Contents

 Inductors vs capacitors – how are they different?

o What does an inductor do?
o How does a capacitor work?
o Common capacitor types to know
 How does a speaker crossover work?
o Crossover slopes explained
o 1st order crossover with a capacitor (high pass filter)
o 1st order crossover with an inductor (low pass filter)
o 2nd order 2-way crossovers with capacitors and inductors
o How they work
o Crossover design types
o Crossover phase concerns
 What is a crossover Zobel network?
Inductors vs capacitors – how are they
different?

https://soundcertified.com/what-does-capacitor-inductor-do-in-speaker-
crossover/

Inductors and capacitors are the basic components behind all kinds of wonderful audio
electronics including speaker systems at home or in the car. Each is considered a passive
component parts type as they don’t require a power supply to function.
What’s really cool is how they’re somewhat like polar opposites but can work together in a
non-powered crossover (passive crossover) to greatly improve the sound quality & get the
most from your speakers.

What does an inductor do?

An inductor is a tightly wound coil of wire with a specific number of loops used to take
advantage of a property called inductance. Inductance is the tendency of a conductor
(specifically a coil of wire in this case) to oppose a change in the electric current flowing
through it due to the magnetic fields it generates.
They’re used in electric motors, solenoids, spark plug coils, and, of course, speaker crossover
designs. Inductors behave like the opposite of capacitors: an inductor opposes high
frequency signals meaning it passes low frequency audio signals more easily.
An electrical behavior called reactance is what makes this possible. When the frequency
changes, so does an inductor or capacitor’s opposition to the flow of electricity.
 Inductive reactance: An inductor builds up a higher resistance (impedance) to
current as the frequency increases due to the inductance it has.
 Capacitive reactance: a capacitor builds up a higher resistance (impedance) to
current as the frequency decreases due to its capacitance.
Units of measurement for inductance

For inductors, the unit of measurement is the Henry. By convention, inductors are usually
sold in units of milliHenries (1/1,000th of a Henry, or .001 Henries). A typical speaker
crossover might use an inductor with a value of 10 mH, for example.
How does a capacitor work?
Capacitors store an electrical charge using extremely thin and tightly wound electric
conductors separated by an insulator. This can be an electrolyte, mica, or several other
types of materials. While they do not allow a direct current (DC) signal to pass, they do
allow alternating current (AC) voltage & signals to pass.
They have an interesting characteristic: capacitors allow only high frequencies to pass –
they increase their impedance (frequency-based resistance) when lower frequencies are
applied.

The point at which this takes place is carefully chosen to be the crossover frequency. There
are two fundamental types of capacitors with one in particular used for audio.

Units of measurement for capacitance

For capacitors, the unit of measurement is the Farad. By convention, capacitors are often
sold in units of microFarads (1/1,000,000th of a Farad, or .000 001 F), sometimes written
with the Greek letter mu “µ” to represent “micro.” As an example, when buying capacitors
for your own crossovers you’ll see capacitors listed in “µF” sometimes.
Smaller ones may use picoFarads (pF) or nanoFarads (nF) which are even smaller and are
used in electronics.

Capacitors used in audio nearly always tend to be in the microFarad range. For example, a
bass blocker to a tweeter may use a 47 µF capacitor.

Common capacitor types to know

1. Electrolytic capacitor

Electrolytic capacitors are essentially the most common and the most affordable type, hence
their popularity in all kinds of electronics and speaker applications. You’ll frequently find
them in a passive crossover or directly connected to a tweeter as a high pass crossover.

They have a thin metal case and contain an electrolyte between the super-thin charged
conductive plates inside.

A non polarized electrolytic capacitor allows passing an alternating current (AC)

waveform like that used for a musical signal. They’re also called “bipolar” capacitors. DC
types, on the other hand, can’t and should only be used for direct current like in a power
supply.
Note: non polarized capacitors normally are marked as such and are the only types you
should use for audio. Polarized capacitors can be damaged (and even explode!) if used for
audio applications. At the very least you’ll have a distorted sound that you’ll hate.
2. Film capacitor

A film capacitor uses a thin film material to separate its charged plates and is typically a bit
more expensive. They’re also longer-life, have better performance (in some cases) for audio,
and may have higher temperature limits.

Film capacitors are also offered in high voltage types which are great for vacuum tube audio
designs. They’re also a good upgrade for cheaper electrolytic capacitors if you’re handy with
a soldering iron.

3. Ceramic capacitors
Ceramic capacitors aren’t normally used in crossovers because their capacitance values are
usually quite small (in the picoFarad range, for example) while we often need microFarad
range values for speakers.

They’re used for other purposes typically, such as a bypass capacitor in a power supply to
control electrical signal noise or in an active crossover.

How does a speaker crossover work?

A crossover uses a capacitor, inductor, or both to limit the frequency range of audio sent to
one or more speakers. This is extremely useful for preventing bass frequencies from reaching
a tweeter or harsh-sounding midrange and treble from reaching a subwoofer.

The crossover point is often recommended by the loudspeaker manufacturer or picked as a

good compromise between the limits of the frequency response of each speaker used. The
filter frequency (also called “the corner frequency” or “Fc” at times) is directly related to the
speaker impedance.
If you know the values of the inductors and/or capacitors you can determine the speaker
crossover frequency based on the impedance it’s used with.

Crossover slopes explained

When we talk about the “order” of a crossover network, we’re referring to the number of
stages (sections). This affects how effectively the slope – the audio filtering ability – is.

A 1st order design uses a single inductor or capacitor while 2 make up a 2nd order, three a
3rd order, and so on. Each stage (order) has a -6dB per octave slope with -12dB/octave being
one of the most commonly used both for speaker or amplifier crossovers.

1st order crossover with a capacitor (high pass

filter)
A high pass filter works by passing higher frequencies to a speaker and opposing lower
frequencies. At lower frequencies the impedance of a capacitor has a very high Ohm value,
greatly reducing output voltage to the speaker.

Likewise, the opposite is true at high frequencies. You’ll often find them installed on a car or
home audio tweeter set to block distorting and potentially damaging bass from being played.

1st order crossover with an inductor (low pass filter)

A low pass filter works by blocking higher frequencies to a speaker and allowing lower
frequencies to pass. At higher frequencies, the impedance of the inductor means it has a very
high Ohms value, greatly reducing the output to a speaker.

This type is normally used with a woofer or mid range speaker to prevent treble or “highs”
they can’t produce well from being played.

2nd order 2-way crossovers with capacitors and

inductors
How they work
A 2-way, 2nd order speaker crossover network is essentially a high pass and low pass
crossover filter combined in parallel.
However, as they add a second stage (2nd order, -12dB/octave filter), they’re better
performers than a simple 1st order design. They’re the most common type also, using a
woofer & tweeter to create a 2-way speaker system for home audio speakers or a component
speaker set for car audio.

Even inexpensive 2-way 2nd order component sets can sound excellent with speaker drivers
of decent quality and adequate design.
Because of the overlap point at their cutoff frequency, things can get a bit more complicated
when it comes to 2 way crossovers as I’ll explain further below.

How it works:

 Capacitor C1 reduces voltage output to the tweeter below the cutoff

point. Inductor L1 reduces high-frequency signals that have reached it even
further by passing them to the ground/negative (-) amplifier return path.
 Inductor L2 reduces voltage output to the woofer above the cutoff point,
passing low frequency signals to the speaker. Capacitor C2 passes additional
high-frequency signals that have reached it to the ground/negative (-) amplifier
return path.
The end result is two crossover stages that are staggered in series, meaning they compound
together for a crossover slope that’s 2x as effective as a single stage (-6dB/octave) design.
This is a -12dB/octave slope.

Why are 1st order crossovers used if 2nd order ones are better?

While 1st order crossover networks are less common now, they’re still around. You’ll often
find them in:

 Budget 2-way speaker cabinet systems

 Designs where the speaker’s natural rolloff (falling frequency response) can be
used to reduce the components required to achieve the same effect.
 Simple inline speaker crossovers like bass blockers for car stereo use and
related applications where something quick & easy is ideal.

Generally speaking, however, 12dB/octave crossovers are the most popular as they’re a good
compromise between cost, parts count, and complexity. In fact, car amplifiers and home AV
receivers commonly use a 12dB/octave (12dB/octave) design even in their electronic active
crossovers.

Crossover design types

There’s a range of possible crossover network designs a designer might choose from, but a
few are preferred over others:

 Butterworth: In its standard configuration, the Butterworth design sums to

+3dB at the cutoff frequency overlap. This can be handy for various design
needs like accounting for different speaker drivers’ offsets (spacing between
speakers at their central acoustical point).
 Linkwitz-Riley: This type sums to a flat (0dB) output at the crossover frequency
overlap and is one of the most commonly used. It’s an ideal choice in most
cases.
 Bessel: A Bessel design is not considered “all pass” like a Linkwitz-Riley and
does not sum flat at the frequency point.
 Chebyshev: Not often used, this can be helpful where there’s a need for a
boosted output at the crossover point as the Chebyshev provides a +6dB sum.
Why does this matter? Ordinarily, it doesn’t matter much for the average person.
However, if you’re interested in making your own speaker crossover it’s helpful to
understand the options possible.
Each crossover network type uses a slightly different mathematical formula set to calculate
the parts values you’ll need. In all cases, a good speaker design book can help you do it
yourself if you’re interested in getting the best performance or if you’re interested in do-it-
yourself (DIY) speaker projects.
Want to learn more advanced speaker design skills or make your own crossovers? I highly
recommend the Loudspeaker Design Cookbook by Vance Dickason. It’s full of excellent
information!

Aftermarket home or car speaker crossovers generally use a Linkwitz-Riley design and sum
to a 0dB level at their crossover point. One reason is that it’s assumed the speakers you’re
using will have an adequate output near that point. If that’s not the case you can investigate
more advanced designs yourself and compensate for a particular speaker’s weaknesses.
Crossover phase concerns
Crossovers have another issue to contend with: each single crossover component adds a
90° phase (shift) to the signal sent to the speaker. Capacitors and inductors have a “phase
shift” when a signal passes through them.
For simple single-stage crossovers, this isn’t really a concern as it’s not something you’re
very likely to notice, although it’s a detail for more advanced speaker designs. However, for
2nd order designs, this means there’s a 180° difference between the two outputs, often
resulting in an “odd” sound and also means the sound isn’t arriving at the listener’s ears at the
same time.

To remedy this, 2nd order/even order crossovers normally have the tweeter output
reversed. This reversal of one speaker output means both speakers are “in phase” and there’s
no longer an issue with sound delay. If you’ve ever bought 2-way speaker crossovers you
probably didn’t even know it was designed that way on purpose!
What is a crossover Zobel network?
A Zobel network is an impedance equalization network used to compensate for the rise
in speaker impedance over the frequency range due to voice coil inductance. For
example, many speakers with a voice coil commonly show a rise in their total impedance as
the frequency increases. You can see this on impedance graphs plotted for speakers.
Because a crossover’s behavior is directly affected by impedance, this simple design can
improve the speaker system’s performance by compensating and “flattening” the normal rise
in the speaker Ohm load that the crossover sees.

The network uses a simple RC (resistor-capacitor) network in parallel with a speaker driver to
offset the impedance, resulting in the crossover seeing a nearly flat impedance over the
frequency response range.

The resistor ensures the minimum total impedance is always met while the capacitor works to
decrease the total crossover Ohm load as the frequency rises.
The Crossover - Brain of your Loudspeaker
System
https://www.audioholics.com/loudspeaker-design/crossover

by Paul Apollonio — June 09, 2009

Contributors: Gene DellaSala

RBH Sound T30-LSE Crossover

Have you ever wondered how the tweeter and woofer have their responses
combined in a speaker system, or what the crossover network is and how it
works? If so, this article will shed some light on the least seen and perhaps
most undervalued part of the speaker system, the crossover network.
Unlike the electronic world where the filters have gone from using tubes, to
analog computers to digital signal processors, modern day passive
crossovers, a pair of frequency selective filters designed to work together
as a single network, are basically the same as they were about 100 years
ago when the speaker business was in its infancy. While crossover
components and materials and construction have changed, the underlying
theory and practice has not.

What does the Crossover Do?

Like your own brain, the crossover network is a director. In addition to its
primary function, the passive crossover in practice is expected to do more
than simply split the frequency band in two.
1) The crossover allows you to place two 8 ohm speakers in parallel, one
for the highs and one for the lows, and still present an 8 ohm, not a 4 ohm
load to the amplifier (which is what the amp would see without the network)
2) The crossover allows you to match two drive units of different
efficiencies so that the combination of the two does not result in a large
shelf in the frequency response.
3) The crossover allows you to equalize the frequency response of either
the woofer or tweeter, making the composite response smoother than the
unfiltered curves of either driver (woofer or tweeter) would seem to indicate
is possible.
Let’s get back to the basic job of the crossover, and see if we can
understand how it accomplishes this by learning a few new but simple
concepts. The basic job is to send the high frequency information to the
tweeter, and eliminate this same band of high frequencies from the woofer.
Next, send the low frequency information to the woofer, and eliminate this
same band of low frequencies from the tweeter. A crossover is in essence
two frequency filter sections working in parallel. A high pass section and a
low pass section. These terms are self explanatory. The high pass passes
high frequencies, the low pass passes low frequencies. The high pass is
placed in series with the tweeter, and the low pass is in series with the
woofer. A normal two way crossover will have six terminals. Two in, and
four out. Usually the two sections (high-pass and low-pass) are in parallel
and both sections ONLY work properly when the woofer and tweeter are
both attached and operational. To understand how the crossover does its
job, we must first understand the concept of impedance.
The three main passive elements used in all crossovers are resistors,
inductors, and capacitors. Impedance is a measure of how much the
resistor, capacitor or inductor impedes the flow of electrons at any given
frequency. Resistors are the simplest components, and the easiest to
understand. The have a constant resistance to the flow of electrons which
does not vary with the frequency of the signal. Resistors do not change the
phase of the signal, the time relationship between voltage (electrical
pressure) and current (flow of electrons). They only consume power and
convert it to heat. That is all they do. In a perfect speaker system there is
no need for resistors at all. Of course, while there are no perfect speakers,
all the basic crossover theory's we use start with that premise.
The second element we shall consider is a capacitor. The capacitor
impedes the flow of electrons through it in a way which is NOT independent
of frequency. (It would not be technically correct to use the word resistance,
but you can certainly think of it as such to conceptualize this). The
capacitor has an impedance which is inversely related to frequency. This is
to say when you double the frequency of a signal applied to the capacitors
terminals, its impedes that electrical flow by half as much as it did at the
lower frequency. If you halve the frequency applied, the capacitors
impedance (How much it impedes the flow of electrons) is doubled. This is
a very useful characteristic which we use to create networks which are
frequency selective (like a crossover).

The Maths
The impedance of a capacitor is determined by the equation: Z =
1/j(2piF*C)
Where Z = Impedance magnitude,
J = the square root of (-1) ("Imaginary" component to describe the phase)
pi = 3.1414 etc
C = Value of Capacitance in Farads; 1 Farad = 1,000,000 ufd
I think the first time I read about the square root of (-1) I had an allergic
reaction. So, for those of us who are finding it difficult to comprehend phase
for the first time, we can pretend it is not there, and to simplify the
discussion we will admit to this equation being somewhat incomplete
without the “j”, but for the determination of the steady state and not
transient value of the impedance, we are OK.
When a capacitor is put in series (shares one electrical terminal) with a
resistor, the capacitor acts as a high pass filter. At some frequency it has
the same impedance as the resistor, and at a much higher frequency it is
like a short, just a piece of wire with little to no resistance (at least in a
perfect capacitor). At very low frequencies the impedance of the capacitor
becomes so high, that essentially nothing can pass through it. It is this
characteristic which is used in endless electronic circuits to allow AC
signals to pass through yet completely block DC voltages from one part of
the circuit to another.
 Figure 1

Lets say we have a 10 ufd capacitor and put in in series with an 8 ohm
resistor. At some frequency, the capacitors impedance magnitude will be
equal to the 8 ohm impedance of the resistor. (Which of course is still 8
ohms at any frequency from DC to daylight.)
Lets do the math.
Z = 8 = 1/{(2)*(3.14)*(F)*(10)*(10-6)}
The capacitor is MICRO Farads, which is millionths of a Farad, so we must
multiply the 10 by 0.000001( 10-6) to convert microfarads into Farads, so we
can solve for the frequency (F) in Hz. By cross multiplying (multiplying both
sides by F and dividing both sides by 8) we get:
F = 1/(2*3.14*8*10*10-6)) = 1990 Hz.
What this tells us is that at 1990 Hz (2 kHz if you are among friends) the
impedance of the capacitor is equal to that of the resistor, so at this point
the power into the resistor is half what it would be without the capacitor. Put
another way, at 1990 Hz, the power into the resistor is 3 db less and this is
your crossover point. Above this frequency the attenuation (reduction in
power to the resistor) is less, and as we go lower in frequency the
capacitors impedance grows larger, while the resistor is unchanged, so the
attenuation of signal to the resistor increases. At a frequency of 995 Hz, the
attenuation will be 6 db. At 498 Hz, it will be 12 db. We call this a first order
or 6 db per octave crossover. That is, in a nutshell, how this works. It grows
in complexity when more parts are used, but the basic function is the same.
Now if your speaker is a perfectly resistive load, your 10 ufd series
capacitor will be a perfect passive 6 db per octave High Pass (HP) filter.
Don't worry, life is not that simple.
The other important element used in a crossover network is the Inductor;
sometimes called a choke, and sometimes called a coil. They are all refer
to the same thing. The Inductor is in many ways the inverse of the
capacitor. It too changes impedance with frequency, but in a way entirely
opposite to the capacitor. The inductors impedance grows LARGER with
increasing frequency, not smaller like the capacitor. Its impedance grows
larger at the same rate as the capacitor grows smaller, and the similarity
ends there. So if we replace our capacitor with an inductor in the first case
we have a first order low pass filter. (As shown in figure # 1)
Note that in higher order networks, second order and above, both inductors
and capacitors are used in the same circuit to provide high-pass (HP) or
low-pass (LP) filtering. In the HP filter, capacitors are put in series with the
load while inductors are put in parallel. (A resistor represents the load in
figure 1, and the speakers represent the load in figure 2). In the LP filter the
opposite is true, the inductors are in series and the capacitors in parallel
with the load (or speaker if you are making a crossover network). Hopefully
this makes sense if you keep in mind the direction of the impedance
changes of the parts with increasing or decreasing frequency.
 Figure 2
In the network pictured in figure 2, the HP blocking capacitor, C1, increases
in impedance with decreasing frequency, while at the same time the
shorting inductor L1, decreases in impedance with decreasing frequencies,
hence shorting out the low frequencies while the capacitor is increasingly
blocking them. The inverse is true of the LP filter. L2 is blocking highs and
passes lows, while C2 is shorting out the highs and passing the lows. This
is how the LP filter blocks highs and passes lows while the HP filter is
performing the inverse function. It is this dual action that allows this kind of
network to create twice the blocking action (12 db) per octave of frequency
than the simpler 6 db/octave network. If you did not understand this on the
first read, that is OK, you have a lot of company. Think about how the
capacitor and inductor change their impedance with frequency, and take
another look at this circuit diagram. If you try, you can figure out what is
going on as the frequencies change.
Back to the simplest network, the 6 db per octave. If I have an 8 ohm
resistor, and a 0.001 Henry (1.0 millihenry inductor), at what frequency do I
have a 3 db attenuation across my resistor? Let's calculate the frequency
where the inductor equals 8 ohms.
The value of the impedance of an inductor is determined by the formula
Z = {(2pi)*(J)*(F)*(L)} where F= Frequency in Hz, L = Inductance in Henries
Z= Impedance in Ohm
Like we did for the capacitor, we shall ignore the imaginary "J" term and
push ahead.
8 = (2pi)(F)(0.001H) Therefore: F = 8/{(2pi)(0.001) = 1273 Hz
If we want to set the Low Pass frequency equal to the High Pass frequency
of 1990 Hz, what inductor do we need? We set F = 1990 and then we can
solve for L.
L = 8/{(2pi)(F)} = 6.4 + 10-4 Henries = 0.64 millihenries

We have calculated the parts needed for our first crossover network for our
8 ohm tweeter and 8 ohm woofer with a crossover frequency of 1990 Hz
Network below:
 FIGURE 3

Loudspeaker Crossovers cont.

So What’s all the Fuss?
Now the problem with an exercise like this is that the chance that the
crossover network will do what you might expect of it is somewhere
between slim and none. The main reason for this is that loudspeakers do
not have a constant impedance, (like an 8 ohm resistor) nor do they posses
flat frequency response through the crossovers region immediately on
either side of the crossover point. Let's take a look at a real world example.
Let’s say we want to build a system, and we go to a parts vendor and buy a
woofer and tweeter we think should work well together. Not being gurus on
the subject, we decide to keep it simple and use a simple 6 db per octave
network. It is the least complicated, least expensive, and least sensitive to
component tolerance, plus it is the only network able to offer (theoretically)
perfect transient response characteristics. We will load the driver data from
the Xover Pro 3 database and use this program to calculate the crossover
parts for a 2000 Hz crossover point as shown below.
 FIGURE 4
Unfortunately there are characteristics inherent in the speakers we use
which cause issues with this kind of simple network. Let's take a look at
these results from Xover Pro version 3.

Figure 5 -Peak in tweeter Impedance results in frequency response peak in

combined response
This particular tweeter has a resonance at 900 Hz. As we approach the
tweeter resonance the Impedance magnitude peaks. The capacitor can do
nothing to prevent that, so it does not do a good job of attenuating the
signal across the actual speaker the way it would across a simple resistor
whose impedance does not change with frequency.

Figure 6 -Impedance of Tweeter and series capacitor (red line) peaks at Tweeter
Resonance
If you refer to the original Audio Cyclopedia by Howard Tremaine, it refers
to the set of equations widely available in reference texts like the
Loudspeaker Cookbook to calculate crossover component values as
"constant resistance networks". While a good number of real speakers
have relatively flat impedance regions, we rarely cross over those speakers
in that region. It is almost always at the extreme upper frequency range of
woofers where they have a mostly inductive impedance; or the lower end of
tweeters where the HF devices have resonances and therefore impedance
peaks as in our example above. In fact many woofers have a voice coil
inductance considerably higher than the series inductance required for the
first order crossover network. This is why you will often find impedance
compensation networks called Zobel networks used in parallel with the
woofers.
 FIGURE 7 - 1ST ORDER LP WITH 2ND ORDER HP FILTER
Since our Dayton woofer is small as is its voice coil, it has a very low self
inductance, so in our example we do not need a Zobel network for a good
result. On the other hand, our problematic tweeter has to be attenuated a
bit to lower its sensitivity to match the woofer, and it also needs a parallel
choke to short out that resonant peak so the series capacitor can do its job.
This is a second order network and is a bit more complex than a first order.
While it brings with it the issues of increased phase shift, more parts and
more expense, it also brings with it a greater amount of control over the
responses of the speakers in and around the crossover region. By fudging
a bit the elements generated by Xover Pro 3, I am able to massage this
curve to look very flat.
Figure 8 - High Pass, Low Pass and Combination Response using Network shown
in figure 7
Now this beautiful curve is not the result of a measurement, but in fact a
simple modeling tool which makes a lot of assumptions that may not be
true. (Both speakers are in the same horizontal and vertical plane, both are
minimum phase, etc.) That said, it sure looks a lot better than what we
started with. As you can see by the impedance curve below, there are no
sudden jags in the curve, and the impedance of both the HP and LP
sections are smooth and quite symmetrical.

Figure 9 -Woofer, Tweeter and Combination Impedance Magnitudes using Network

in Figure 7
Let's take a look at one more example before we finish. We are going to
replace our woofer with one having a much more inductive voice coil,
arbitrarily chosen to be 1.0 millhenry. In this example, we will use a second
order network, and see what kind of change we get from using or not using
a Zobel network, an impedance compensation network to make the woofer
look non-inductive to the crossover. The Zobel network is the series
combination of a resistor and capacitor together in parallel with the woofer.
In the schematic below, they are shown as Rz and Cz.
 Figure 10 - Woofer with 1st Order LP and Zobel Network
Below is a graph that shows the woofer in three instances. No network. 1.7
mh in series with no network. The same woofer and 1.7 mh and with the
Zobel network.

Figure 11 -Woofer raw response - first order LP no Zobel - first order LP with Zobel
The two curves which parallel each other at very high frequencies are the
raw woofer (top) and the woofer with the single series inductance (1.7 mh -
lower). You will notice that the inductor simply adds about 6 db of
attenuation and then offers no more rejection. (This is evidenced by the
curves tracking each other at high frequencies). The lowest curve is that of
the woofer using both the series inductor and the Zobel network. By
returning the apparent impedance of the speaker to that of a simple
resistance, the Zobel network allows the series inductor to do its job, that is
to offer ever increasing attenuation with increasing frequency. Below is the
Impedance magnitude of the speaker with the Zobel Network (flat at high
frequencies), the speaker with the series inductor, (highest impedance at 2
kHz) and the speaker with the series inductor and the Zobel network both
(middle curve). What this shows is that the inductance of the speaker and
the inductor add together to present a very high impedance to the driving
amplifier, but not a very effective filter of high frequency information.

Figure 12 - Impedance Magnitude of Vented woofer with Zobel (bottom)

Zobel and Choke (middle) and No Zobel network and choke (quickest to the top)
In practice, very few woofers give us what is called in filter speak, a
monotonic decay. This means the response goes away and does not return
(as is shown in graph # 7). There are no ripples in that portion of the
spectrum we want to deny the speaker, called in filter-speak, the stop-
band. This is a really important characteristic of woofers, and very few are
well enough behaved in the midrange to not be a problem in this regard. If
the woofer has peaks at 3 KHz, and we attempt to cross it over at 2 KHz,
you will likely hear the peaks, especially if you are in the woofers direct line
of sight. For sure you will see their effects in the combined response
through the crossover. As we can see from the above example, making
sure the filter can actually attenuate the signal increasingly with increasing
separation from the crossover frequency is a very important function with
real world speakers, and the Zobel network is an important tool in the
arsenal of crossover designers.

So What’s Next?
Now at the outset of this article, I said that we were making certain
assumptions about the actual parts we use in crossover networks being
theoretically perfect. In the real world, capacitors, inductors and resistors
exhibit real world behavior which is neither ideal or perfect. In this part, we
considered the importance of real loudspeaker impedance, and how it acts
as a monkey wrench to complicate and frustrate the function of constant
resistance type crossover networks. In part two of this article, I will discuss
non-ideal behavior of real world parts. This I hope, will allow the reader to
gain some insight into the kinds of mistakes made in crossovers that can
be seen by simply examining the networks, and some flaws which are not
quite so obvious. We will also look at a real world example of a crossover I
build, and how far it actually deviates from the theoretically predicted
network.
Filter & Crossover Types for Loudspeakers
https://www.audioholics.com/loudspeaker-design/filter-crossover-types-for-
loudspeakers

by Gene DellaSala — August 29, 2004

Contributors: Shane Rich

The filter type can be described in several different ways. Low-pass and
high-pass filters in two-way crossover networks are often identified by their
"Q". The Q is the resonance magnification of the filter and it is recognized
by the shape of the "knee" of the amplitude response. Filters with a high Q
tend to "ring" and exhibit poor transient response. Unlike drivers and boxes
which use only numerical values for Q, filters are sometimes named after
the engineer(s) who first described them. Some examples are shown in the
amplitude response graph below.

Filter Types

The filters in three-way crossover networks (and some two-way networks)

are often identified as either "APC" or "CPC" depending on the way they
combine. APC stands for "All-Pass Crossover" and it refers to those
crossover networks whose filters sum to create a flat voltage output. APC
networks are generally considered the best choice because they make it
possible for the speaker to have a flat on-axis amplitude response.
Common APC networks include 1st- and 3rd-order Butterworth filters and
2nd- and 4th-order Linkwitz-Riley filters. CPC stands for "Constant-Power
Crossover" and it refers to those crossovers whose filters sum to provide a
flat power response. The power response of a speaker is the total of both
its off-axis and on-axis amplitude response. In other words, it is the total
acoustical power that is radiated into a space. CPC networks can be
beneficial in reverberant environments where the off-axis response is
important.
The difference between APC and CPC networks can be understood
electrically by a comparison of their input to output voltages. APC networks
satisfy the following expression:
[VI] = [VL + VM + VH]
This means the absolute value of the input voltage will equal the absolute
value of the sum of the output voltages of each filter at all frequencies. CPC
networks satisfy the following:
VI^2 = VL^2 + VM^2 + VH^2
This means that the square of the input voltage will equal the sum of the
squares of the output voltages of each filter at all frequencies.

Filter Summary
These generalizations assume that the drivers are properly aligned at the
crossover frequency. This means that they are mounted in such a way that
the direct sound from each driver arrives at the listener's ear at the same
time at the crossover frequency. Another important assumption is that the
impedance response of each driver has been equalized so that it appears
to be approximately resistive to the crossover network. Also, the sensitivity
of the drivers is assumed to have been equalized with an appropriate L-
pad.
Finally, the following descriptions assume that all filters in the crossover
network are of the same type. If a two-way crossover network has a 4th-
order Linkwitz-Riley low-pass filter, it is assumed that it also has a 4th-order
Linkwitz-Riley high-pass filter. If you choose to use mismatched filters,
you'll have to rely on the your own measurements and experience to
determine the results.

1st-order Filters
Advantages: Can produce minimum phase response (Butterworth only) and
a maximally flat amplitude response. Requires the fewest components.
Disadvantages: Its 6 dB/octave slope is often too shallow to prevent
modulation distortion, especially at a tweeter's resonance frequency.
Achieving minimum phase and a maximally flat amplitude response
requires very careful driver alignment and only occurs when the listener is
located at exactly the same distance from each driver. It has a 90 degree
phase shift which can result in lobing and tilting of the coverage pattern.
Two-Way
1st-order Butterworth: Produces a -3 dB crossover point to achieve a
maximally flat amplitude response, minimum phase response and flat
power response that qualifies it as both an APC and CPC network. The 90
degree phase shift results in a -15degree tilt in the vertical coverage pattern
if the tweeter and woofer are vertically separated by no more than one
wavelength at the crossover frequency and if the acoustical depth of the
tweeter and woofer are carefully aligned at the crossover frequency. The tilt
will increase and lobing can become severe if the drivers are separated by
a greater distance or are misaligned. These problems appear as a ripple in
the amplitude response. Filter Q = 0.707.
Two-Way & Three-Way
1st-order Solen Split -6 dB: A custom version of the 1st-order Butterworth
filter (twoway crossovers) or 1st-order APC filter (three-way crossovers)
that uses a -6 dB crossover point to minimize the disadvantages of a
crossover network with standard 1st-order Butterworth or APC filters.
Three-Way
Note. 1st-order filters are usually not recommended for three-way
crossover networks because their shallow 6 dB/octave slopes do not
provide adequate separation. 1st-order APC: Produces -3 dB crossover
points to achieve a flat amplitude response.
1st-order CPC: (Seldom used.) Produces -3 dB crossover points to achieve
a flat power response.

2nd-order Filters
Advantages: Can produce a maximally flat amplitude response. Requires
relatively few components. Has a 180 degree phase shift which can often
be accommodated by reversing the polarity of the tweeter and which
produces minimal or no lobing or tilt in the coverage pattern. Is less
sensitive to driver misalignment than 1st-order filters.
Disadvantages: Although the 12 dB/octave slope is better than a 1st-order
filter, it may still be too shallow to minimize the modulation distortion of
many drivers.
Two-Way
2nd-order Bessel: Produces a -5 dB crossover point to achieve a nearly flat
(+1 dB) amplitude response. The summed group delay is flat. It has a low
sensitivity to driver misalignment and resonance peaks. Filter Q = 0.58.
2nd-order Butterworth: Produces a -3 dB crossover point that sums to a +3
dB amplitude response and a flat power response that qualifies it as a CPC
network. It has a medium sensitivity to driver misalignment and resonance
peaks. Filter Q = 0.707.
2nd-order Chebychev: (Seldom used.) Produces a 0 dB crossover point to
achieve a
+6 dB amplitude response with about ±2 dB of ripple. The summed group
delay has a significant peak just below the crossover frequency. It has a
medium sensitivity to driver misalignment and resonance peaks. Filter Q =
1 .0.
2nd-order Linkwitz-Riley: (Very popular.) Produces a -6 dB crossover point
to achieve a maximally flat amplitude response that qualifies it as an APC
network. It has a -3 dB dip in the power response. The summed group
delay is flat. It has a medium sensitivity to driver misalignment and
resonance peaks. Filter Q = 0.49.
Three-Way
2nd-order APC: Produces -6 dB crossover points to achieve a flat
amplitude response but the power response will have approximately 3 dB
of ripple.
2nd-order CPC: (Seldom used.) Produces -3 dB crossover points to
achieve a flat power response but the amplitude response will have
approximately 3 dB of ripple.

3rd-order Filters
Advantages: Can produce nearly flat amplitude response. With an 18
dB/octave slope, it is better able to minimize modulation distortion. Less
sensitive to driver misalignment.
Disadvantages: Requires more components. Has a 270 degree phase shift
which can result in lobing and tilting of the coverage pattern.
Two-Way
3rd-order Butterworth: (Popular for some D'Appolito mid-tweeter-mid
designs.) Produces a -3 dB crossover point to achieve a maximally flat
amplitude response and flat power response that qualifies it as both an
APC and CPC network. A 270 degree phase shift results in a + 15 degree
tilt in the vertical coverage pattern if the tweeter is wired with normal
polarity and a -15 degree tilt if the tweeter is wired with reverse polarity.
(D'Appolito mid-tweeter-mid designs overcome much of this tilt problem
and produce a more symmetrical coverage pattern.) It has better group
delay than a 1st- and 2nd-order Butterworth network. Filter Q = 0.707.
Three-Way
3rd-order APC: Produces -3 dB crossover points to achieve a flat amplitude
response but the power response will have a modest ripple (usually less
then 1 dB) that increases slowly as the spread between the two crossover
frequencies increases.
3rd-order CPC: (Seldom used.) Produces -3 dB crossover points to achieve
a flat power response but the amplitude response will have a varying
amount of ripple (typically 1 to 3 dB) depending on the spread between the
two crossover frequencies.

4th-order Filters
Advantages: Can produce a maximally flat amplitude response. With a 24
dB/octave slope it provides the best isolation between drivers resulting in
the least modulation distortion. Has a 360 degree phase shift which results
in "in-phase" response and which promotes minimal or no lobing or tilt in
the coverage pattern. Is the least sensitive to driver misalignment.
Disadvantages: Requires the most components. The increased number of
inductors can result in substantial insertion loss because of inductor DCR.
Two-Way
4th-order Bessel: Produces a -7 ½ dB crossover point to achieve a nearly
flat (-1 ½ dB) amplitude response. The summed group delay produces a
moderate bump just below the crossover frequency. Filter Q = 0.58.
4th-order Butterworth: Produces a -3 dB crossover point that sums to a +3
dB amplitude response and flat power response that qualifies it as a CPC
network. The summed group delay has a significant peak just below the
crossover frequency. Filter Q = 0.707.
4th-order Gaussian: (A seldom used filter that is constructed with an
asymmetrical filter topology.) Produces a -6 dB crossover point to achieve
a nearly flat amplitude response with moderate ripple. The summed group
delay produces a moderate bump just below the crossover frequency.
4th-order Legendre: (A seldom used filter that is constructed with an
asymmetrical filter topology.) Produces a -1 dB crossover point that sums
to a +5 dB amplitude response with minor ripple. The summed group delay
has a significant peak just below the crossover frequency.
4th-order Linear-Phase: (A seldom used filter that is constructed with an
asymmetrical filter topology.) Produces a -6 dB crossover point to achieve
a nearly flat amplitude response with moderate ripple. The summed group
delay produces a moderate bump just below the crossover frequency.
4th-order Linkwitz-Riley: (Very popular. Sometimes called a "squared
Butterworth" filter. Also used for some D'Appolito mid-tweeter-mid designs.)
Produces a -6 dB crossover point to achieve a maximally flat amplitude
response that qualifies it as an APC network. It has a -3 dB dip in the
power response. The summed group delay produces a moderate bump just
below the crossover frequency. Filter Q = 0.49.
Three-Way
4th-order APC: Produces -6 dB crossover points to achieve a flat amplitude
response but the power response will have approximately 3 dB of ripple.
4th-order CPC: (Seldom used.) Produces -3 dB crossover points to achieve
a flat power response but the amplitude response will have approximately 3
dB of ripple.
Source : Xover Pro Harris Technologies
Acknowledgements: I would like to thank Shane Rich (Technical Director
of RBH Sound, Inc) for helping with the compilation of this information to
serve as a tool in forthcoming technical articles and reviews of
loudspeakers.
Series vs Parallel Networks - First Order
Comparison
https://www.audioholics.com/loudspeaker-design/series-vs-parallel-networks-first-
order-comparison

by Gene DellaSala , Rod Elliot — August 29, 2004

Despite many of the myths that surround series networks and their
acclaimed superiority over conventional parallel networks for loudspeaker
design, both networks can be designed with identical transfer functions if
the load impedance remains constant. Most of the claims regarding series
networks are either grossly overstated or blatantly wrong and may cause
deleterious effects on system performance. As with all aspects of design,
there are compromises that must be made, and it is impossible to make an
informed decision if you are unaware of the facts.
This article is intended to show that there are no greatly enhanced features
in a series or parallel network - if properly designed their performance is
essentially identical in terms of response, phase and (by extension)
transient response. It is unwise to claim that one type of network is superior
to the other, when simple logic dictates that if amplitude and phase
response are the same, then all of the filter's other characteristics are also
the same.
There are other factors than just the response, and this is where the
differences between the network topologies exist. Each has good and bad
points that must be considered.
Schematic Representation of an Equivalent First Order Series and
Parallel Network

Illustrated above are equivalent series parallel first order crossovers with
1kHz crossover points for a fixed load. Note, resistive loads were used in
order to minimize analysis variables. Complex load impedances, typical of
a loudspeaker, will be covered in latter parts of this article.
Input Impedance Comparison between First Order Series & Parallel
Networks
Summed Output Impedances of First Order Series and Parallel
Networks
Summary
As we can clearly see, both first order series and parallel networks have
identical input impedances and very similar summed output impedances.
The Output impedance of the parallel network has a bump at the crossover
frequency of a mere 0.8mdB which can mostly be considered irrelevant.
Frequency Response Comparison (Constant Woofer and Tweeter
Loads)

Summary
As we can see both series and parallel networks have identical frequency
response and 1kHz 3dB points for the HP and LP filters when the loads are
held constant.
Frequency Response of Parallel Network (Variable Woofer Load;
Constant Tweeter Load)
Summary
By varying the woofer impedance ± 2 ohm from the nominal (8 ohms), we
observe the nominal 1 kHz -3dB point of the LPF change from 759 Hz to
1.25 kHz, respectively. Since this is a parallel network, the changing woofer
impedance has no affect on the tweeters crossover frequency as the HPF
is electrically isolated from the woofers load impedance. However, for the
case where Rpwoof = 6 ohms, we see the overlapping bandwidth between
the HPF and LPF decrease while we see it increase when Rpwoof = 10
ohms.
Frequency Response of Series Network (Variable Woofer Load;
Constant Tweeter Load)
Summary
For the series case, when varying the woofer impedance ± 2ohms from
nominal (8 ohms), we see a slightly less 3dB point variance in the LP filter
response, with a convergence at about ½ a decade past its corner
frequency. We now also have variance of the HP filter response. This in
turn has less affect on overlapping bandwidth changes between the two
drivers since they somewhat track each other (an advantage to the series
crossover).

Summary of Series vs Parallel Networks

In the parallel network, only the driver whose characteristics have changed
is affected. In the series network, however, the change is somewhat
complimentary.

Series vs Parallel Networks - First Order

Comparison - page 2
Phase Response (deg) Comparison between First Order Series &
Parallel Networks
Summary
As we can see, both series and parallel implementations of a first order
network exhibit identical phase response with 90 degrees phase difference
between LPF's and HPF's and a 90 degrees phase change between start
and stop band roll off of each filter.
First Order Series and Parallel Networks Comparison with Reactive
Woofer Load
For the next phase of our comparison between first order series and
parallel networks, we expand our models to include a reactive load such as
those typically found in a woofer as illustrated below.
Frequency Response Comparison of First Order Series & Parallel
Networks with Reactive Loading

Summary
We can see that a reactive woofer load has had a and minor impact on the
HPF 3dB point, as seen by the shift to 1.08kHz. More importantly, we note
the profound affect the reactive woofer load had on the LPF of the parallel
network as evident by the 3dB point shift from 1kHz nominal to 836 Hz, and
its degenerative stop band attenuation rolloff. The series network also
experiences 3dB point shift to a much lesser degree, with no ill effects on
stop band rolloff. Note, Zobel compensation (series R & C) in shunt with the
LPF of the parallel network is mandatory to restore comparable filter
characteristics and is also recommended for the series network to restore
optimal crossover points.
To figure out the correct R and C values for the Zobel network, the
following equations are usually used:
Rz = 1.25 * Rs where Rs is the series resistance of the driver
Cz = 1*10^6 / (2*pi*Rs*fd) where fd is the frequency where the
impedance of the driver doubles
However, by tweaking these values using trial and error, a more damped
response was achieved.
Calculated values: Rz = 10 ohms , Cz = 2.97uF

Tweaked Zobel Network for Optimal performance

Frequency Response Comparison of First Order Series & Parallel

Networks with Reactive Loading (Zobel Compensated)
Summary
As evident in the graph above, both Series and Parallel networks exhibit
identical frequency responses under reactive loading when Zobel
compensated.
Summed Output Impedances of First Order Series and Parallel
Networks with Reactive Loading (Zobel Compenstated)
Summary
The series network exhibits a perfectly flat summed output impedance
while the parallel network has minor variations, which are mostly
inconsequential given their magnitude (44mdB).

Driver Induced EMF

By changing the driver impedances, two things happen. The filter Q
changes, and the reflected change affects the behavior of the other filter
section. Although the individual response, Q and phase varies, the net
result is that the effective crossover frequency is changed, but nothing
more. This is a remarkable property, and the series first order is the only
crossover filter circuit that has this ability.
Remarkable though it may be, it is still advisable to design the series
network correctly, and maintain everything as close as possible to the
design values. Should the woofer impedance increase (with voice coil
temperature, for example), the crossover frequency will move upwards,
thus providing a small measure of added protection for the tweeter at
sustained high power levels.
However, all is not completely rosy. Everything in electronics is a
compromise, and the selection of a crossover is no different. There is one
final test that needs to be applied, and that is to examine the amount of
woofer back EMF that reaches the tweeter. This is an area where the
series network is inferior to the parallel.
 Series, Woofer Back EMF Attenuation
Summary
With a parallel network, only the amplifier's output impedance plus the
impedance of the cable allows any cross coupling between high and low
pass sections. With a zero ohm source, attenuation is infinite, and is not
shown above.
A series network relies solely on the isolation of the crossover filters, and
as a result, the back EMF from the woofer is not attenuated as well. This
may not be a major problem, since the attenuation of back EMF is the
same as for amplifier power (actually, it is 3dB better), and the latter is at a
far greater amplitude. It is a consideration nevertheless, so be aware that it
may increase tweeter intermodulation.
Transient Response Comparison between First Order Series &
Parallel Networks
Although not illustrated to save space, both first order series and parallel
networks exhibit identical transient response under a fixed resistive load.
However, when the load becomes complex, such as a real world
loudspeaker load, the results are quite different as illustrated below.
Summary
Injecting a 1kHz square wave into each network loaded with a reactive load
on the woofer, and looking at their summed response, we see the series
network electrically passes the signal unadulterated while the parallel
network exhibits overshoot on the rising and falling edges of the square
wave. However, again by simply applying a Zobel network (series R and C)
in shunt with the woofer load, we see the compensated parallel network
can now pass the square wave just like the series network.
Conclusions
These simulations have revealed that first order series and parallel
networks can be designed to exhibit very similar transfer functions as
evident by their similar input impedance, frequency response, transient
analysis, summed output impedance, and phase response. However,
under a reactive load such as a loudspeaker, both parallel and series
networks must be Zobel compensated to restore equivalent filter responses
to their original responses during purely resistive loading. The series
network is probably a better choice than parallel for a number of reasons. It
retains a flat response even when the driver characteristics change, and is
to an extent "self correcting". Implementation is no more difficult than for an
equivalent parallel network, and the same component values are used. On
the negative side, woofer back EMF suppression is significantly worse than
with a parallel network - it is up to the designer to determine if this is likely
to cause a problem.
Finally, it must be remembered that any first order network dictates that the
drivers will have significant power applied at frequencies where their
performance will be rapidly deteriorating, however for a system that will
never be operated at high power, the performance can be very satisfying.
The Loudspeaker Crossover Part II: The Brains of
your System
https://www.audioholics.com/loudspeaker-design/the-
loudspeaker-crossover
by Paul Apollonio — September 30, 2009

Contributors: Gene DellaSala

Madisound Kit
Do Better Quality Parts Really Matter?
In part one of "Crossover, Brain of your speaker system", we introduced the
concepts of inductance, capacitance, and resistance. We then examined
how these three basic passive elements relate and combine to create
frequency selective networks called High pass and Low pass sections, the
building blocks of the crossover network. We also considered in part one,
the effect of real loudspeaker impedance, and how, unlike a resistor, its
amplitude and phase vary with frequency to complicate and frustrate the
function of constant resistance type crossover networks. These real
loudspeaker impedance variations result in frequency and phase
responses which end up being very different than what our textbook
equations would have us expect, because they assume a speaker behaves
like a simple resistor. We also made the assumption that the parts used in
our crossover networks were theoretically perfect and without flaws. In part
two, we will discuss how in the real world, capacitors, inductors and
resistors exhibit behavior which is neither ideal nor perfect. We will
determine if better quality parts truly yields better performance.

Some of my more recent efforts

Real world parts, the kind you will actually find in your own crossovers,
suffer from many flaws. In part two, we will discuss and illustrate the
effects of some of these. We will also examine how simple mistakes, like
the physical orientation and location of inductors on the crossover board
can result in non-ideal behavior like cross-talk. This article will allow the
reader to gain some insight into the kinds of mistakes made by amateur
and professional crossover designers alike, and allow us to recognize
compromises in crossovers by simply looking at the networks. We also
hope to gain some understanding into flaws which are not quite so easy to
see with the naked eye. While this article is not going to be an exhaustive
study of crossover component parts, it will touch on most of the major flaws
present in the three basic components used in all real world crossovers,
resistors, capacitors, and inductors.

I am hopeful this light shed on crossover networks will make you all better
and somewhat more cynical consumers, ones who understand the
importance of the passive crossover parts used in their speaker system.
Reading some of the more ardent audiophile press, one can be left with the
opinion that there is all sort of magic going on in this network. In fact the
enemies of these passive components are basically the same as the
enemies of all electronic parts; hysteresis, loss, tolerance, insufficient
power handling capacity, insufficient space, and compromises made on
behalf of cost.

Resistors & Tolerances

Let’s start by considering the simplest of the three electrical components
used in our crossover, the resistor. It will, in combination with inductors
and capacitors create time constants used in frequency selective circuits,
although by itself the resistor does nothing other than to consume power.
In a crossover network, resistors are usually used in combination with other
components to control either impedance magnitudes or the relative levels
between different drivers in a system. Resistors are most often used in
"padding" a tweeter which is more efficient than the woofer, so the overall
system frequency response will be flat. The resistor, in series or parallel
with capacitors and/or inductors, is often used as part of a Zobel or
impedance compensation network.

A good Meter is the best way to "Trust, but verify"

Of all flaws with which we must deal, the simplest to understand is
tolerance; the allowable variation of the components value, whether that
component is a resistor, inductor, or capacitor. No surprises here,
everyone can understand how a part with a small 1% tolerance will lead to
a more uniform and reliable frequency or amplitude response performance
than a part with a 10% tolerance. The tolerance issue, while seeming
obvious, becomes more critical as we increase the order of the network.
Remember, a first order network has one part with tolerance, while a third
order network is going to have 3 which vary with tolerance. It is for this
reason that the higher order the network, the greater the need for a tight
component tolerance. Said another way, for a given amount of allowable
variation in response, (plus or minus 1 db for example), a second order
network requires tighter tolerances from its components than does a first
order network, and a third order network requires tighter tolerances than a
second order network. As we increase the complexity (order)of the
network, the sensitivity of the network to component tolerance increases.
So, as we increase the network order, not only do we add additional parts,
for a given crossover frequency, we require both larger size (value)
components and tighter tolerance in those components in order to keep the
frequency response window tolerance the same as the simpler network.
This is often a hidden and un-calculated cost in using higher order
networks. This exponential rise in part size and cost should explain why
crossover networks are almost never found in complexity above fourth
order.

Resistors normally deviate from their design values within a window of

anywhere from 0.1% to 1%. If you buy a 5% 10 ohm resistor from an
electronic store, you might go back there complaining you measured it with
your meter and found is was only 9.5 ohms, but if you get a refund it is
because they are hoping you don't return to buy more stuff. You will find
neither the highest or lowest tolerance parts in most crossover networks, as
the typical tolerance specification is either 10% or 5 %. The letters (K) and
(J) on the part will indicate if it is 10% or 5% respectively. The effect of this
variation is one of magnitude and is important to hold close enough so that
there is not much variation from one speaker system to the next. Lets
consider an example.

We have an 8 ohm tweeter which is 6 db hotter than the woofer in the

system. If we put an 8 ohm resistor in series with the tweeter, the
combination of the 8 ohm series resistor and the 8 ohm tweeter presents
16 ohms as a load to the amplifier. Since power = V2 /R and since we
have doubled R, we have halved the power the entire network (resistor plus
speaker) consumes from the amplifier. The amp is delivering 1/2 the
current to the loudspeaker load. Now half of the power that is delivered is
consumed in the series resistor, the 8 ohm resistor in series with the
tweeter. So, we have cut the power in half twice, and therefore get (- 3 db)
+ (- 3 db) = ( 6 db) attenuation. Let's say we pick a 10% tolerance for our
resistor. This generally defines a 20% allowed variance, since the
specification is +/- 10%. We are making a stereo pair, and the two
resistors we use are 8.8 ohms, and 7.2 ohms respectively, both within our
10% tolerance window. In the first case, (8.8 ohms in series) we attenuate
6.44 db, and in the second case (7.2 ohms) we attenuate the signal 5.57
db. This means we have a mismatch between our pair of 0.87 db, and a
definitely audible difference. This same magnitude variation (tolerance)
when the part is used in conjunction with a capacitance or inductance will
also cause a shift in the frequency corner of the network.

Another very well documented issue with resistors is inductance. While

high impedance, small wattage resistors are most often made from a metal
film, higher wattage parts of low impedances (the kind most likely to be
used in crossover networks) are often wire-wound parts. (If you have never
noticed before, the symbol for the resistor is a bunch of wire scrunched
up). Some old wire-wound types have a inductance high enough to cause
issues in a crossover at very high frequencies. You may often find wire-
wound resistors being referred to as "non inductive" to let the buyer or
engineer know these parts have eliminated this potential flaw. Modern day
parts are often wound with a serpentine pattern so the windings have self
canceling inductance without having an effect on their intrinsic resistance.

The largest real world problem you will run into with resistors, is a universal
problem for all components, heat sensitivity. As we read the specifications
for any component, we must bear in mind that these specifications are only
met within a certain allowable window of temperature.
 An American Made Resistor Bank - 3 times 300 Watts

Temperature dependence of resistance

The electrical resistance of a metal is approximately proportional to its
temperature over a limited range. Resistivity of materials is usually
specified at normal room temperature, 20 degrees C (68 degrees F). If one
knows the resistivity of the material at room temperature, and the rate at
which it changes, we can calculate the resistivity at other temperatures with
the following formula:

Where:
R = Resistance at temperature T
Ro = Resistance at temperature To
To = Temperature at Reference T (usually 20 degrees C)
alpha (the Greek letter with the bracket and outside parenthesis) =
Percentage change in Resistivity per unit temperature
Let’s work one example. Suppose we put an 8 ohm resistor into a
crossover network, and use it to drop
the sensitivity of the tweeter so it will match the woofer. Lets say we are
driving the speaker system pretty hard, and we are heating the resistors so
that they rise to be 200 degrees F, (a not uncommon operating temperature
for a resistor in a crossover network). 200 degrees F is equal to 93.33
degrees Celsius. (Celsius and Centigrade are the same). If the resistor is
wound from Copper wire, the temperature coefficient would be (3.9 * 10-3))
/ deg C. Since the resistor was 8 ohms at 20 degrees C, and has now
heated up to 93 degrees C, the new value of resistance would be:
8 * [(3.9 * 10-3 / oC(93.33C - 20C) + 1] = 1.286 * 8 = 10.288 Ohms
We can plainly see this increase of 25% is more significant than the
component tolerance of 10%. Suppose we do not use copper in the wire-
wound resistor. If we use a material called Nichrome, which is basically an
alloy of nickel and chromium, our change in resistance will be considerably
less. Nichrome has a low resistance variation (alpha) with temperature,
alpha = 0.4 e-3/ (deg C) Using this material to make the resistor gets us an
eventual resistance of 8.23 ohms at the elevated temperature of 93
degrees C. This is an increase of only 3%. This is no doubt, a small
increase compared to the change of resistance in the voice coil which is
either copper, aluminum, or a combination of those two.
As the power dissipated in a resistor increases, so does its temperature. As
the temperature increases, so does the parts resistance (as illustrated
above). As the resistor is heated, its ability to absorb power is
compromised, and its value in the circuit is not as designed. That resistors
get hot, and can burn out is a well known phenomena. What is not as well
known is that running high power into resistors at or near their limits brings
with it audible effects on your music.
Since the resistor manufacturer has no control over how the part is
mounted to the printed circuit board (PCB), or its orientation, proximity to
other parts on the PCB, they cannot predict with accuracy at what point the
power through the circuit will cause the resistor to go outside of its
allowable range. Take a look at the resistor in the photographs below:
The 20 ohm resistor shown above (white rectangular part) is mounted
directly to the printed circuit board (PCB) which is an effective insulator of
heat. The path for the heat generated inside the resistor to escape has lost
1 of 4 large sides. Now lets take a look at another PCB which has
eliminated this issue.

On this PCB we see the terminals of the resistor are designed to stand the
part off the board. This has the disadvantage of making the part taller, but
the advantage of creating space between the hot resistor (whose only job is
to dissipate power) and the PCB (whose only job is to connect the different
components without routing the heat from one part to another). With the
resistor mounted off the PCB, the hot air can circulate around the part more
easily, dissipating the resistors power more efficiently.

Suppose both of these parts shown above are 20 watt resistors. Which
one burns out first? Now suppose one of the boards is mounted to the
bottom of the cabinet so the heat from the resistor can rise into the entire
cabinet volume, and escape through a nearby port. Suppose the other
PCB is mounted upside down on the top of the cabinet, inside a sealed
speaker box, and under the cabinet stuffing which is fiberglass? Although
both parts are 20 watt resistors, the one in the heat containing environment
is going to burn out faster than the one with good ventilation when it needs
to handle all 20 watts. Power ratings on resistors are NOT independent of
the way in which they are mounted to a PCB.

The Loudspeaker Crossover Part II: Capacitor

& Inductor Issues
Capacitors imperfections
Now let's discuss the capacitor. (For those of you who want to read more
detail on the performance issues with real world capacitors, there is an
excellent treatment of the subject in Wikipedia
@ http://en.wikipedia.org/wiki/Types_of_capacitor).

There are many non ideal behaviors that capacitors show, but perhaps the
worst is dielectric absorption. This is a fancy name for capacitor memory.
The dielectric is the material which insulates the two electrically conductive
rolled plates in contact with the capacitor terminals. This dielectric material
is how we specify the capacitor type, e.g., mylar, polypropylene,
electrolytic, ceramic, etc. With Dielectric Absorption (DA) even after the
initial voltage gets removed from the capacitor terminals by shorting out the
capacitor, the capacitor returns to its prior state of charge without any
signal being supplied. The energy which returns comes from the electrical
insulating film, or dielectric, in which it was stored. Problem is, the
dielectric is not supposed to store a charge. Use of polystyrene,
polypropylene and Teflon as dielectrics will keep this effect to a minimum.
This effect (DA) is also known as dielectric absorption hysteresis.
Hysteresis is the storage of energy in a medium. It is what makes our
permanent magnets work, causes speakers to drift off center position, and
what make a magnetic cored inductor distort even without saturating.
 A 27 ufd polypropylene and two 22 ufd non polar electrolytic capacitors

In the photograph above are three capacitors. The two smallest are both
22 ufd, 5% tolerance and 100 Volts ratings. The blue one in the center is
about twice as large (volume not length) as the small black one. You can
see the relative size of the 27 ufd Solen 400 Volt Polypropylene capacitor
at the top of the photo. Which one of these three do you think will handle
the most power? Which will handle the least? If you want a good seat of
the pants approach for determining capacitor quality, here is a clue. For a
given capacitance, bigger is almost always better.

The Q or quality factor (Also known as Dissipation Factor or DF) of a

capacitor is a measure of its losses of signal due to leakage and Equivalent
Series Resistance (ESR for short). The HIGHER the Q of the capacitor,
the less losses it will impose on the signal. Secondary to the signal loss, is
that high DF capacitors will create more heat while they are passing the
electrical signal. An ideal capacitor will not dissipate or absorb any
electrical energy. A real capacitor will, and the resulting heat it generates,
will likely mean the capacitors ESR will increase, meaning it will be easier
to generate more heat and more loss. This is a little bit like the situation
with a transistor getting hot, and increasing its native gain, which makes it
get even hotter. With transistors this is known as thermal runaway. In
short, the effect from the heat exacerbates itself, like positive feedback in a
microphone feeding a loudspeaker.
 Some of us will always prefer to "Roll our Own"
In a loudspeaker, the heat which increases the voice coil resistance means
that the voice coil draws less power from the voltage amp as the voice coil
gets hotter. (Heat in this instance acts like a compressor). With a capacitor
getting hot, it starts to dissipate even more power, not less. While heat
tends to make the voice coil self limiting, it makes semiconductors and
capacitors likely to run away with an accelerating problem. While this effect
(DF) is measurable in film dielectric capacitors, like mylar and
polypropylene, it is not likely to be large enough to be audible. Dissipation
Factor (DF) IS a serious issue with electrolytic capacitors. My very first
consulting job in 1983 was fixing a crossover for a DJ in New Hampshire.
When I opened up his speaker, I found his electrolytic capacitors had
gotten SO HOT they literally blew up. All that was left of the electrolytic
capacitors used was the plastic covering over the metal capacitor cases.
The parts failed and exploded! (In the DJ business, it is often about how
loud you are more than how good you sound).

When you are comparing electrolytic capacitors, and you notice some of
these parts are much smaller than others which seem to bear the same
specifications, it means you have left out one very important specification,
that of DF (Dissipation Factor). This is often a good measure of the life of
the part, along with its Voltage and temperature rating. This imperfection is
THE compromise made with capacitors in order to keep both the size and
the cost low. The initial cost that is. If it blows up in your crossover, it may
not have been worth saving the initial $2 you saved at retail. Since this is
something hidden inside the speaker box, the manufacturer often believes
the buyer ignorant of this, so they elect to pocket the savings, hoping the
cheap parts used will outlive the life of the warranty.

For a sense of scale, a flip phone, 18 ufd electrolytic (green) and 18 ufd Mylar
(white)
Like the resistor, the capacitor can also be inductive at a high enough
frequency. This ill effect is known as ESL (Equivalent Series Inductance).
(We use the letter L to designate an inductor in a circuit diagram). In case
you have wondered why Choke, Inductor and Coil is represented in
electronic circuits by the letter "L" it is not solely because "C" was already
taken by the capacitor.
For some background on this go to: http://en.wikipedia.org/wiki/Lenz's_law

In the world of the crossover, ESL is not likely an audible phenomenon,

given the high inherent inductance of most loudspeakers. Audiophiles will
likely want to argue with me on this point, so let me deflect this by saying
when I discuss audibility I do not include the most sensitive audiophiles in
the discussion. Some audiophiles claim hearing abilities rivaling that of
bats and sonar equipment. That said, if measured you will find the typical
tweeter has 100 times or more inductance than the most inductive
capacitor likely to be in series with it. If one is using a very large capacitor,
such as a 200 ufd electrolytic, then it may be helpful to bypass it (parallel)
with a 0.1 - 0.5 ufd film capacitor to effectively eliminate the ESL. Unlike
DF or DA, ESL can be eliminated by a lower value capacitor being used in
parallel.

Inductors - The Most expensive & Most problematic

Crossover Element
Let's consider the subject of chokes. Once again I am inclined to thank
collectively the authors of Wikipedia for an excellent treatment of
Inductance, and its close relatives, EMF, magnetic flux, and magnetism.
There is enough physical science behind an inductor to write not just one
article but several large textbooks, and many scientific journal articles.
For an excellent general treatment and links enough to keep the
scientifically curious busy for days: http://en.wikipedia.org/wiki/Inductor This
page shows no less than (6) different formula for calculating inductance of
an air core choke. (So try not to go nuts....)

The primary figure of merit for a crossover coil is its Quality factor (or Q) for
short. The quality factor of a coil is determined by:

Where
W(Omega) = 2pi*Frequency
L = Inductance in Henries
R = Resistance in Ohms

Ignoring all else but the resistance of the choke we can see as this figure
goes to zero, the Quality factor (Q) approaches infinity. We can also see
from this, that a choke of 1 millihenry inductance having a resistance of 0.5
ohms, has the same quality factor of a choke of 2 millihenries inductance
having a resistance of 1.0 ohms. This should make it obvious to those who
understand that as the crossover frequency goes lower, the size of, cost of,
and importance of the quality factor (Q) of the inductor becomes more and
more important. Two ohms in series with a four ohm woofer, is still a
bigger problem than one ohms will be with regard to losses. This is one
important reason why manufacturers sometimes shy away from systems
having very low crossover points. Even when it is better for performance, it
is often so costly to do it right, that a higher crossover point gets chosen for
budgetary reasons.

2.0 mH air core Jantzen from Parts Express = $11.00 each = 0.80 ohms
DCR
2.0 mH iron core Erse from Parts Express = $ 7.49 each = 0.26 ohms DCR

On the face of it, the magnetic core choke seems to be a better deal.

A laminated Steel Core Choke

Inductors are the biggest attention getters in the passive crossover and for
good reason. They are usually the largest and most expensive parts used,
and they are present in sizes and masses that have few real analogs in the
electronic world except for transformers in power supplies or charging coils
used in magnet chargers. Inductors are subject to losses much more so
than capacitors. While one can buy a high quality mylar or polypropylene
capacitor today for a few dollars, a very high quality large value air core
choke is still many times more expensive than that high quality capacitor.

Like electronic components on a PCB, sometimes the inductors are not just
inductors, sometimes they are transformers as well. (Just like sometimes
loudspeaker voice coils are inductors when they were trying to be simple
resistors). For those of you familiar with the construction of transformers,
you will remember they are two separate electrical windings both put on a
common magnetic core which is intended to link them together by
electromagnetic field coupling, also sometimes called inductive coupling.
Let’s take a look at the following circuit:
 An inexpensive LCR meter will Define the deficiencies of your parts
Let’s say we used two chokes, both 3 mH and both randomly placed on the
PCB. Here in my lab on the prototyping board, they look like this:

If we curve the HP and LP function of this network, we find the following

frequency responses:
 Frequency Response of System with Separated Chokes
We have our expected High Pass and Low Pass function, but we might
notice there is more loss in the LP than the HP filter. Of course, we say,
the inductor is in series with the resistor load in the LP but in parallel on the
HP circuit. That results in a significant series resistance being placed
between the amplifier and the load, in this case a 5.6 ohm resistor. Now,
lets’ move the inductors so that they are physically on top of one another,
(see photo below) stacked so they share the same diameter and run the
test again.
 HP and LP Chokes Stacked so they inductively couple

Change in HP response due to stacking the HP and LP chokes

Looking at the green and red HP curves, we might notice two things here.
First, we have an increasingly divergent stop-band on the HP filter section,
and second, we have less rejection (below 400 Hz). Why? Because the
close proximity of the two inductors means they are inductively coupled so
we are getting crosstalk. Because the signal is coupled between the two
inductors, the series LP choke, and the parallel HP choke, we get a
lessening of the out of band signal rejection of the HP section. Two
conductors are inductively coupled when they are configured such that
change in current through one wire induces a voltage across the other
wire. By placing one inductor on top of the other, I am inducing currents to
flow in the HP choke because of the current flowing through the LP choke.
It is for this reason that you will often find crossover chokes places at right
angles to one another as pictured below when laid out on a PCB. This
orientation eliminates most of the inductive coupling that would occur
inadvertently.

A 5.6 mH magnetic core and a 3.6 mH air-core choke side by side

In the photo above we have two different kinds of Inductors. Magnetic and
Air core. The magnetic core (left) is 5.6 mH and has a DCR of 0.28 ohms.
It is wound from 15 AWG copper wire. The 3.0 mH air core choke, wound
from 16 AWG copper wire has a DCR of 0.75 ohms, or three times the
resistance and about 64% of the inductance. This means the loss in the Air
core is about 4.66 times as great as the iron core choke. In order to both
minimize the inductors resistance and cost, vendors have wind chokes on
magnetic cores. There are different types of magnetic core materials used,
depending on the frequency range the choke is to be used in. The beauty
of using a magnetic core, is the huge savings in copper wire for a given
inductance. (If I say laminated steel or ferrite or powdered iron, someone
will feel compelled to explain to me the difference, and why one is superior
to the other, so I shall simply differentiate inductors by saying magnetic or
air core.) Because the magnetic core increases the permeability of the
inductor, confining the magnetic field more closely and with greater
intensity; you reach a given inductance with fewer turns of wire, allowing
you to make a part which, for a given size, is going to have a much lower
DCR than its air core counterpart. This means you will lose less power in
your series LP chokes. It is this reason (plus cost) that we use magnetic
cores in inductors.

On the face of it, the magnetic core choke seems to be a better deal all the
way around. So, why then with all these advantages would anybody use
anything but Magnetic core chokes? There are two reasons. Saturation,
(an effect of running out of permeability by the ferrous core) and hysteresis,
the storage of energy, which is present with all magnetic materials. Which
problem is worse for the crossover designer; loss of signal due to higher
resistance, or the nonlinearity of the magnetic core choke? This depends
in large part upon the magnitude of the difference. This is why small value
chokes are almost always air core, while very large value chokes are
almost always magnetic core. I dislike taking sides in such debates when
both approaches have distinct advantages. Personally, I can easily hear
the distortion made by iron and steel core chokes, and would rather live
with the higher cost and higher (but relatively constant) resistive losses in
the air core choke than hear the distortion in a magnetic core choke.
Unfortunately, at very high inductance values, air core chokes quickly
become impractical because of their weight, size and cost.

Let's take a look at the following circuit, a simple first order low pass filter
attached to some lab gear.

The Sine wave symbol indicates the power amp, in this case a bridged
"3000 watt" power amplifier, and the V is the voltmeter, a Fluke model 45.
The 5.6 mH choke is shown here as air core, but both air and magnetic
core chokes were used for this experiment. Clio is a popular
hardware/software system designed specifically for measuring audio. The
input to the Clio box is put across a small 1 ohm resistor, placed between a
pair of 40 ohm resistors so as not to load the bridged power amp, or cause
an input voltage overload to the very expensive Clio box, which might
cause the Clio to smoke, and me to lose my day job.
The point of this circuit is to see if I could measure distortion in the resistor
load, and identify that distortion as originating from the changing
impedance of the magnetic core choke. The theory is, if the core
permeance is changing as it nears saturation, then the impedance of the
choke will also change. Since an AC waveform varies its magnitude with
time, the amount of saturation that occurs will be greater at the peak than
elsewhere in the waveform. If that is true, this will generate a distortion in
the voltage waveform and therefore in the resistor load, as the current
flowing will not be faithful to the original input voltage. As the AC waveform
varies, the current should follow it. Since the AC waveform varies its
amplitude with time, even with a relatively invariant AC signal like a pure
tone, how bad would it be if the signal had a crest factor of 20 db, not
uncommon in music, instead of the relatively modest 3 db crest factor of a
pure tone? Through the entirety of the AC waveform, the current will only
be proportional to the voltage if the impedance does not vary (and like 99%
of all amps, the power amplifier is a voltage amp, not a current amplifier).

The Loudspeaker Crossover Part II:

Measurements & Conclusions
The screen-shot below, shows the Clio input shorted to its output so we
have a reference of the signal used for the magnetic vs air core choke test.
The noise floor of the Clio is below the -90db level at the bottom of the top screen
The screen-shot above is a dual view, with the top screen a spectrum
analysis, amplitude vs frequency; and the bottom screen an oscillogram, or
the voltage vs time view. Both views are two ways of examining the exact
same data. If our choke and amplifier are perfect, the measurements will
look exactly like the view above. Since our 4 ohm resistor is able to
dissipate 900 watts RMS, and the bridged amplifier is rated to deliver 3000
watts into this 4 ohm load, we will limit the power into the load bank to 900
watts. To further limit the resistor heating, we will use a sine wave which is
on for one second and off for two. With this arrangement we can take a
stable measurement while reducing the long term (3-second avg.) power to
only 1/3rd of the signal-on condition.

Now if my theory is correct, we do not go from "the choke is perfect", to "the

choke is saturated" like a light gets switched on or off. If the magnetic core
saturates, this effect should happen gradually as the alignable magnetic
domains available become more rare, as the magnetic material runs out of
them (this is saturation). We should be able to see this effect by viewing
either the time or spectral data. When we examine the sine wave above, it
reminds us that the AC signals are not constant, but vary with time. That
115 volts we plug our computers and televisions into goes from zero to 162
volts, to zero to minus 162 volts 60 times every second. If we attempt to
drive the 4 ohm load with a 900 watts RMS sine, we will require a current of
15 amps RMS and 21.2 amps peak! Although the manufacturer does not
list this parts saturation point on the data sheet, we should be able to find it
for ourselves. Unlike a transistor amp without compression run into
clipping, the shape of a magnetic cores hysteresis curve limits gradually as
it approaches it saturation point, so there is no exact point of onset of
saturation.

At the frequency of 120 Hz, the choke calculates out as a little more than 4
ohms. Since our load is also 4 ohms, and since music tends to have a lot
of power occurring in this frequency range, we shall use 120 Hz as the test
frequency. Under this condition, there will be approximately as much
voltage across the inductance as the resistor. Choosing a much lower
frequency will tend to take the choke out of the equation, and choosing a
much higher frequency would seriously limit the amount of current flowing
through the circuit or put very high demands on the voltage required of the
amplifier. The signal is raised until the voltage across the 4 ohm resistor is
just above 60 volts (61.6 rms), and the power absorbed by the load is 900
watts. This conditions requires about 95 volts from the amplifier. (In the
perfect world, 87 V would do, but nothing is perfect).
The signal is measured and shown below:

High level of 3rd Harmonic Distortion due to the Magnetic core of the 5.6 mH
inductor
With the Clio Meter maximized, the THD measures 6.9% and IMD
measures 1.2%. As can be seen from the measurement above, the
distortion is symmetrical, as expected, so the dominant distortion is 3rd
harmonic, with 2nd harmonic low. The waveform is starting to take on the
form of a triangle wave as a result of the changing impedance of the
magnetic core choke.

Well, perhaps this distortion is a result of the amplifier being driven so

high? Let's test that. Since I do not have an air-core choke this large, I
construct one out of two 2.4 mH cokes, stacked one on top of the other. It
ends up being 5.7 mH. (The reason two 2.4 mH chokes add up to more
than 4.8 mH, is due to mutual, not self inductance). This is why you do not
have a linear relationship between turns and inductance in a choke. (And
why there are (6) different equations to solve for inductance on the
Wikipedia page referred to earlier). If both of these chokes were wired in
series and placed far apart, instead of stacked one on the other, they would
in fact be 4.8 mH. Now, at 1.7 ohms, the DCR of our air core choke is
considerably higher than the magnetic core choke which measures only
0.28 ohms. This increased resistance requires us to deliver 115 volts to
the series circuit to achieve the same 60 volts (62 V) across the 4 ohm test
resistor.
The measured response is below:

The distortion Products from the combination of Amplifier and 5.7 mH Air Core
Choke
The distortion measured 0.952% THD. As can be seen above, the
spectrum across the resistor looks quite different when driven through the
air-core choke than it does when driven through the magnetic core choke.
The oscillogram too looks free of distortion as well. So, let's ask ourselves,
is this better because the DC resistance is higher by 1.5 ohms? Maybe the
issue was current limitation in the amplifier. This amp is rated to deliver
3000 watts into 4 ohms, which requires 27.3 amps RMS. The 62 volts
across 4 ohms only requires 15.5 amps RMS, just a little more than half
what the amplifier should have available. Maybe the 45 degree phase
angle between voltage and current at 120 Hz is straining the amp? What is
the limit of this amplifier? Let's see. Without a series choke, with the 4 ohm
load alone attached to the amplifier, I drive the amp to clipping, and
measure 88 volts RMS across the resistor terminals.
The following measurement data was taken.

Amplifier in Clipping
As can be seen here, the sine wave is clipped, as the top of the waveform
is flattened. Since we were able to deliver 115 volts unclipped to the higher
load impedance, we have to assume this amp has run out of current. At 88
volts RMS, a 4 ohm load will draw 22 amps RMS. Close inspection of the
amplifier specification claims 3000 watts @ 1% THD when bridged into 4
ohms but only at 1 kHz. Clearly at 120 Hz, this amplifier has clipped at
1936 watts. As had been said before, trust but verify... So, lets drop the
input by 1db, and measure the signal across the resistor again.
 Amplifier 1 db below Clipping - No Choke
This time we manage to get 80.6 volts across the resistor, with very low
distortion. Clearly with a zero phase angle, this amp can deliver 20 amps
RMS and 28.3 amps peak without clipping into 4 ohms. We can be pretty
confident that the difference we have measured is the result of the
magnetic core of the choke, and not the amplifier or signal generator.
Suppose we drop the input drive to the magnetic choke to 6 db less than
the 900 watts RMS delivered to the test load. Lets say we drop the voltage
down to 28.5 volts RMS, or 203 watts RMS. What kind of distortion will we
measure? Lets see..
 6 db below clipping with Magnetic Core Choke

Even at 6 db below clipping, the Magnetic core choke results in more than 1% THD

While the waveform looks pretty clean, we see the same distinctive line up
of harmonics, with 3rd being very dominant. When metered, the distortion
at 25.5 volts RMS measures 1.3%. This is still higher than 0.952% THD
measured with 62 volts across the resistor, and 115 volts from the amplifier
using the air core choke. At this point we are well below the limitations of
the amplifier, and delivering a current of 7.125 amps RMS, less than half as
much current (1/4 the power) as was delivered through the air core choke
and yet despite our reducing the power to 203 watts RMS, we find we have
more THD in our resistor than what was present at 961 watts when driven
through the air core choke.
The use of magnetic cores in chokes is a two edged sword. While
magnetic materials in chokes will reduce size, cost, and the DCR of the
choke, it comes with a disadvantage in terms of distortion. While amplifier
manufacturers are busy reducing their distortions to numbers well below
0.01%, loudspeaker manufacturers still struggling to build systems which
can play at or below 1% distortion anywhere at or near their output limits.
To add an additional 1% distortion by using magnetic core inductors in
series with the speaker is a price I am personally not willing to pay.

So, what conclusions are to be drawn from this

article?
 Temperature is an enemy of all three components, and temperature
has real and measurable effects upon the transfer function of the
crossover network (and your music!)
 Using Nichrome as a resistor material provides a huge advantage
over copper or aluminum by virtue of its relative constancy with
temperature fluctuation
 Electrolytic Capacitors suffer hysteresis effects which can negatively
impact audio
 Bypass caps can be used to improve performance of Electrolytics by
lowering ESL
 Electrolytic Capacitors suffer from high dissipation, which cannot be
minimized without using multiple capacitors in parallel
 PCB layout is an important consideration and effects the overall
crossover power handling and performance, and is especially
important in terms of heat and inductive coupling
 Saturation is not a digital phenomenon - It happens gradually and the
magnetic core choke increases distortions at levels considerably
below those at which the cores will "Saturate"
 Magnetic Cores can reduce crossover cost, DCR and size by
increasing inductance
 Magnetic core Inductors are inherently nonlinear, even before
saturating

• Air cores chokes are preferable, but suffer from high losses and high
cost
While this is not an exhaustive study of chokes, or of every type of
magnetic core material, we have seen how the use of a non-linear material
can affect the sound of the music we hear. If we are looking to improve our
systems, and purchasing upgrade crossovers, we need not only be
informed about the specifications of the parts used in them, we also need
to consider the PCB layout as well. While perhaps the simplest electrical
circuits in use in audio today, crossovers can play a major role in the sound
we hear coming from our speakers, and our speakers are the most
important link in the audio chain in the vast majority of high performance
home theater systems. The sum of the parts and their quality truly does
matter, especially if you're a golden ear and desire the very best
performance.
Paul Apollonio is a Southern California based consulting engineer, with 25
years of experience in the audio industry designing loudspeakers and
crossovers. He may be reached at papollonio@yahoo.com and his resume
and profile viewed online at http://www.linkedin.com/in/paulapollonio.