• Connect 10MHz OUT from e.g. a PM6650 to input
A and TIME-BASE OUT to input B. The PM 6650
should be set to PERIOD A with a 1 us time base.
• Release CHECK.
• The display read-out should be 6.
• Push SLOPE for input B. The display read-out should
be 4.
• Release SLOPE for input B.
Check of COUNT A Start/Stop by B
• Select COUNT A. Start/Stop by B -Tl—H—
• The display read-out should be 10. Push SLOPE
for input A if necessary due to the internal phase
condition in the LSI counter-ona-chip.
• Release SLOPE for input A.
Check of PULSE WIDTH A
• Select PULSE WIDTH A.
• Remove the cable from input A. Shift the cable from
input B to input A.
• The display read-out should be 0.0006 ± 0.0001 ms.
• Push SLOPE for input A. The display read-out should
be 0.0004 ±0.0001 ms.
• Release SLOPE for input A.
Check of TIME INT SINGLE A-B
• Select TIME INT SINGLE A-B.
• Push COM via A and SLOPE for input B.
• The display read-out should be 0.0006 :
• Push SLOPE for input A and release
input B.
• The display read-out should be 0.0004 ± 0.0001 ms.
0.0001 ms.
SLOPE for
Check of RPM
• Select RPM A.
• Connect a 50Hz, 500mVpp square-wave to input A.
• The display read-out should be approx 3000.
Check of PERIOD A, FREQ A
and RATIO A/B
• Select required function and push CHECK.
• Adjust the measuring time and check the display
read-out as follows.
MEAS.
TIME
1 0ms
100ms
Is
PERIOD A
100.000ns
100.0000ns
100.00000ns
FREQ A
10.0000MHz
10.00000MHz
10000.000kHz
RATIO A/B
1 .00000
1.000000
1.0000000
• Release CHECK.
Check of PHASE
• Select PHASE A-B. Push COM via A and SLOPE
for input B.
• Connect a 1.6MHz, 10OmVpp sine-wave to input A.
• The display read-out should be 1 8Cf± 3° .
• Push SLOPE for input A and release SLOPE for
input B.
• The display read-out should be 1 80°± 3° .
• Release SLOPE for input A.
Check of HOLD OFF
• Select TIME INT SINGLE A-B and push SLOPE for
input B.
• Connect a 50Hz, lOOmVpp square-wave to input A.
• The display read-out should be 1 0ms.
• Turn the HOLD OFF potentiometer slowly clockwise,
until the display read-out is 30ms.
• Push CHECK. The Hold-Off time should be approx
1 0ms.
• Reset HOLD OFF to 0. Release COM via A, SLOPE
for input B and CHECK.
Check of the low-pass filter
• Push the 50kHz FILTER button.
• Connect a 1MHz sine-wave to input A. Check that
the counter requires a signal amplitude ± 40dBm
(20x) higher than without filter.
Check of TIME INT AVERAGE A-B
• Select TIME INT AVERAGE A-B.
•Set the measuring time to Is. The display read-out
should be approx 400ns.
•Release SLOPE for input A and push SLOPE for
input B. The display read-out should be approx
600ns
• Remove the cable from input A.
Check of the dead-time between
stop and start
• Select PULSE WIDTH A.
• Push COM via A and SLOPE for input B.
• Connect a pulse generator to input A and adjust
the output to 1 MHz with 0.8V amplitude.