May 1966 Electronics World
Table
of Contents
Wax nostalgic about and learn from the history of early electronics. See articles
from
Electronics World, published May 1959
 December 1971. All copyrights hereby acknowledged.

Here, just in time for Thanksgiving, is a real
cornucopia of attenuator information. If you need circuits
diagram and equations for "T," BridgedT, Ladder, Pi, BalancedH,
Balanced Ladder, Potentiometer, and Balanced
(Dual) Potentiometer type
attenuators, then you've come to the right place. A discussion
is included on attenuator selection and specification for ordering
rather than designing and building your own. The distinction
between a 'pad' and an 'attenuator' has always been vague to
me and I, like most people, use the terms interchangeably. Author
Chester Scott seems to believe a 'pad' always has a fixed value
whereas an attenuator can be either fixed or variable.
Resistive Attenuators and Pads
Circuitry, design, and selection of these devices which vary
signal levels without disturbing the impedance, frequency response,
or other system characteristics.
By Chester F. Scott
Chief Engineer
Daven Division,
McGrawEdison Co.
This article covers the circuitry and design of resistive
attenuator networks. These are used to vary signal levels without
disturbing the impedance, frequency response, or other characteristics
of the system. They are referred to by many names such as faders,
pads, mixers, masters, pots, etc., but the basic function remains
the same.
A typical unbalanced bridgedT variable resistive
a.f. attenuator.
Attenuator Classifications
Attenuators are divided into two main classifications: fixed
and variable. Fixed attenuators are used where the level of
a signal must be reduced by a known number of decibels which
will not have to be changed. Variable attenuators are used where
rapid or frequent changes in signal level are required. They
are also used when the required amount of signal reduction must
be determined at the time of system installation. Where frequent
variations in level are necessary, an attenuator which is variable
by a rotary or a straightline switch is used. When the level
can be adjusted and set with a minimum of future changes, a
type that is variable by means of soldered interconnections
is normally used.
The basic circuits for both the variable and fixed types
are the same and the formulas for resistor values are equally
applicable to both. In variable attenuators which are controlled
by a switch, the variation is generally accomplished by adding
or subtracting resistors in the series and shunt circuits. The
exceptions are the ladder and
π
types. In the ladder type, all resistors are fixed and the series
resistor tapped at correct intervals. The
π
operation is covered later. (Inexpensive
variable attenuators are also available in which the various
series and shunt elements consist of volumecontroltype potentiometers
that are ganged together to produce the proper resistance changes
as the control is operated  Ed.) Attenuators that are
variable by means of changing soldered interconnections consist
of a group of fixed pads with different attenuations. The attenuation
variation is accomplished by connecting attenuators with the
correct losses in series to provide the required total attenuation.
Electrically, attenuators are divided into two main classifications:
balanced and unbalanced. A balanced attenuator control is one
where equal resistance is inserted in each side of the circuit
and where the center point of any shunt resistance can be grounded
or left floating depending upon the other circuit characteristics.
An unbalanced control is one where the resistance is added in
only one side of the line and where the other side of the line
is "common" or grounded.
In designing the equipment where attenuators are to be used,
there are two schools of thought as to whether balanced or unbalanced
controls are more satisfactory. One group claims that balanced
circuitry is better because hum and noise due to ground loops
and improper grounding is done away with. The other group contends
that unbalanced circuitry is better because, with proper grounding,
stray capacities are overcome and better control is obtained.
This article will not attempt to deal with the merits of overall
system design. There are instances where economics will be a
determining factor in system design, and we will only indicate
that balanced controls are usually approximately twice as expensive
as their unbalanced counterpart since twice the number of resistors
and twice the switch facility are required.
An unbalanced control cannot be used where other circuitry
is balanced because, to operate correctly throughout the required
frequency range, an unbalanced control must have one side of
the line grounded. A balanced control cannot be used in an unbalanced
circuit since ground on one side of the unbalanced circuit would
affect the control operation. Onehalf of a balanced control,
i.e., one input terminal, one output terminal, and the centertap
terminal, can be employed in an unbalanced circuit but only
if the particular circuit impedance is onehalf that of the
balanced control.
The most frequently used attenuator circuits are the T, bridgedT,
ladder, and potentiometer types or their balanced counterparts,
the balancedH, balanced bridgedH, balanced ladder, and dual
or balanced potentiometers. In highfrequency applications,
the unbalanced
π
type is most generally used. The schematic diagrams for these
types are shown in Fig. 1.
Fig. 1  Various types of unbalanced and
balanced attenuators.
The following information covering the design and use of the
various styles is given for the unbalanced types only. The formulas
for determining the resistance values in a balanced control
are the same as those indicated for the unbalanced control except
that, in inserting the value for Z in the formula, the full
value is used for an unbalanced control but only onehalf the
value is used for the balanced control. Thus, in determining
the values for a 600ohm attenuator, the value for Z in the
formulas for an unbalanced control is 600 ohms, but that for
a balanced control is 300 ohms.
Typical fixed attenuator T pad for 600ohm
audio lines.
The T attenuator is generally the most stable and, within
reasonable limits of terminal impedance and range of attenuation,
can be made the most accurate through fairly wide frequency
ranges. Since it requires three variable resistors and their
associated switching, it is generally the most expensive of
the unbalanced types used at frequencies up to 1 MHz. Both the
input and output impedances remain constant, within the specified
resistor accuracy, throughout the attenuation range. The insertion
loss of this type control is zero dB when the input and output
impedances are equal.
These characteristics make the T attenuator especially useful
as controls in instruments and other applications where accuracy
and constant impedance are required. The formulas for determining
the resistance values required are shown in Fig. 2A.
Fig. 2  Circuits and design formulas for
various attenuators.
Where equal input and output impedances are involved, Z_{in}
and Z_{out} are equal and the resulting values for R1
and R2 are also equal. For a fixedloss attenuator with either
equal or unequal input and output impedance, and for a variable
loss attenuator with equal input and output impedance, the formulas
are used without further consideration. For a variable loss
attenuator with unequal input and output impedance, the resistor
values are normally determined for one impedance, and an impedancematching
network is used to match the impedance at the other end. The
resistance values for this impedancematching network are determined
as indicated later in this article.
The bridgedT network is a variation of the T network which
requires the use of only two variable resistors. It is thus
somewhat less expensive than the T but has practically equal
characteristics up to somewhat lower attenuation values. The
bridgedT has a constant input and output impedance throughout
the attenuation range and has a zero dB insertion loss when
the input and output impedances are equal. In the higher losses
above 30 to 40 dB, the value for R1 (see Fig. 2B) becomes so
high that it is more susceptible to the effect of any stray
capacitance across it.
The bridgedT attenuator is normally used only for attenuators
which are variable by means of rotary or linear switching. They
are normally used between equal input and output impedances
but, with the use of an impedancematching network at one end,
they can be used between unequal impedances with the necessary
insertion loss due to the impedance match. They are used mainly
in instruments or in communications equipment where constant
impedance and zero insertion loss is important and where the
attenuation range is not too great. The formulas for determining
resistance values are shown in Fig. 2B.
Ladder Attenuators
The ladder type attenuator is actually a variation of the
π
type and is used only in the variable loss styles. It requires
the use of only one variable resistor and is the least expensive
of all the types except the potentiometer. Ladder attenuators
are most generally used as mixer controls in public address
systems and broadcast consoles, and also as controls in audiometers.
The circuitry is such that the output impedance falls to
about twothirds of the nominal value at the low loss end and
the input impedance falls to approximately the same amount at
the high attenuation end. The curves for a 30step, 1 1/2 dB/step
ladder with a 600ohm input and output impedance are shown in
Fig. 3. Throughout the midattenuation range, where the control
is normally operated in mixing applications, both the input
and output impedance are fairly constant and practically nominal.
Fig. 3  Characteristics of a 30step ladder
attenuator.
The resistor values in a ladder attenuator are such that
they remain within practical limits throughout the attenuation
range. Thus, the high attenuation values required for audiometer
applications are obtainable in a single control. With an equal
input and output impedance, the ladder attenuator has an insertion
loss of 6 dB. This is partially due to the fact that, unlike
a T or bridgedT attenuator, all the resistors of the ladder
attenuator are in the circuit at all times and control is obtained
by connecting to various taps on the series resistor by means
of the switch. The remainder of the insertion loss is due to
the fact that a resistor, equal in value to onehalf of the
input impedance, is connected in series with the input to provide
a more constant impedance. When the input impedance of a ladder
attenuator is onehalf of the output impedance, the resistor
in series with the input is omitted and the insertion loss is
approximately 2dB.
The ladder attenuator is unique because its circuit is such
that any possible switching noise is attenuated along with the
signal. This feature is helpful in mixer applications where
the signaltonoise ratio is especially important. The formulas
for the resistor values in the ladder attenuator are shown in
Fig. 2C.
For optimum decibel and impedance accuracy, dB in the formula
is normally selected as 6 dB. For a 2 dB per step control, R2
is tapped for three equal sections and for 1 1/2 dB per step
it is tapped for four equal sections.
Potentiometers and
π
Networks
The potentiometer is a simple potential divider and is designed
to operate into a very high impedance such as the grid circuit
of a vacuumtube amplifier. It is used to control voltage in
specific decibel steps or in special voltage ratio steps. It
has the simplest construction and is therefore the least expensive
of any of the attenuators. The formulas for determining the
resistor values for this attenuator are shown in Fig. 2D.
The formulas are for potentiometers calibrated in decibel
steps. When voltage ratios are required, the ratio of R2 to
R1 is the same as the ratio of the required output voltage to
the input voltage. Again, R3 = R1  R2.
The
π
network is comparable to the T network in stability and accuracy.
It is also a network which requires three variable resistors.
It is not generally used in low and mediumfrequency variable
attenuator applications because the switching is somewhat more
complicated than that for the T network. It is more generally
used in highfrequency attenuators because it is easier to shield
and the resistor values are in a range more readily adaptable
for the lower impedances used.
In the highfrequency variable attenuators, the variation
in attenuation is accomplished by switching
π
attenuators of the correct loss in and out of the circuit. For
this type of switching, the resistors values in a
π
network remain within a range which is not overly affected by
stray capacitance. The formulas for the
π
network are shown in Fig. 2E.
Attenuation Per Step & Matching
The selection of the dB per step of any attenuator depends
upon the usage. For controls used in instruments, the dB per
step can be any value up to 10 dB, and sometimes more, depending
upon the instrument requirements. In broadcast and recording
operations, the normal maximum dB per step is 2.
In the higher quality recordings, no switching noise can
be tolerated. If the dB per step is too high, there will be
a noticeable "pop" if the control is switched between steps
during a sustained note. The lower the dB per step, the more
steps are required to obtain a comparable total attenuation
and more steps means higher cost. Thus, economics enters into
the choice, but in recording applications, steps of 1.5 dB or
less, if practical, are commonly used.
For optimum overall operation, the impedance of all portions
of a system should match. Thus, with the exception of the potentiometer,
the input impedance of the attenuator should equal the impedance
of the source and the output impedance of the attenuator should
be equal to that of the load.
When the source and load impedances are unequal, a resistive
network can be utilized to match them. This impedancematching
network is generally a T network or its counterpart, the balancedH.
Where a resistive network is used to match unequal impedances,
some loss is necessary. The amount of this loss is dependent
upon the ratio of the higher impedance to the lower impedance
being matched.
Power & Impedance Available
The normal power dissipation of most attenuators is 0.5 or
0.6 watt continuous. Such attenuators will handle up to approximately
2.5 watts peak in audio circuits where the level fluctuates
considerably. Attenuators with continuous power rating up to
20 watts and, in special cases, higher are available. Attenuators
which have a low dB loss, such as decade T types with 10 steps
of 0.1 dB per step, can be used in circuits where the wattage
is considerably above the normal rating of the control. This
is due to the fact that only a relatively small portion of the
power is dissipated in the attenuator whereas most of the power
is dissipated in the load.
In general, balancedH, T, and ladder attenuators are available
with impedances up to about 1000 or 2000 ohms. For special applications,
controls with higher impedances are obtainable but attenuation
and frequency ranges must be considered since internal capacitance
and capacitance in the external wiring will have greater effect
on the attenuator accuracy. The attenuators used in highfrequency
applications generally have an impedance of 50, 52, 73, or 75
ohms. Attenuators with 93 and 125ohm impedances can be made
but, for a specified accuracy, the frequency range is limited
by capacitance effects. Potentiometers normally are made with
impedances in the range from 1000 ohms to 1 megohm. Lower impedances
are available with a somewhat reduced attenuation range due
to resistance values becoming too low for practical manufacture.
Capacitance effect becomes a limiting factor for impedances
over 1 megohm.
Attenuation & Frequency Range
The attenuation range available in an attenuator depends
upon impedance, circuit type, and frequency range. In bridgedT
and bridged balancedH attenuators the variable series resistor
becomes exceedingly high at the higher losses and the variable
shunt resistor becomes too low for practical manufacture. Thus
the highest loss normally supplied in the bridgedT and bridged
balancedH types is approximately 40 dB. This is not a firm
figure because impedance and dB per step must be taken into
consideration.
In T and balancedH attenuators the limiting factor on the
dB range is the resistor value between steps in both the series
and the shunt arms.
At the higher losses these become too low for practical manufacture.
The approximate maximum loss normally supplied in the T and
balancedH attenuators is 60 dB but here again the actual maximum
depends upon impedance and dB per step.
In the above types the total attenuation range of a variable
attenuator may be increased by combining two or more attenuators
on the same shaft and connecting them in series. For example,
a 100dB attenuator can be made by connecting two 50dB attenuators
in series. In operation, the first group of steps varies the
first 50 dB and the second group of steps varies the second
50 dB. Up to 50 dB, the attenuation is varied in one section
while the second section remains at zero dB. Above 50 dB the
first section remains at 50 dB while the second section inserts
the remainder of the attenuation.
In ladder and balanced ladder attenuators the circuitry is
such that the resistor values do not get too high or too low
as loss is increased and therefore they are not a limiting factor
in the dB range. Thus, in these types, attenuation ranges of
somewhat over 100 dB are obtainable.
Attenuators are available covering the frequency range from
d.c. to 10 GHz. Socalled audiofrequency attenuators will operate
satisfactorily up to at least 100 kHz and in some instances,
depending upon the decibel range and impedance, much higher.
Videofrequency attenuators generally operate at frequencies
up to 10 MHz and radiofrequency attenuators cover the remainder
of the range to 10 GHz.
Attenuator accuracy is dependent upon the accuracy of the
resistors used in their construction. At audio frequencies,
the attenuation accuracy is the same as the resistor accuracy.
Thus, if resistors with a 5% accuracy are used, the attenuation
will be within 5% of nominal. Actually the attenuation accuracy
will probably be much closer than 5%.
Attenuators used in broadcasting and recording operations
normally have a 5% accuracy while those used in instruments
and in measuring applications generally have a 1% accuracy and
frequently have tolerances as close as 0.1%. The higher frequency
attenuators normally have the accuracy specified in dB through
a range of operating frequencies. Since this accuracy depends
upon the amount of loss, impedance, and frequency range, a general
statement of available accuracies is not feasible. For these
applications the manufacturer should be consulted before completing
an equipment design.
With the exception of a few covering fixed highfrequency
types and two or three covering attenuation networks (two or
more attenuators connected in series and mounted in a portable
case or on a relay rack panel), there are no MILSpecs on attenuators.
However, attenuators are available which will meet most of the
military environmental requirements such as: temperature extremes,
humidity, shock, vibration, and salt spray. For detailed information,
the manufacturer should be consulted.
Attenuator Ordering Information
Most attenuator manufacturers have catalogues which show
the electrical and mechanical specifications of their "standard"
attenuators. When items are required which are not included
in the catalogue, the information supplied to the manufacturer
should be as complete as possible to avoid delay or misunderstanding.
As much as possible of the following information should be included.
1. Complete electrical specifications including whether the
attenuator is to be fixed or variable. If fixed, include type
of circuit, decibel loss, impedance (both input and output),
required accuracy, frequency range, and operating level (power).
If variable, include type of circuit, impedance (both input
and output), required accuracy, frequency range, operating level
(power), number of steps and dB per step, linear or tapered,
are "off" or "cue" positions required? (both input and output
terminated but no signal goes through).
2. Mechanical specifications. If fixed, include maximum physical
size (if space is important), and type of terminals and connectors.
If variable, include physical size or manufacturer's basic
type, with or without detents (indexing device), any required
shaft deviations from manufacturer's standard, and type of terminals
or connectors.
Finally, specify any military requirements which must be
met.
Posted November 20, 2014