Core
Technology Group Application Note 10
AN-10
Speaker Crossover Analysis
Paul A. Teseny
Introduction
A Vector Network Analyzer (VNA) is one of the most important aids for analog
designs. It can provide accurate information about active and passive network
performance in the frequency domain, such as frequency response, phase
response, impulse response and so on. Designers heavily rely on these
measurement results to locate design problems and to evaluate network
performance.
This application note discusses how to measure the impedance of loudspeakers
as well as how to measure frequency response of audio crossovers using the
SA/RA Series Frequency Response Analyzer. After a short introduction to the
measurements this document will show how to configure the Frequency
Response Analyzer and how to perform the measurements.
Notes:
1. All measurements in this application note have been performed with the
SA/RA Series Frequency Response Analyzer software Suite V5.50.
Speakers
Physics sets the limits for audio reproduction systems and compromises have to
be made. As no loudspeaker driver can reproduce all the audible frequencies
from 20 Hz to 20 kHz without amplitude reduction, splitting the spectrum for
multiple drivers is required. Loudspeaker crossovers are used to accomplish this,
but at the same time crossovers introduce several problems. Some have more
than others. The crossover is an electronic filter, which is typically passive and
mounted inside the speaker enclosure. In this Application Note, I have
concentrated on a two-way crossover network.
An ideal crossover would filter exactly the defined range of frequencies to the
woofer and respectively to the tweeter, so that no overlap occurs between the
drivers. It would, as well, keep the filtered output signals from the woofer and
tweeter intact, in terms of magnitude and phase. In reality, such filter is not
140 Independence Lane
Chalfont, PA 18914
www.CoreTechGroup.com -1–
Core
Technology Group Application Note 10
AN-10
possible due to physical limitations of components. The schematic shown in
Figure 1, is the crossover used in the speakers tested.
In a loudspeaker, the transfer function
of a crossover filter consists of the sum
of low-pass and high-pass filters’
transfer functions in a two-driver case.
It must be remembered that the total
reproduction consists of the transfer
functions of both the crossover filter
and the loudspeaker drivers, although
here the crossover filter is only
Figure 1
discussed.
Crossover Used in Testing
The filter’s transfer function is written by the following equation:
H(s) = H(s)L + H(s)H
Where H(s)L and H(s)H are the transfer functions of the low-pass and high-pass
sub-filters. In order to achieve a good response on the listening axis, the drivers
are time aligned for coplanar radiation, otherwise, there is a lobe error (tilting) in
the loudspeaker’s radiation pattern towards the lagging driver, which is outside
this discussion.
Figure 2 shows the magnitude response of a simulation of the crossover used in
these speakers. The magnitude response can be considered the most important
parameter in crossover design. Where high-pass and low-pass signals cross over,
they overlap and affect each other’s reproduction accuracy by interference.
How much, depends on the phase relation between the speakers.
When the woofer and tweeter are both contributing to the reproduction in the
crossover frequency region, being in the same phase means that they boost
each other, or cancel each other when in the opposite phase. In addition, their
mutual phase through the whole frequency range is of importance.
The final design of a good loudspeaker means the whole transfer function of the
system, not just the crossover. In practice, the properties and desired parameters
of a crossover depend on the driver’s physical properties, and that of the
enclosure.
140 Independence Lane
Chalfont, PA 18914
www.CoreTechGroup.com -2–
Core
Technology Group Application Note 10
AN-10
Figure 2
Simulation of Crossover in Figure 1
It is safe, for now, to assume the manufacturer of the loudspeakers used in this
testing have followed the design practices discussed above in developing this
crossover.
Speaker Details
The measurements discussed in this document were performed on a real full-
range passive bookshelf loudspeaker (the manufacturer shall remain nameless).
These loudspeakers are a high-end consumer product and are owned by the
author of this Application Note. The speakers consists of a tweeter and a mid-
range / woofer as well as a crossover. Both speakers are installed in the same
enclosure along with the crossover.
Measurements
The following measurements are discussed in this application note:
Impedance measurement of loudspeakers.
Impedance at the input to the crossover.
Frequency Response of each path of the speaker crossover.
140 Independence Lane
Chalfont, PA 18914
www.CoreTechGroup.com -3–
Core
Technology Group Application Note 10
AN-10
Impedance Measurement
Impedance Setup
One of the most difficult measurements to make in audio is impedance.
Whether measuring the impedance of an individual component or a complete
network, the Frequency Response Analyzer, makes this a simple task.
For the input impedance measurement we perform a logarithmic frequency
sweep from 10 Hz to 40 kHz and use an output level of 0 dBm such that the
sound of the loudspeakers during the measurement is not too loud and so that
we do not overload the analyzer inputs. The Input Over Voltage bargraphs in the
Sweep Parameters tab on the virtual front panel of the analyzer indicate an
input level that is too high by displaying a red bar. It is important that there are
no red bars, which indicates clipping, during a sweep. If clipping occurs, simply
decrease the level and re-run the test.
We set the number of
points to 200 in order to
speed-up the frequency
sweep process while still
delivering enough
resolution. Under the
Advanced tab, we set
the IF Bandwidth to 5 Hz,
which is below the lowest
frequency in the scan.
Changing the IF
Bandwidth helps to
reduce noise while
making the sweep take
longer due to the
number of calculations.
The image in Figure 5
shows the connection of
Figure 3 the Frequency Response Figure 4
Sweep Parameters Tab Analyzer (FRA) to the Advanced Tab
speaker terminals.
140 Independence Lane
Chalfont, PA 18914
www.CoreTechGroup.com -4–
Core
Technology Group Application Note 10
AN-10
Figure 5
Connection for measuring Impedance
Reference resistor, RREF, in Figure 5 is used to
sense the current through the network. The
value of the reference resistor is entered into
the Calibration tab on the virtual front panel
of the analyzer.
Input Impedance
Once everything is connected as shown in
Figure 5, the Input impedance test can be
run by clicking on the “Run” button. As the
test progresses, you will see the graph being
drawn on the screen. When the graph is
complete, the numbers below the graph will
be updated, see Figure 6.
At the end of this test, we positioned Cursor
1 (C1) at the low frequency peak. In the C1
column under the graph, we can read the
impedance and frequency where the cursor Figure 6
is positioned. In this case, 57.575 Hz and Speaker Input Impedance
25.19 ohms. By moving the cursor the
impedance at any frequency can be found. The true impedance of this 8 ohm
speaker is about 4.8 ohms at 20 Hz and reaches 142 ohms at 20 KHz.
140 Independence Lane
Chalfont, PA 18914
www.CoreTechGroup.com -5–
Core
Technology Group Application Note 10
AN-10
Keep in mind, this measurement includes all the components in the system
including both the woofer and tweeter along with the crossover. A word of
warning is worthwhile here. Never power-up an amplifier connected to a
crossover network with the drivers disconnected. It may be tempting during
testing, but at a frequency equal to the series resonant frequency of the
inductor and capacitor, the network may present an almost dead short circuit
to the amplifier (depending on the filter type), which could damage the
amplifier, crossover or both.
Woofer Impedance
One area where measurement is essential when designing passive crossovers, is
the loudspeaker driver itself. There is usually very little information in the
manufacturer’s data that will prepare you for the behavior of a loudspeaker /
crossover network combination, and these data are usually derived empirically.
In almost every case, the crossover frequency selected for the woofer and
midrange driver will be at a frequency where the voice coil inductance peaks.
As frequency increases, the effect of the voice coil inductance increases the
driver's impedance, and this plays havoc with the crossover’s performance.
Because the voice coil is surrounded by conductive metal (steel), its inductance
is not linear with frequency. This creates a lossy inductor, and as such it cannot
actually achieve the 6dB/ octave (20dB/ decade) impedance rise that one
would normally expect. The actual rise varies from one driver to the next, and
cannot be simulated in general terms - it must be measured.
Using a Frequency Response Analyzer,
the impedance of each driver can be
individually tested. Simply connect the
speaker, without the crossover, directly
across the wires from the Frequency
Response Analyzer. The Woofer
impedance and phase are shown in
Figure 7. The graph looks a lot like low
frequency portion of the system
impedance in Figure 6, which indicates
the woofer impedance is effecting the Figure 7
Woofer Impedance
crossover; which it should.
140 Independence Lane
Chalfont, PA 18914
www.CoreTechGroup.com -6–
Core
Technology Group Application Note 10
AN-10
Note: The impedance of any single component or any network may be tested
with this setup. With reactive components, such as capacitors and inductors, we
can use the magnitude and phase to calculate reactance of the component.
The graph can also be changed to “Impedance (Real, Imaginary)” and the
reactance can be seen directly as “Z Img”. For more details refer to Application
Note (AN-6) for more information on Measuring Inductors and Capacitors.
Tweeter Impedance
Unfortunately, it is extremely difficult to get the required details for most tweeters.
The resonance is almost always quoted, but most of the time it is difficult to find
the voice coil inductance, which would enable one to accurately calculate the
required crossover network.
The Q of the crossover circuit must be the same as the resonance Q, or it will
simply form a sharp notch in the middle of the resonance peak (Q too high) or a
broad notch that spans the resonance (Q too low). The traditional way to
measure Q is to measure the bandwidth between the -3dB frequencies, then
divide the resonant frequency by the bandwidth. However, a tiny measurement
error will cause a large final calculation error. Fortunately, this is rather easy to
measure accurately with the SA/RA Series Frequency Response Analyzer.
Further discussion in this area leads us into the Thiele – Small Parameters which is
outside the topic of this application note.
Figure 8
Frequency Response Setup
140 Independence Lane
Chalfont, PA 18914
www.CoreTechGroup.com -7–
Core
Technology Group Application Note 10
AN-10
Frequency Response Measurement
To measure frequency response of a network requires changing the
connections at the Frequency Response Analyzer as shown in Figure 8.
Frequency Response Setup
When measuring frequency response of an active or passive circuit, connect
the output of the Frequency Response Analyzer across the input of the device
as you would any signal generator. We select Bode (Phase, Magnitude) as the
graph mode. No changes are required to the Sweep Parameters or Advanced
settings unless the test requires a different frequency range. For our purposes,
the frequency range will remain 10 Hz to 40 Khz. For these tests, Input A is used to
monitor the frequency input to the network and Input B monitors the device
output.
NOTE: Should Input A and Input B be reversed, the resulting waveforms will be
reversed vertically.
Woofer Frequency Response
With the probe from Input B connected to
the woofer positive output of the crossover,
we can see the frequency response through
the crossover and across the woofer in Figure
9. The Bode graph shows the attenuation of
the input signal in dB across the frequency
spectrum. It looks just like you might expect a
crossover to work.
We can move a cursor to the – 3 dB point
and directly read the frequency, as shown in
Figure 9. In this test, we were unable to get to
-3 dB because there were not enough
sample points in the test we performed.
Should this occur, simply increase the
number of Points in the Sweep Parameters
tab and re-run the test. Figure 9
Woofer Frequency Response
140 Independence Lane
Chalfont, PA 18914
www.CoreTechGroup.com -8–
Core
Technology Group Application Note 10
AN-10
Tweeter Frequency Response
To measure the Tweeter frequency response,
simply move the probe to the positive tweeter
terminal. In Figure 10, the frequency response
of the tweeter can be seen. Here we can
measure the -3 dB point at 11 KHz.
Unlike the woofer with its smooth roll-off, the
tweeter frequency response has several
humps below the crossover frequency. The
peak of these humps are at about -38 dB,
which this manufacturer determined was Figure 10
enough attenuation for their speaker. Tweeter Frequency Response
By making the tweeter frequency
response the Reference (Data /
Current --> Reference) and running
the woofer test again, we can see
both response curves simultaneously
in Figure 11.
You can clearly see the crossover
frequency (Red lines) is at 2 KHz. You
can also see a bump in the tweeter
frequency response at about 150 Hz.
There is a 46.6 degree phase shift
between the speakers at this
frequency indicated by Cursor 2
(Yellow). To find the resulting output
level at this frequency requires
converting the db levels back to
voltage as follows:
𝑑𝐵
Figure 11 𝑉 = 10 20
Overlaid Crossover Waveforms
140 Independence Lane
Chalfont, PA 18914
www.CoreTechGroup.com -9–
Core
Technology Group Application Note 10
AN-10
Then perform vector math on the two phasors: 1.099 V (V1) @ -21.33° (Φ1) and
0.013 V (V2) @ 25.35° (Φ2).
𝑅𝑒𝑎𝑙 𝑇 = (𝑉1 𝑐𝑜𝑠Φ1 ) + (𝑉2 𝑐𝑜𝑠Φ2 )
𝐼𝑚𝑎𝑔𝑖𝑛𝑎𝑟𝑦𝑇 = (𝑉1 𝑠𝑖𝑛Φ1 ) + (𝑉2 𝑠𝑖𝑛Φ2 )
Now that both the real and imaginary values have been found the magnitude
of voltage, VT is determined by using Pythagoras’s Theorem for a 90o triangle as
follows:
𝑉𝑇 = √(𝑅𝑒𝑎𝑙)2 +(𝐼𝑚𝑎𝑔𝑖𝑛𝑎𝑟𝑦)2
Once we have the total we convert it back to dB as follows:
𝑑𝐵 = 20𝐿𝑜𝑔(𝑉𝑇 )
Using Excel for the calculations, the spreadsheet looks like this:
From FRA dB → Volts Calculations
Magnitude 1 0.8186 dB 1.098829 V1 Real -0.86729
φ1 -21.33 Deg Imaginary -0.67471
Magnitude 2 -37.94 dB 0.012677 V2 Real 0.012379
φ2 25.35 Deg Imaginary 0.002732
RealT -0.85491 V
ImaginaryT -0.67197 V
VT 1.087394 V
Converted to dB 0.727741 dB
Figure 12
Excel Spreadsheet of Waveform addition
140 Independence Lane
Chalfont, PA 18914
www.CoreTechGroup.com - 10 –
Core
Technology Group Application Note 10
AN-10
The resulting waveform at 150.159 Hz is about 0.73 dB which reduces the woofer
output by about 0.091 dB. Of course, room acoustics will play a much larger role
in the final frequency response of the speaker.
Summary
This Application Note has demonstrated how to use the SA/RA Series Frequency
Response Analyzer to measure the impedance and frequency response of
crossovers and speakers. When designing crossovers, engineers perform
calculations and make simulations regarding crossover performance. These
calculations can be verified and the data gathered can be used by the
engineers to improve their simulation models.
In addition, speaker performance can be quickly verified in production against
a saved test, to see how manufacturing variations and component tolerances
are effecting circuit performance.
In short, the SA/RA Series Frequency Response Analyzer is an essential tool for
engineers to verify their designs and to insure component variations do not
render performance outside product specifications.
For additional Information on how to measure the characteristics of audio
amplifiers can be found in application note “AN-9 Audio Circuit Testing”.
140 Independence Lane
Chalfont, PA 18914
www.CoreTechGroup.com - 11 –
