STANDARD AUDIO TESTS

BY
CLIFFORD SCHROCK

ACKNOWLEDGEMENTS

The author would like to thank Linley Gumm and Gordon

Long for their exceilent technical assistance in the prepara-
tion of this paper. In addition, I would like to thank Joyce
Lekas for her editorial assistance and Jeanne Gaiiek for the
illustrations and layout.

CONTENTS
PRELIMINARY INFORMATION
Test Setyps__ ; , page 2
Input- Outpyt Load hatching ; page 3
TESTS
Power Oytpyt ; , page 4
Frequency Response : page 5
Harmonic Distortion , page 7
IntermO'dulation Distortion ; page 9
Distortion vs Outpyt ; page 11
Power Bandwidth ; page 11
Damping Factor ; page 12
Signal to Noise Ratio page 12
Square Wave Response page 15
Crosstalk , page 16
Sensitivity page 16
Transient intermsdulation Distortion page 17
SERVICING H1MTS page 19
 PRELIMINARY INFORMATION
Maintaining a modern High-Fidelity-Stereo system today
requires much more than a "trained ear." The high
specifications of receivers and amplifiers can only be
maintained by performing some of the standard measure-
ments such as:

1. Power output
2. Harmonic distortion
3. Intermodulation distortion
4. Frequency response
5. Signai-to-Noise measurements

Unfortunately, because of the available test equipment and

lengthy procedures that are required to "spec" a new or
repaired amplifier, it usually doesn't get done. Figure 1 5L4N Low Frequency Analyzer and Mainframe with 5A15

This pamphlet describes an alternative test device and

procedures that permit rapid, easy to understand,
"spec'ing" and troubleshooting. The device, a Low Fre-
quency Spectrum Analyzer, is now available, at a moderate
price, to the Audio industry. In addition, besides the
standard measurements, the spectrum anaiyzer may be used
effectively for expanding the standard tests or for special
measurements such as the CCIF distortion or Bell Tele-
phone multitone audio tests.

These descriptions are presented so that they can be

foMowed by anyone with a technical background. Proce-
dures apply to all the standard tests typically performed on
high fidelity equipment.

Test Setups Figure 2 SG502, TM503, DM502.DC-504 Audio Test Package

The only major test equipment that is really required to

"spec" an amplifier with our new techniques are; the
Tektronix 5L4N Low Frequency Spectrum Analyzer piug-
-in unit with a 5100-Series mainframe, and an audio
oscillator, such as the Tektronix SG502 (Figures 1 and 2).

Some complementary test equipment that may be useful

would be a vertical amplifier plug-in, such as the 5A15, to
be inserted beside the 5L4N so the mainframe can be used
as a standard time domain oscilloscope: a digital multi-
meter, such as the DM502, that reads out in dB's, for
troubleshooting; and a frequency counter, such as the
DC-504. An attenuator with a Q to 60 or 80 dB range can be
a handy test device for audio measurements. Two kinds to Figure 3 Construction Details for a Step Attenuator
consider are the variable (potentiometer) type or a step
attenuator. You may want to build attenuators. Construc-
tion details are given in Figure 3 and 4 for both the step
and variable types.

To ensure accurate measurements, certain precautions must

be observed. AC power or high level RF fields from local
radio or TV transmitters can interfere with the low level
measurements encountered in Hi-Fi systems. The typical
test setup, Figure 5, should be well shielded. Use a copper
sheet for the test bench top and ground ali equipment
through short pieces of copper braid to the copper sheet.
The use of an AC line filter on the bench plugs is also
recommended. Try reversing the AC pfugs on equipment to
obtain minimum hum. Figure 4 A Variable Attenuator
Figure 5 Typical Test Setup with Good Shielding

Load Matching
As a final note, wherever possible perform measurements The degree to which amplifiers and preamplifier inputs and
with input signals about 30 dB above the measured outputs are matched or loaded wiil affect the accuracy of
reference sensitivity, using the volume control on the the overall test.
preamplifier to obtain reference power output.
Power amplifiers must be loaded to their characteristic
impedance, within plus or minus 1% (usually 8 ohms). The
resistor bad should have no more than 2% reactive
components. This restriction often precludes the use of
CAUTION some types of wirewound resistors. An accurate non-re-
Some consumer high fidelity equipment can constitute a active, high power load can be constructed for the power
shock hazard Transformerless audio equipment can level (wattage) and impedance required by connecting a
have line AC on the chassis, the control shafts, the input, number of one or two-watt carbon resistors in parallel. A
and speaker leads. Transformer equipment can also have photo of a homemade load constructed with 1 watt carbon
leakage to ground or defective bypass capacitors. The resistors immersed in oil to dissipate extra power Is shown
following precautions are recommended when testing in Figure 6.
equipment.

Before applying AC power to the bench make sure that

1 All equipment is securely grounded to the bench top
through the ground braid leads
2 The bench ground is secure Use No 12 or larger wire
fo r bench ground lead Some water pipes ARE NOT
ground Test should be performed to insure a good
ground
3 The bencn power should be well fused A 5 amo
piugstnp fuse snould be used The use of a Ground Fault
interrupter (GFI) :s also recommended on the plugsmp
circuit such as the 3M Mode! 2701

4 When testing transformerless equipment, an isoiatior

transform >r should be usea on the AC supply circuit
After appl 'ing AC power, exercise caution white ad-
justing amplif-.er and test equipment controls since
hazg'dous potentials will always be present wher
makino tests
Figure 8 Load Matching Chart
 The amplifier output must be matched to the 600 ohm
spectrum analyzer input. A chart of values for 600 ohm
amplifier matching pads as well as for other common values
TESTS
is also shown in Figure 6.
Power Output
Preamplifiers should see a 100 kilohm slightly capacitive
Power Output of an amplifier can be roughly defined as the
load such as the device in Figure 7 would provide.
maximum power an amplifier can deliver per channel
before distortion begins to impair the audio (music, etc.).
Sometimes a test probe is the best way to test or Unfortunately, there is wide disparity in how much
troubleshoot a circuit. To maintain frequency response and distortion different people will accept. Therefore, the most
minimal loading, use a 10X probe connected directly meaningful power output figures are obtained when the
through the input of the analyzer. Select H! Z input. To output is plotted against the amount of harmonic and
measure higher voltages, construct a pad as illustrated in intermodutation distortion.
Figure 8.
Power measurements can be performed in many ways. A
The input signal supplied by a signal generator should be
common procedure is to measure the continuous RMS
applied in series with an impedance equal to the source
impedance for which the equipment was designed. With output power of an amplifier for at least 30 seconds (to
auxiliary inputs, it is considered standard to apply the sig- account for changes due to heat, power supply sag, etc.).
nal in series with a resistance of 5000 ohrns plus or minus Power measurement of stereo amplifiers is then made for
5%. Undriven channels should have their inputs terminated each channel with both channels driven.
as described above.
The Federal Trade Commission (FTC) in November of
1974 enacted a regulation on measuring amplifier power
output specifications.

Under this regulation, the amplifiers must be precondi-

tioned by simultaneously operating all channels at one third
of rated power for one hour using e 1000 Hz tone.
Complete descriptions of these new rules are available
through the FTC.

Other power tests exist, such as the IHF2and Music Power

tests. These claim to give a better picture of an amplifier's
Figure 7 Preamp Load Pad ability to respond to sudden or instantaneous demands;
however, the specifics will not be covered in this procedure.

Power Output Test Procedures

1. Select a load-'matching resistor combination for the

amplifier being tested from the chart in Figure 6. If the
maximum power output is unknown, select a load-matching
combination high enough to handle any anticipated power.

2. Connect the equipment as shown in Figure 9. Set a!!

amplifier controls to their flat positions and set the volume
control to maximum.

3. Select 1 kHz on the SG502 oscillator and carefully

increase the output level until the tone is visible on the
analyzer screen in the 10 dB/DI V mode.

Figure 8 Using a probe with the analyzer 4. Watch the 2nd (2 kHz) and 3rd (3 kHz! harmonics of the
1 kHz tone. Increase the output ievei until the amplitude of
the 2nd or 3rd harmonic increases faster than the 1 kHz
tone. The ieve! at which the harmonics (distortion) begin to
increase radically (faster than the tone reference) is
generally considered to be the MAXIMUM UNDISTORTED
POWER OUTPUT (Figure 105.

1. "F.T.C, Standard," as published in Stereo Review, Page 81,

November 1874.
2. Institute of High Fidelity Standard Methods of Measurements for
Figure 8 Equipment Setup for Measuring Power Audio Amplifiers, SHF-A-201 (1966).
Figure 10 Distortion Crossover Point for Measuring Power Figure 12 Visible Clipping of sine wave

5. For the new FTC test,'all channels of a multi-channel Harmonic Distortion are contained in this application note
(stereo! amplifier should be driven to the maximum power and should be followed to satisfy the entire FTC regulation.
point before power is measured.

6. Power is determined by assuming the full screen display

to be the value of the load-matching combination. Switch Frequency Response
to the 2 dB/DIV mode and note the number of dB down
from the top graticule line, as shown in Figure 11. This is
the power output in dB below the full screen display and Frequency Response is a measure of the amplifier's ability
can be converted directly to watts. to pass a wide range of frequencies in the audio spectrum.
Icteally, one would strive to achieve a fiat response; that is,
all frequencies would pass through an amplifier with equai
amplification. A Hi-Fi amplifier may have controls to
modify the response. These may include tone controls (bass
ant treble), rumble and hum filters (low frequency rolloff),
scratch filters (high frequency rolloff), and a variety of
tailoring devices such as the RIAA, FM de-emphasis, and
tape head equalization filters. The frequency response test
should provide response information of the amplifier in the
flat position and should also represent the limits and
interaction of the tone controls and filters.

Response of a modern Hi-Fi system is generally measured

from below 20 Hz to we!! beyond the 15 kHz audible limit.
It is measured in dB of deviation across the audio spectrum.

The 5L4N Low Frequency Analyzer is ideally suited to

frequency response testing since it has a self contained
tracking generator and a log sweep 20 Hz to 20 kHz mode.
Figure 11 Calculating Power Output An amplifier can be swept under a variety of different
conditions in a matter of seconds, eliminating the need for
tedious measurements and point to point plots. Multiple
Notes:
traces of conditions can be built up either on film or on a
storage oscilloscope to obtain one picture of the complete
1. An alternative procedure for demonstrating maximum
response performance of an audio device.
undistorted power is to simultaneously apply the 1 kHz
tone through a plug-in vertical amplifier in the compart-
ment next to the 5L4N and display the time domain sine The rated frequency response is the frequency range over
wave on the scope. Increase the tone level untii the sine which the amplitude response does not vary more than plus
wave visibly clips, as shown in Figure 12. The power level in or minus 3dB from the amplitude at 1000 Hz.
dBV or dBM is then read on the spectrum analyzer display.
3, E.A Standard Methods of Measurement for Audio Amplifiers
2. To satisfy the FTC requirements, the rated power must Used in Home Equipment RS-234-C (1971).
be obtainable at all frequencies within the rated power 4. Larry Klein, "Amplifier Power-Output Ratings: A New FTC
band (width) without exceeding the total rated maximum Trade Regulation Rule," Stereo Review, Voi. 35 No. 5, Page 79,
harmonic distortion. Procedures on Power Bandwidth and November 1874.
Figure 13 Frequency Response Equipment Setup

Figure 15 Tone Control Ranges

Fjprs 14 Frequency Response at 10%, 50% and 75% Power

Frequency Test Response Procedures

1. Select a load resistor matching pad combination for the

rated power of the amplifier being tested. Connect the
equipment as illustrated in Figure 13.

2. All tone controls and filters on the amplifier should be Figure 16 Loudness Control Effects
initially set for flat response. The volume (and/or ioudness)
controls should be set to maximum.

3. Use the LOG span on the 5L4N, An internal circuit

jumper in the 5L4N provides either 20 Hz to 20 kHz, or a
100 Hz to 100 kHz LOG span. Select the 20 Hz to 20 kHz
LOG span. Consult the instrument manual for details. Set
the SPAN/DIV to LOG and the display mode to 10dB/D!V
(20 Hz corresponds to the left edge of the graticule). Install
the Audio Graticule contained in this application note.

4, Begin the test with all the attenuation IN on the

attenuator.

5. With the analyzer sweeping, reduce the attenuation of

the attenuator until the rated power of the amplifier is
achieved (as indicated on the display, taking the load-
matching combination into account). The response, from
20 Hz to 20 kHz, should now be displayed.

6, Different power levels should also be checked such as

10% (-10 dB}, 50% (-3 dB), and 75% f-1.25 dB) to see if
variations occur in the response (Figure 14). Figure 17 Response Action of Doibv "B" Encoder
7. The 2 dB/DIV display mode provides higher resolution 3, Harmonic distortion appears as multiple signals above the
of the response flatness. fundamental 1000 Hz tone. Switch to 10 dB/DIV mode
and observe the position and amplitude of the 2nd
8. The range and effect of the tone control and filters can harmonic (2000 Hz) relative to the 1000 Hz reference tone
be checked by varying these controls. If the 5L4N (Figure 19). The amplitude ratio between the two, in dB, is
mainframe has storage or a camera is available, sequential the second order harmonic distortion. This ratio can be
displays of responses can be "built u p " as shown in Figure converted to percentage of distortion by referring to the
15. chart in Figure 20.

iSiotes: 4. Similarly, the amplitude ratio between the reference and

higher order components may be calculated and converted
1. The effects produced by varying the loudness contour to percentage of distortion.
control can be observed {Figure 16) by sweeping the flat
response each tirnt the loudness control is increased by 1/8
turn.

2. The action of a Dolby B system encoder or decoder can

be checked or noted by sweeping the device each time as
the input level is reduced by 10 dB steps (Figure 17),

3. EIA standards3 recommend that the frequency response

be measured at a power output not higher than 1OdB below Figure 18 Harmonic distortion method of measurement
the rated power output and not lower than 20dB above
residual hum and noise.

Harmonic Distortion
Harmonic Distortion or THD (total harmonic distortion) is
determined by measuring and summing the ampiitude level
of the various h air ionics that occur when a single, pure
tone is passed through an amplifier. This is the most
common distortion test performed on amplifiers. Harmon-
ics can be predicted to occur in sequence (2nd, 3rd, 4th,
etc.). Therefore, a 1000 Hz tone would have a 2nd
harmonic of 2000 Hz, a 3rd harmonic of 3000 Hz, etc. The
harmonics represent various amounts of distortion as shown
In Figure 18?

The low frequency analyzer permits evaluation of the

components contribu ting to THD and also makes it possible Figure 18 Harmonic Distortion with Strong Second Order
to visually and graphssally reference output leve! (power) to Components
distortion. The 5L4N in combination with the SG502 audio
oscillator can make 70 dB THD (.034%) measurements. %of % Of
RATIO in dB RATIO in dB READING
READING
With auxiliary fixed filters, this range can be extended to
measure 100 dB ( .001%) THD, Complete details are 20. (40:60! 10% (1% .1%| 30 150,705 3.16% (.31,.031%)
contained in the note s that follow.
21 8.S 31 2.87

22 7.94 32 2.51
Harmonic distortion is generally plotted against frequency
at different power levels across the audio spectrum. The 23 7.08 33 2.24
following procedure describes one method of measuring 24 6.31 34 2.00
harmonic distortion.
26 S.62 35 1.78

Procedures 26 6.01 36 1.59

2? 4.47 37 1.41
1. This tesi is usually conducted when the power output of
28 3.98 38 1.26
an amplifier is measured. The equipment setup is shown in
Figure 9. Set the amplifier controls for a flat response and 28 3.SB 3S 1.12

maximum
Figure 20 Chart for Conversion from dB's to Percentage Readings
2. Select a 1000 hz tone from the audio generator
(SG502). increase the output level unti! the reference 5. James S. Aagaard, "An Improved Method for the Measurement of
power rating of the amplifier is reached, as indicated by the Nonlinear Audio Distortion," IRE Transactions on Audio, Page 121,
2 dB/DiV display on the spectrum analyzer screen. 1958.
 8

5. THD (Total Harmonic Distortion) is calculated by 2. The 70 dB (.031%) range of measurement that is possible
comparing the reference tone level to the RMS sum of a!! with the 5L4N Low Frequency Spectrum Analyzer can be
harmonic ieveis. The sum of the harmonics can be extended to 100 dB or more (.001%) by adding a band stop
determined by using the chart in Figure, 21.If a!! other filter tuned to suppress the fundamental test tones between
harmonics are 8 dB or more down from the 2nd harmonic, the amplifier and 5L4N INPUT. The oscillator must also be
one can disregard the higher order harmonics and use only clean to 100 dB and may require extra bandpass filtering.
the second to obtain an accurate THD figure (see Figure Complete details for extended range measurements are
22). shown in Figure 24.

6. The harmonic distortion can be plotted against 3. In some situations, it may be difficult to distinguish
between the harmonics and the noise floor. The noise floor
frequency at various power levels such as 10%, 50% and
may be moved down by selecting a slower sweep speed
100% on a suitable chart. This will produce results similar
which will, in turn, automatically seleci a narrower RESO-
to those illustrated by Figure 23.
LUTION B.W.
4. EIA standards discuss a parameter called Low Power
Distortion. It is the highest vaiue of toai harmonic distor-
Mote: tion(THD) measured on a 1kHz tone from 6dB to 26dB
below the rated power output.
1. THD figures closer to actual operation can be obtained
by setting the volume control at less than maximum, then 6. W.J. Warren and W.R. Hewlett, "An Analysis of the Inter-
increasing the output level of the audio generator to bring modulation Method of Distortion Measurement," Proc. IRE, Vol.
the power output of the amplifier up the rated reference. 36, Page 457, February 1948.

Figure 22 Calculating Percentage of Total Harmonic Distortion

23 rmonic Distortion Plotted vs Frequency at Dif

• 21 Correction Factors for Addition of Compc
 9

Figure 24 Extending the Range of THD Measurements at 1 kHz

Intermodulation Distortion
InTermodul. tion Distortion is determined by putting two or
more pure tones into an amplifier and measuring the
amount one tone is transferred (cross modulated) onto the
other.

Two commoniy used intermodulation tests are described by

the chart in Figure 25.

Measurement of Intermodulation Distortion'has always

generated a lot of controversy. AH involved agree that the
various measurements produce numoers that relate to
performance of audio equipment; however, everyone has
his/her own idea about what frequencies to use, how many,
Figure 25 1M Methods of Measurement
the ieveis, e'-c.

Fortunately, the low frequency analyzer can handle all of

the different known methods. The procedure presented
below uses the Society of Motion Picture and Television
Engineers (SMPTE) modulation method using two tones.
The specific tones and ratios are recommended by the
institute of High Fidelity Manufacturers (IHFM).

Intermodulation Distortion Test Procedures

1. Trsis test is usually conducted in conjunction with the

power output measurement of an amplifier. Two tones
must be used and carefully combined for intermodulation
Distortion measurement. The SMPTE method uses the
filament traisformer arrangement as shown in Figure 26 for
the 60 Hz source, instead of a second audio generator,

7. Gerald Stanley and David Mclaughlin, "Intermodulstion Distor-

Figure 28 Equipment Setup for tntermoduiation Measurement tion: . . . Audio, Page 36, February 1972.
 10

2. For the SMPTE Modulation method, select 60 Hz and

6000 Hz (7000 Hz for IHFM) in a voltage ratio of 4 : 1 . This
means that the 6000 Hz tone will be down 12 dB from the
60 Hz tone. This two tone generator setup can be checked
directly with alow frequency analyzer (Figure 27).

3. Set the amplifier controls for a flat response, with the

volume control set to maximum.

4. Drive the amplifier using the two tone generator to

indicate 14 dB less than the rated power using the 6000 Hz
tone as a reference point The sum of the two tones will be
equivalent to the rated power since the 60 Hz tone is 12 dB
higher and the component of the 6000 Hz tone adds
approximately 2 dB.
Figure 27 SMPTE Modu _... . . . . . . .

5. Select 10 dB/DI V display mode and a SPAM/DIV of 100

Hz. Tune the 6000 Hz signal to center screen with the
FREQUENCY control.

6. Figure 28 is a typical display. The sidebands that appear

around the base of the 6000 Hz tone are modulation
components of 60 and 120 Hz, generated fay the 60 Hz
tone. Use the chart in Figure 21 to calculate the sum of the
60 Hz, 120 Hz, and other visible modulation components
that appear on one side of the carrier only. Do not add the
upper and lower 60 Hz or 120 Hz components together.
Percentage of intermodulation distortion can be calculated
from the,chart in Figure 20.

Notes:

1. The, CCIF Difference method of measuring requires a

similar setup that uses two equal amplitude input signals
Figure 28 SMPTE Modulation Method of Measurement
from the signal generators. A typical display is shown in
Figure 29.

2. Other intermodulation measurement methods, such as

the Modified CCIF (symmetrical distortion), SMPE Three-
-Component, Noise and Notch, Impulse, and Multitone, can
a!! be handled similarly.

3. The amount of IM distortion may decrease or change

when the volume control setting is other than maximum,
with the power output being maintained by increasing the
signal generator output level.

4. A parameter called Low Power Intermodulation Distor-

tion is discussed in the EiA standards. This is the highest
IM distortion measured from power outputs from 6dB to
26dB below the rated power output. Figure 28 CCIF Difference Method of Measurement
 11

Distortion vs Output Power Bandwidth

Distortion vs Output is a- measure of the distortion for Power Bandwidth, for high fidelity test purposes? is defined
every power level of an audio amplifier system. By plotting as the frequency range between the two points where the
the output power or voltage against the percentages of both distortion at a power output 3 dB below the reference
harmonic (THD) and intermodulation ftft/S) distortion, one power intersects the reference distortion line at full power.
can readily determine the power capability for any The reference distortion at full power is usually measured
distortion level. using a 1 kHz tone. Then the power output is held to 3 dB
less than full power while the harmonic distortion THD is
plotted for tones from 10 Hz to the upper frequency limits
A Federal Trade Commission rule requires complete of the amplifier.
disclosure of harmonic distortion (THD) using a 1000 Hz
tone for all power output levels from 250 mW to rated This test is performed using the same procedure as for THD
power. This procedure has been followed for the THD except that the distortion curve is plotted at 3 dB less than
measurement. IM specifications are not required to satisfy rated power, and the frequency range of the measurement
the FTC but are induced in the following procedure. is much wider.

Power Bandwidth Test Procedures

Distortion/Output Test Procedures 1. Use the equipment setup shown in Figure 9. Use the
procedures discussed under Harmonic Distortion for
reference.
!. Use the equipment setup and the procedure for harmonic
distortion measurement; and make a series of readings by 2. Measure the distortion of a 1 kHz tone at the rated
using a 1000 Hz tone with a power output of 250 raW* power output of the amplifer. This is the REFERENCE
increasing the output until the rated power of the amplifier DISTORTION and should be plotted on the chart as shown
is exceeded. Piot this data on a graph similar to that shown in Figure 31.
in Figure 30.
3. Reduce the power output by 3 dB {to 50%) and carefully
begin plotting the distortion for frequencies of 20 Hz, 40
2. Use the equipment setup and the procedure for the Hz, 60 Hz, 100 Hz, etc.
intermodulation distortion measurements and make a series
of readings that start with a power output of 250 rnW 4. It is important to reset the 50% power reference as
increasing the output until the rated power of the amplifier indicated on the low frequency analyzer, since the fre-
is exceeded. Plot this data on the graph that was used to quency response of the amplifier will vary.
plot harmonic distortion.
5. Continue to measure at regular frequency intervals until
the percentage of distortion becomes at least twice as high
as the REFERENCE DISTORTION.

6. The frequency between the two crossover points is the

Power Bandwidth of the amplifier.

Figure 30 IM and THD Distortion of an Amplifier Figure 31 Power Bandwidth of an Amplifier

8. Edward C. Palmer, "8 Ways to Test Hi-Fi Amplifiers," Radio-

Recommended reference tone by FTC, November 1974. Electronics, Voi. 44 No. 7, Page 37, July 1973.
 12

Damping Factor
Damping Factor is measure of output impedance versus
frequency relative to a constant load R. it is an indirect
measure of an amplifier's ability to remain stable while
encountering speaker impedance changes (at different
frequencies).

A simple way to measure damping factor is to measure the

ratio at maximum power of the ioaded output voltage to
the unloaded output voltage. Care must be exercised during
this measurement because some amplifiers cannot be driven
to maximum power with a no load condition for more than
a few seconds without damage.

Procedures
Figure 33 Damping factor In
1. Use the equipment setup that is illustrate..) in Figure 32.
Use a specially constructed or modified load-matching
combination as shown. The load portion of the pad must be
switch able.

2. Use a 2 dB/DIV display mode and apply a 1 300 Hz tone.

Increase the signal level to its rated output powar.

For the EIA standard measurement, use a 100 Hz tone and

use an output 6 dB less than the rated power QU: put.

3. Switch to LIN mode and accurately note the voltage

amplitude of the signal.

4. Disconnect the amplifier load with the switch and Figure 34 A Plot of Damping Factor vs Frequency
remeasure the signal voltage as shown in Figure 33.
Duration of the no load condition shouSd be kept short to
protect the amplifier.

5. To calculate the damping factor: Damping Factor equals Signal to Noise Ratio
EL over ENL minus EL; where EL is voltage with load and
ENL is voltage with no load.
Signal to Noise Ratio is a measurement of the ratio of the
6. If desired, the damping factor can be piottsd for all rated output of an amplifier to noise (mostly thermal
frequencies from 20 Hz to 20 kHz as shown in Figure 34. noise). Ideally, one would like to be able to differentiate
between noise and other problems such as hum.

Signal to noise ratio measurements are performed with a

variety of techniques. Some of these techniques include the
use of weighting filters that take into account the response
sensitivity of the human ear {the Fietcher-Munsen effect),
and other techniques use lew pass filters to measure noise
flatly across the audible range.

The Spectrum Analyzer can measure Signal to Noise Ratio

easily if measurements do not specify the use of a weighting
filter. In addition the analyzer permits rapid display of a
noise parameter not normally attempted with traditional
audio measuring sets. Spectral-Noise DensitytOcan fae easily
determined and displayed. This is a plot of noise against
frequency. This permits much more accurate evaluation of
Fipre 32 Damping Factor Measurements systems that alter or improve noise performance such as the
 13

Dolby systems, tape systems, FIVf receiving systems, and

some modern preamplifiers.

Signal to Hum Ratio is measured while checking Spectral-

Noise Density by comparing the residua! 60 and 120 Hz
components to the rated output of an amplifier. Hum is
generally output signal components from the poweriine
frequency. In some modern amplifiers a high frequency
power supply of 1 kHz or even supersonic (above 20 kHz)
frequencies may be used. The residual signals from these
frequency supplies must also be measured.

To producs a standard specification, it may be necessary to

Figure 35 Equipment Connections to Measure S/N measure tie weighted signal to noise ratio (WSNR)'.' The
mathemat caliy inclined can calculate this from the
Spectrai-hosse Density readings or one car. use weighting
filters w i t i the tangential method of measuring noise.

This technique is described briefly at the end of the Signal

to Noise procedure.

Procedures

1. Set up tiie equipment as shown in Figure 35. Apply a

1000 Hz tone from the audio signal generator to the
amplifier input. Adjust the signal level so the amplifier is
driven at rat >ti power output (volume control maximum! as
indicated b'• the 2 dB/D!V display of the spectrum
analyzer.

2. Set the RESOLUTION bandwidth to 3 kHz and select a

span of 2 kHz'DIV.

Figure 3S Measuring the Noise Floor 3. Switch to the 10 dB/DIV display mode and note the
amplitude of the 1000 Hz tone. This will be the Signal
Reference Level (VS).

4. Switch the 1000 Hz tone off and observe the noise

on the display. Use the analyzer VIDEO FILTER to average
the noise so ihe amplitude of the noise floor {VN) can be
measured as shown in Figure 36.

5. The difference in amplitude (Figure 37} between the

signal reference level and the noise floor, less a correction
from the measured bandwidth (3 kHz) to the desired
bandwidth (15 kHz), is the signal to noise ratio. The
correction factor to be subtracted is 7 dB, Example: S/N
equals VS minus VN minus Correction Factor; S/N equals
-2 dBm minus (-7 dBm) minus 7 dB; equals plus 65 dBm.

6. if the noise floor does not appear flat, use a narrower

RESOLUTION bandwidth to determine if the noise is, in
fact, uneven, or if another problem such as hum is causing
an uneven dispiav. The shape of the noise fioor from 20 Hz
Figure 3? Example of S/N and Hum Measurements to 20 kHz is the Spectral-Noise Density, Figure 38.

9. "Standards on Electron Devices: Methods of Measuring Noise," 11. RETMA Sta.idard Amplifiers for'Sound Equipment, SE-101-A,
Proc. IRE, Vol. 41, Pages890-896, July 1953. July 1947, Reprinted by EIA, May 1973.
10. "Standards on Sound Recording and Reproducing: Methods and
Measurement of Noise," Proc. iRE, Vol. 41, Pages 508-512, April 12. Gary Franklin and Troy Hatley, "Don't Eyeball Noise,"
1953. Electronic Design 24, Pages 184-187, November 22, 1973.
 14

Notes: Various weighting filters may be specified for different

kinds of equipment. The schematic of one commoniy used
1. if the noise floor is below 80 dB, the gain of the analyzer type is shown in Figure 39.
can be increased in 10 dB steps until the noise floor is
visible. Every step adds 10 dB spread between the signal To measure noise as recommended by the EIA standard the
"and noise floor. B section of the weighting filter shown in Figure 39 should
be inserted in the test setup as shown in Figure 40.
2. Signal to hum ratio can be determined by measuring the Procedures
amplitude difference between the Signal Reference Level
(step 3) and any 60 or 120 Hz hum components. An 1. Set up the equipment as shown in Figure 40. Short the
example is shown in Figure 37. amplifier input. A dual trace vertical plug-in such as the
Tektronix 5A18N must be used, if this procedure is used
3. Signal to noise and signal to hum should be rechecked without a weighting filter, some form of a frequency
through different amplifier inputs and with reduced volume limiting filter (usually a 15 kHz low pass) MUST be used.
control settings to determine the exact characteristics of
these parameters. 2. Connect the noise to both vertical channels simul-
taneously in the alternate-sweep mode. Two noise traces
will appear (Figure 41}.

3. With both channels identically calibrated, adjust the

I. WEIGHTED SIGNAL TO NOISE voltage offset until the dark area between the two traces
just disappears (Figure 42).
Weighted Signal to Noise measurements can be made using
an oscilloscope display to measure the RMS amplitude of 4. Disconnect the inputs and measure the separation
the noise. Unfortunately, it is not easy to "eyeball" the between the two traces in volts. This is VN (Figure 43).
amount of noise on the screen; therefore, the tangential
technique was derived to permit rapid, repeatable 5. The ratio between the voltage at rated power output and
measurements of noise. the noise VN is the signal to noise ratio.

~i§«re 39 Weighting Network Figure 40 Equipment Setup for Weighted Signal to Noise Tests

Fifure 38 Spectral-Noise Density Display Figure 41 Dual Trace Display of Noise

 15

Fipire 42 Offset Adjusted until Dark Area just Disappears ; 43 Measuring Separation between Two Traces

Square Wave Response

Square Wave Response is measured by passing first a 50 Hz

then a 10 kHz square wave through an amplifier operating
at its rated output. The resultant waveform provides a
quick check of a number of parameters including frequency
response, transient response, group delay, and distortion.
This is strictly an osciiSoscope type measurement once the
rated power is determined.

Procedures

Figure 44 Equipment Setup for Square Wave Response

1. Set up the equipment as shown in Figure 44. The square
wave output of the Audio Generator (SG502) is applied to
the ampiifer. The output of the amplifier is connected
through the load-matching combination network and split
into both the low frequency spectrum analyzer and an
amplifier plug-in unit. Set the amplifier controls for a flat
response and the volume control to maximum.

2. The amplifier is driven at rated power output using a 50

Hz square wave. Rated power indication on the analyzer
will be 2 dB below the rated power for a single tone,
because the square wave spectrum contains less funda-
mental frequency power than a sine wave.

3. Observe the resultant square wave response of the

vertical amplifier plug-in.

4. Change the square wave generator frequency to 10 kHz

and observe the vertical amplifier response (Figure 45).

5. Some information to look for is: Rounded corners

indicating frequency response deficiencies; transients which
indicate damping or phase problems; and axis shift of the
50 Hz square which indicates poor low frequency coupling. Display of 50 Hz and 10 kHz Square Waves
 16

Crosstalk
Crosstalk is the amount of signal that leaks or spreads from
one channel of a system into another channel or channels
of the system. It is measured by driving al! channels except
one to rated output, then measuring the amount of signal
that leaks into the idle channel. Crosstalk is usually
expressed in dB's and is the ratio of the rated power output
to the signal level on the idle channel.

With the Low Frequency Spectrum Analyzer, we can easily

go one step further and measure the crosstalk at all
frequencies from 20 Hz to 20 kHz, obtaining a display of
crosstalk versus frequency in dB.

The EIA standard calls for separation (crosstalk) to be Figure 47 Crosstalk from One Channel to the Other
measured at 3 dB less than the Rated Power Output.

Procedures

1. Set up the equipment as shown in Figure 46. The

switches may be deleted by plugging the input signal into
Sensitivity
various jacks as desired. All channel outputs must be
properly terminated and unused inputs have a shielded Sensitivity checks determine how much signal {in volts)
matching resistor and not a shorting jumper connected must be applied to each input terminal to drive an amplifier
from input to ground. Set all controls for a flat response to its rated or some reference output. The overload point
and the volume control to maximum. for each input is usually determined while making the
sensitivity checks.
2. When testing four channel amplifiers, ail the channels,
except the one being measured, should be driven at rated Procedures
power output,
1. Connect the equipment as shown in Figure 48. The
3. Drive one channel to rated power output, using a 1000 switch is convenient though not absolutely necessary for
Hz tone from the tracking generator output. Then sweep this measurement.
the channel using the LOG SWEEP to obtain a reference
response.
2. Start with any amplifier input and'use a 1000 Hz tone to
4. Switch the analyzer input to the other channel and, using drive the amplifier to its rated power output. The volume
the LOG SWEEP MODE and log graticule, observe the control and any input gain controls MUST be set at
crosstalk signal level in dB with respect to the reference maximum.
response. The drop in dB from the rated power reference
established on the other channel can be measured by using 3. Once rated power output is reached., switch the signal to
the 10 dB/DiV display mode on the spectrum analyzer the analyzer input. Select the LIN display mode and read
(Figure 47). the input voltage amplitude required to achieve amplifier
rated power output. This voltage Is the SENSITIVITY
5. Repeat the check on each adjoining channel. rating of the amplifier input tested.

Figure 4S Setup for Measurement of Crosstalk Figure 48 Sensitivity Equipment Setup

 17

4. Using the THD procedures {part III), begin reducing the crepancies. Transistor amplifiers test excellent using steady
volume on the amplifier to indicate below rated output state harmonic and intermodulation tests. However, music
while increasing the level of the 1 kHz tone until the 2nd material generates amplifier distortion because of its
and 3rd harmonics increase more rapidly than the input transient nature.
signal. Remeasure the input voltage. This is the overload
point for that input. A popular explanation of the source of TIM is that the
transient reaches or exceeds the slew rate of the amplifier
5. Continue this procedure for all inputs on the amplifier causing an instant, severe intermodulation condition until
recording the results {Figure 49!. the negative feedback signal catches up with and corrects
the distortion.

No measurement standards exist to date!4However, a square

wave with a high frequency sine wave has been used to
observe this distortion.

Presented below is a technique that used a 8 kHz sine wave

mixed with a 500 Hz square wave to demonstrate TIM
distortion. However, no singie number results to adequately
indicate the amplifier's performance.

Procedures

1. Connect the equipment as shown in Figure 50. Two

SG502 Audio Generators must be combined as shown, one
set to produce 500 Hz square waves., the other 6 kHz sine
waves. A FG501 may be substituted for the square wave
source. The square wave used should have excellent
symmetry.

Figure 49 Recorded Results of Sensitivity and Overload Tests

. Transient
Intermodulation
Distortion Figure 50 Equipment setup for TiM demonstration

Transient Intermodulation Distortion (TIMPis distortion in 2. Temporarily bypass the amplifier and set up the
amplifiers that occurs principally during ioud, high generator levels in a voltage ratio of about 5:1. Then
frequency passages. Most music contains some material that carefully adjust the frequency of the 6 kHz generator until
can. cause TIM distortion. Amplifiers with large amounts of a stationary pattern is obtained similar to the top trace in
negative feedback are prone to TiM distortion because the Figure 51.
amplifier loop, if improperly designed, requires too much
time to respond tp rapid transients. 3. Reconnect the amplifier and remove attenuation until
50% of rated power is Indicated by the 500 Hz fundamental
Ever since the introduction of the transistor power on the analyzer display.
amplifier, the "transistor sound" has been discussed. Even
though in many cases a transistor amplifier tested better in
terms of distortion than a tube counterpart, during a 13. Matti Otala, "Transient Distortion in Transistorized Audio
listening test the tube unit would unmistakably perform Power Amplifiers," IEEE Transactions on Audio...., Vol. AU-18,
better. TIM distortion is one explanation of these dis- No. 3, Page 234, September 1970.
 18

4. One of the characteristics of TIM may appear on the

oscilloscope display as in the lower trace of Figure 51. This
display shows a temporary "swamping" of the 6 kHz tone
immediately after the 500 Hz transition.

5. Bypass the amplifier again and select 1 kHz/DIV in the

LINEAR sweep mode on the spectrum analyzer. Manually
seiect a resolution bandwidth so that good sideband
resolution is obtained as shown in Figure 52.

6. Observe the pattern obtained directly from the source.

Note that the even harmonics of the 500 Hz signal are
suppressed (Figure 52).

7. Reconnect the amplifier and note that TIM may be

obse-ved as 1 kHz sidebands of the 6 kHz signal as shown in
Figure 53.
Figure SI TIM Display in the Time Domain
14. W. Marshall Leach, "Transient IM Distortion in Power Ampli-
fiers,*'Pages 34-41, Audio, February 1975.

Figure S2 500 Hz 8 kHz source signal for TIM demonstration Figure 53 TIM Display in the Frequency Domain
 19

SERVICING HINTS
The combination Low Frequency Spectrum Analyzer- C. STAGE BY STAGE GAIN CHECK
-Oscilloscope offers tremendous flexibility to permit rapid
servicing of high quality Hi-Fi and stereo equipment. An Stage by Stage Gain Check are often used to find the source
oscilloscope probe can be connected directly to the front of a weak or distorted channel in an amplifier system. A
panel of the 5L4N making it easy to pinpoint problem modification of the stage by stage check can be performed
areas. with the Low Frequency Spectrum Analyzer using the
tracking generator output. Insert the tracking generator
While it is impossible to list ail the teps and techniques one into an ampiifier input, and using a X10 oscilloscope probe,
might use, the following are sonv; of the things we came begin at the input stages of the ampiifier and monitor the
across while preparing this applications brochure. input and output of each stage.

A. BIAS VOLTAGE ADJUSTMENT Channels can be compared to each other or gain and
response can be checked against the manufacturers' re-
The adjustment o* bias voltage on the output stage is commendations. Certain problems will be immediately
traditionally done with a voltmete- or a distortion analyzer. obvious if they exist.
The Low Frequency Analyzer can be connected to an
amplifier and using a single tone (like 1 kHz) the bias can 1. Gain differences will show up on the display-
be quickly set for minimum harmonic amplitude. Then the
output power can be reduced anc the low level crossover 2. Low frequency rolSoff will indicate such things as
distortion that sometimes occurs cai be double checked. defective stage coupling or output coupling capacitors.
3. High frequency rolloff could indicate defective
B. INAUDIBLE FREQUENCIES emitter bypass capacitors or other associated problems.

Sometimes a HI-FI system doesn't seem to perform correct- Tne test can then be repeated with no signal on the
ly and the problem can be traied to overloading or ampiifier input, watching the noise floor. A high noise floor
distortion due to INAUDIBLE frequencies being passed by rise between stages is characteristic of a defective or hot
the amplifier. component.

The more common causes of these frequencies are inaccur- Finally, using a 1 kHz tone, a stage by stage check can be
ate adjustment of the stereo traps {setting 19 and 38 kHz performed while watching the 2nd and 3rd harmonics for
into the audio channel): tape recorder bias traps improperly signs of a stage with higher than normal distortion.
adjusted; improper bypassing of inputs letting radio fre- Typically, the output stage should contribute most of the
quency energy into the amplifier; and sometimes an distortion in an amplifying system.
amplifier will just oscillate al! by itself.

By using the 10 kHz/DIV MODE of the analyzer routinely

D. STYLUS PRESSURE ADJUSTMENTS
when checking an amplifier, these kirds o f problems will be
Phono cartridge weight adjustments are often one of the
immediately visible. Figure 54.
least understood areas of Hi-Fi. The manufacture attempts
to recommend the lowest stvius pressure that produces
minimum distortion. By using a continuous tone test
record, and starting with a new stylus, the pressure is
increased progressively until the lowest distortion |2nd and
3rd harmonics; of the tone are noted. This test can be
easily duplicated in the shop using the low frequency
analyzer.

A little imagination, and a low frequency spectrum analyzer

can go a long way toward taking the mystery out of Hi-Fi
repairs. Other equipment such as tape recorders, speakers,
and electronic instruments (like organs) can be analyzed.
Complete PA systems used in auditoriums can be swept and
analyzed, or the acoustics and noise levels of an auditorium
could be checked.

This pamphlet only covers one small application area for

the low frequency analyzer. We hope that you will have an
opportunity to experiment with our applications and
expand them to fit your individual needs.
 STANDARD AUDIO TEST FORM

Response at 10% , 50%, and 100% Tone control response at S0%

power power
Spectral - noise density Square wave response
iO H i -10 kHz

Harmonic distorti&n (THD) Sensitivity & Overload

Power bandwidth Distortion arid power

TEKTRONIX. INC. ENCOURAGES YOU TO COPY OR PRINT THIS TWO PAGE FORM FOR YOUR OWN USE
CcpyngH "g 19"£ Tektronix nc Al, ngnts rese'i/ec P»nted m U S A Foreigr ana U.S A Procuots of Tektronix inc a'e
covered by Foreigr and U.SA Patents and/0' Paients Pending Inforrr.atior ir, this publication supersedes alp previously
published material Specification and once change p"viieges reserved TEKTRONIX, SCOPE-MOBILE, "Tti-EQUIPMEfM1" ana
^ J are registered trademarks of Tektronix, Inc , P C Box 500, Beaverton Oregon 97077 Phone (Area Code 503, 644-0*6'',
TWX. 910-467-8708 Cable. TEKTPON'X Ove'seas O-st"tutors ir over 40 Courfies