Tektronix Differential Oscilloscope Measurements Technical Note

Technical Note

Differential Oscilloscope
Measurements

A Primer on Differential Measurements,
Types of Amplifiers, Applications, and
Avoiding Common Errors

Simulated 4 mV
noise, using a conventional oscilloscope probe (upper). A differential amplifier extracts the signal from the noise

heartbeat waveform can not be measured in the presence of 500 mV

p-p

Introduction

All Measurements Are Two-Point

Voltage is always measured
between two points in a cir­cuit. This is true whether
using a voltmeter or an oscil­loscope. When an oscillo­scope probe touches a point
in a circuit, a waveform usu­ally appears on the display,
even if the ground lead is not
connected. In this situation,
the reference for the mea­surement is conducted
through the safety ground of
the scope chassis to the elec­trical ground in the circuit.

By virtue of their two probes,
digital voltmeters measure
potential between two
points. Because they are iso-

lated, these two points can
be anywhere in the circuit.
This has not always been the
case. Before the advent of the
digital voltmeter, hand-held
meters known as VOMs
(Volt-Ohm-Meters) were
used to measure “floating”
circuits. Because they were
passive, they tended to load
the circuit-under-test. Less
invasive measurements were
made with the high­impedance VTVM (Vacuum
Tube Volt Meter). The VTVM
had one major limitation –
the measurement was always
referenced to ground. The
VTVM housing was
grounded and connected to
the reference lead. With the

, 60 Hz common-mode

p-p

introduction of solid-state
gain circuits, high perfor­mance voltmeters could be
isolated from ground, allow­ing floating measurements to
be made.

Most oscilloscopes today,
like the venerable VTVM,
can only measure voltages
that are referenced to earth
ground, which is connected
to the scope chassis. These
are referred to as “single­ended” measurements – the
probe ground provides the
reference path. Unfortu­nately, there are times when
this limitation lowers the
integrity of the measurement,
or makes measurement
impossible.

If the voltage to be measured
is between two circuit nodes,
neither of which is
grounded, conventional
oscilloscope probing cannot
be used. A common example
is measuring the gate drive in
a switching power supply
(see Figure 1).

Signals which are balanced
(between two leads without a
ground return) such as a
common telephone line can­not be measured directly. As
we shall see, even some
“ground referenced” signals
cannot be faithfully mea­sured using single-ended
techniques.

When Ground Is Not Ground

We’ve all heard of “ground
loops” and been taught to
avoid them. But how do they
corrupt a scope measure­ment? A ground loop results
when two or more separate
ground paths are tied

Copyright © 1996 Tektronix, Inc. All rights reserved.

Contents

Introduction · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 1

Differential Measurement Fundamentals · · · 4

Overview of Differential Measurements · · · · · · · 4
Common-Mode Rejection Ratio (CMRR) · · · · · · · 4
Other Specification Parameters · · · · · · · · · · · · · · 6

Differential-mode range · · · · · · · · · · · · · · · · · · · 6
Common-mode range · · · · · · · · · · · · · · · · · · · 6
Maximum common-mode slew rate · · · · · · · · · 6

Types of Differential Amplifiers and Probes · · · · 7

Built-in differential amplifiers · · · · · · · · · · · · · · 7
High-voltage differential probes. · · · · · · · · · · · 7
High-gain differential amplifiers · · · · · · · · · · · · 8
High-performance differential amplifiers · · · · · 8
Differential passive probes · · · · · · · · · · · · · · · · 8
High-bandwidth active differential probes · · · · 8
Voltage Isolators · · · · · · · · · · · · · · · · · · · · · · · 8

Figure 1. Gate drive signal in a switching power supply is measured between TP1 and
TP2. Neither point is grounded.

Differential Measurement Applications · · · · · 9

Power Electronics · · · · · · · · · · · · · · · · · · · · · · · · · 9
System Power Distribution · · · · · · · · · · · · · · · · · · 9
Balanced Signals · · · · · · · · · · · · · · · · · · · · · · · · 10
Transducers · · · · · · · · · · · · · · · · · · · · · · · · · · · · 10
Biophysical Measurements · · · · · · · · · · · · · · · · 11

Maintaining Measurement Integrity · · · · · · · 11

Sources of Measurement Error · · · · · · · · · · · · · 11
Input Connections · · · · · · · · · · · · · · · · · · · · · · · · 11
Grounding · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 11
Input Impedance effects on CMRR · · · · · · · · · · · 13
Common-Mode Range · · · · · · · · · · · · · · · · · · · · 13
Measuring Totally Floating Signals · · · · · · · · · · 14
Bandwidth · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 14

Glossary · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 15

Related Publications from Tektronix · · · · · · · 16

Figure 2. Ground loop formed by a scope probe. Metal chassis of both scope and device
under test are connected to safety ground and internal power supply common. Scope
probe ground connects to scope chassis at the input BNC connector.

together at two or more
points. The result is a loop of
conductor. In the presence of
a varying magnetic field, this
loop becomes the secondary
of a transformer which is
essentially a shorted turn.
The magnetic field which
excites the transformer can
be created by any conductor
in the vicinity which is car­rying a non-DC current. AC
line voltage in primary
wiring or even the output
lead of a digital IC can pro­duce this excitation. The cur­rent circulating in the loop

impedance within the loop.
Thus, at any given instant in
time, various points within a
ground loop will not be at
the same potential.

Connecting the ground lead
of an oscilloscope probe to
the ground in the circuit­under-test results in a ground
loop if the circuit is
“grounded” to earth ground
(see Figure 2). A voltage
potential is developed in the
probe ground path resulting
from the circulating current
acting on the impedance
within the path.

develops a voltage across any

page 2

Thus, the “ground” potential
at the oscilloscope’s input
BNC connector is not the
same as the ground in the
circuit being measured (i.e.,
“ground is not ground”).
This potential difference can
range from microvolts to as
high as hundreds of milli­volts. Because the oscillo­scope references the mea­surement from the shell of
the input BNC connector, the
displayed waveform may not
represent the real signal at
the probe input. The error
becomes more pronounced

being measured decreases, as
is common in transducer and
biomedical measurements.

In these situations, it’s often
tempting to remove the probe
ground lead. This technique
is only effective when mea­suring very low-frequency
signals. At higher frequen­cies, the probe begins to add
“ring” to the signal caused by
the resonant circuit from the
tip capacitance and shield
inductance (see Figure 3).
(This is why you should
always use the shortest
ground lead possible.)

as the amplitude of the signal

Figure 3. Series resonant tank circuit formed by probe-tip capacitance and ground inductance.

Figure 4. “Floating” battery-powered cellular telephone probed with a grounded oscilloscope. Capacitance
between the phone circuitry and steel bench frame forms a virtual ground loop at high frequencies.

Figure 5. Minute parasitic inductance and resistance in ground distribution system result in VG≠ V'G.

We now have a dilemma:
create a ground loop and add
error to the measurement or
remove the probe ground
lead and add ring to the
waveform!

The next technique often
tried to break ground loops is
to “float” the scope or “float”
the circuit being measured.
“Floating” refers to breaking
the connection to earth
ground by opening the
safety-ground conductor –
either at the device-under­test or at the scope. Floating
either the scope or the
device-under-test (DUT)
allows the use of a short
ground lead to minimize ring
without creating a ground
loop.

This practice is inherently
dangerous, as it defeats the
protection from electrical
shock in the event of a short
in the primary wiring. (Some
special battery-operated
portable scopes incorporate
insulation which allows safe
floating operation.) Operator
safety can be restored by
placing a suitable ground­fault circuit interrupter
(GFCI) in the power cord of
the oscilloscope (or device­under-test) with the severed
ground. However, be aware
that without a low­impedance ground connec­tion, radiated and conducted
emissions from the scope
may now exceed government
standards – as well as inter­fere with the measurement
itself. At higher frequencies,
severing the ground may not
break the ground loop as the
“floating” circuit is actually
coupled to earth ground
through stray capacitance
(see Figure 4).

Even when the measurement
system doesn’t introduce
ground loops, the “ground is
not ground” syndrome may
exist within the device being
measured (see Figure 5).
Large static currents and
high-frequency currents act
on the resistive and induc-

page 3

tive components of the
device ground path to pro­duce voltage gradients. In
this situation, the “ground”
potential referenced at one
point in the circuit will be
different than that referenc­ing another point.

For example, ground at the
input of the high-gain ampli­fier in a system differs from
the “ground” potential at the
power supply by several
millivolts. To accurately
measure the input signal
seen by the amplifier, the
probe must reference the
ground at the amplifier input.

These effects have chal­lenged designers of sensitive
analog systems for years. The
same effect is seen in fast
digital systems. The small

inductance within the
ground distribution system
can create a potential across
it, resulting in “ground
bounce”. Troubleshooting
systems affected by ground­voltage gradients is difficult
because of the inability to
really look at the signal
“seen” at the individual com­ponent. Connecting the oscil­loscope probe ground lead to
the “ground” point of the
device results in the uncer­tainty of what effects the new
path adds to the ground gra­dient. A sure clue that a
change is occurring is seen
when the problem in the cir­cuit either gets better (or
worse) when the probe
ground is connected. What
we really need is a method to

make a scope measurement
of the actual signal at the
input of the suspect device.

By using an appropriate dif­ferential amplifier, probe, or
isolator, accurate two-point
oscilloscope measurements
can be made without intro­ducing ground loops or oth­erwise corrupting the mea­surement, upsetting the
device-under-test, or expos­ing the user to shock hazard.

There are several types of
differential amplifiers and
isolation systems available
for oscilloscopes, each opti­mized for a particular class
of measurements. In order to
choose the proper solution,
an understanding of termi­nology is necessary.

Overview of Differential
Measurements

An ideal differential ampli­fier amplifies the “difference”
signal between its two inputs
and totally rejects any voltage
which is common to both
inputs (see Figure 6). The
transfer equation is:

VO= AV( V

where VOis referenced to
earth ground.

The voltage of interest, or
difference signal, is referred
to as the differential voltage
or differential mode signal
and is expressed as V
is the V
transfer equation above).

Figure 6. Differential amplifier.

page 4

+in

Differential Measurement Fundamentals

The voltage which is com­mon to both inputs is
referred to as the Common­Mode Voltage expressed as

. The characteristic of a

V

CM

differential amplifier to

is referred to

CM

+ i n

– V

– V

)

– i n

term in the

–in

DM(VDM

ignore the V
as Common-Mode Rejection
or CMR. The ideal differen­tial amplifier rejects all of the
common-mode component,
regardless of its amplitude
and frequency.

In Figure 7, a differential
amplifier is used to measure
the gate drive of the upper
MOSFET in an inverter cir­cuit. As the MOSFET switches
on and off, the source voltage
swings from the positive sup-

ply rail to the negative rail. A
transformer allows the gate
signal to be referenced to the
source. The differential ampli­fier allows the scope to mea­sure the true V

signal (a few

G S

volt swing) at sufficient reso­lution such as 2 V / d i v i s i o n
while rejecting the several
hundred volt transition of the
source to ground.

Common-Mode Rejection Ratio
( C M R R )

Real implementations of dif­ferential amplifiers cannot
reject all of the common
mode signal. A small amount
of common mode appears as
an error signal in the output,
making it indistinguishable
from the desired differential
signal. The measure of a dif­ferential amplifier’s ability to
eliminate the undesirable
common-mode signal is
referred to as Common-Mode
Rejection Ratio or CMRR for
short. The true definition of
CMRR is “differential-mode
gain divided by common­mode gain referred to the
input”:

A

CMRR =

D M

A

C M

Figure 7. Differential amplifier used to measure gate to source voltage of upper transistor in an inverter bridge.
Note that the source potential changes 350 volts during the measurement.

Figure 8. Common-mode error from a differential amplifier with 10,000:1 CMRR.

For evaluation purposes, we
can assess CMRR perfor­mance with no input signal.
The CMRR then becomes the
apparent V
put resulting from common
mode input. It’s expressed
either as a ratio – 10,000:1 –
or in dB:

dB = 20log

A CMRR of 10,000:1 would
be equivalent to 80 dB.

seen at the out-

DM

V

D M

(

)

V

C M

For example, suppose we
need to measure the voltage
in the output damping resis­tor of an audio power ampli­fier as shown in Figure 8. At
full load, the voltage across
the damper (V
reach 35 mV, with an output
swing (VCM) of 80 V p-p. The
differential amplifier we use
has a CMRR specification of
10,000:1 at 1 kHz. With the
amplifier driven to full
power with a 1 kHz sine

) should

DM

wave, one ten thousandth of
the common-mode signal
will erroneously appear as

at the output of the dif-

V

DM

ferential amplifier, which
would be 80 V/10,000 or
8 mV. The 8 mV represents
up to a 22% error in the true
35 mV signal!

The CMRR specification is
an absolute value, and does
not specify polarity (or
degrees of phase shift) of the
error. Therefore, the user can
not simply subtract the error
from the displayed wave­form. CMRR generally is
highest (best) at DC and
degrades with increasing fre­quency of V
ential amplifiers plot the
CMRR specification as a
function of frequency.

Let’s look at the inverter cir­cuit again. The transistors
switch 350 V and we expect
about a 14 V swing on the
gate. The inverter operates at
30 kHz. In trying to assess
the CMRR error, we quickly
run into a problem. The com­mon-mode signal in the
inverter is a square wave,
and the CMRR specification
assumes a sinusoidal com­mon-mode component.
Because the square wave
contains energy at frequen­cies considerably higher than
30 kHz, the CMRR will prob­ably be worse than specified
at the 30 kHz point.

Whenever the common-mode
component is not sinusoidal,
an empirical test is the
quickest way to determine
the extent of the CMRR error
(see Figure 9). Temporarily
connect both input leads to
the MOSFET source. The
scope is now displaying only
the common-mode error. You
can now determine if the
magnitude of the error signal
is significant. Remember, the
phase difference between

and VDMis not specified.

V

CM

Therefore subtracting the dis­played common-mode error
from the differential mea-

. Some differ-

CM

page 5