Ripple And Noise, Test Setup - HP BENCH Series Operating And Service Manual

POWER SUPPLY CASE
OSCILLOSCOPE CASE
GND-ii>- 1 D
I -I +c
[f
-Overt icAL
c:
-AC
-ACC
.|-GND
I
L
"G I
A/W
A. INCORRECTMETHOO- ground CURRENTIs PRODUCES 60
CYCLE DROP IN NEGATIVE LEAD WHICH ADDS TO THE POWER
SUPPLY RIPPLE DISPLAYED ON SCOPE.
POWER SUPPLY CASE OSCILLOSCOPE CASE
GND*^^-| A
TWISTED PAIR
VERTICAL
NPUT
USE 3-T0-2
adapter to
BREAK GND PATH
B. A CORRECT method USING A SINGLE-ENDED SCOPE.
3-T0-2 ADAPTER BREAKS GROUND CURRENT LOOP, TWISTED
PAIR REDUCES STRAY PICKUP ON SCOPE LEADS.
POWER SUPPLY CASE OSCILLOSCOPE CASE
AC
ACC
GND-
+0
SHIELDED
TWO -WIRE
&
-O VERTICAL
-O INPUT
c:
•AC
•ACC
hGND
C. A CORRECT METHOD USING A DIFFERENTIAL SCOPE WITH
FLOATING INPUT. GROUND CURRENT PATH IS BROKEN; COMMON
MODE REJECTION OF DIFFERENTIAL INPUT SCOPE IGNORES
DIFFERENCE IN GROUND POTENTIAL OF POWER SUPPLY S
SCOPE, SHIELDED TWO WIRE FURTHER REDUCES STRAY
PICK-UP ON SCOPE LEAD.
ed up from the grounds, the (-h) scope lead should
be shorted to the (-) scope lead at the power sup
ply terminals. The ripple value obtained when the
leads are shorted should be subtracted from the
actual ripple measurement.
5-30 In most cases, the single-ended scope
method of Figure 5-6B will be adequate to eliminate
non-real components of ripple and noise so that a
satisfactory measurement may be obtained. How
ever, in more stubborn cases, or in measurement
situations where it is essential that both the power
supply case and the oscilloscope case be connect
ed to ground (e. g. if both are rack-mounted), it
may be necessary to use a differential scope with
floating input as shown in Figure 5-6C. If desired,
two single conductor shielded cables may be sub
stituted in place of the shielded two-wire cable
with equal success. Because of its common mode
rejection, a differential oscilloscope displays
only the difference in signal between its two
vertical input terminals, thus ignoring the effects
of any common mode signal introduced because of
the difference in the AC potential between the
power supply case and scope case. Before using
a differential input scope in this manner, however,
it is imperative that the common mode rejection
capability of the scope be verified by shorting
together its two input leads at the power supply
and observing the trace on the CRT. If this trace
is a straight line, the scope is properly ignoring
any common mode signal present. If this trace is
not a straight line, then the scope is not reject
ing the ground signal and must be realigned in ac
cordance with the manufacturer's instructions until
proper common mode rejection is attained.
Figure 5-6. Ripple and Noise, Test Setup
however, that the power supply case is still con
nected to ground via the power supply output ter
minals, the leads connecting these terminals to
the scope terminals, the scope case and the third
wire of the power supply cord.
5-31 To check the ripple and noise output, pro
ceed as follows:
a. Connect the oscilloscope or RMS volt
meter as shown in Figures 5-6B or 5-6C.
b. Adjust VOLTAGE control until front panel
meter indicates maximum rated output voltage.
c. The observed ripple and noise should be
less than 200p.Vrms and ImV p-p.
5-28 Either a twisted pair or (preferably) a shield
ed two-wire cable should be used to connect the
output terminals of the power supply to the verti
cal input terminals of the scope. When using a
twisted pair, care must be taken that one of the
two wires is connected both to the grounded ter
minal of the power supply and the grounded input
terminal of the oscilloscope. When using shield
ed two-wire, it is essential for the shield to be
connected to ground at one end only so that no
ground current will flow through this shield, thus
inducing a noise signal in the shielded leads.
5-29 To verify that the oscilloscope is not dis
playing ripple that is induced in the leads or pick-
5-32 Noise Spike Measurement. When a high fre
quency spike measurement is being made, an in
strument of sufficient bandwidth must be used; an
oscilloscope with a bandwidth of 20 MHz or more
is adequate. Measuring noise with an instrument
that has insufficient bandwidth may conceal high
frequency spikes detrimental to the load.
5-33 The test setups illustrated in Figures 5-6A
and 5-6B are generally not acceptable for measur
ing spikes; a differential oscilloscope is neces
sary. Furthermore, the measurement concept of
Figure 5-6C must be modified if accurate spike
measurement is to be achieved:
1. As shown in Figure 5-7, two coax
5-5