Book 3 Module 4

CATEGORY B1 B2
SERVO MECHANISMS

Licence By Post

For best examination

results always use latest
issue number.

Licence By Post © Copyright B EASA 66 4.3 ISSUE 05 1208

© Licence By Post
No part of this study book may be re-produced or distributed in any form or by
any means, or stored in a data base or retrieval system in whole or in part without
prior written permission from Licence By Post.

Books in the LBP series are regularly up-dated/re-written to keep pace with the changing
technology, changing examination requirements and changing legal requirements.
AUTHORITY

It is IMPORTANT to note that the information in this book is for study/training

purposes only.

When carrying out a procedure/work on aircraft/aircraft equipment you MUST

always refer to the relevant aircraft maintenance manual or equipment
manufacturer’s handbook.

You should also follow the requirements of your national regulatory authority (the
CAA in the UK) and laid down company policy as regards local procedures, recording,
report writing, documentation etc.

For health and safety in the workplace you should follow the regulations/guidelines
as specified by the equipment manufacturer, your company, national safety
authorities and national governments.
CONTENTS

Page

Servomechanism principles 1
Open loop control systems 2
Closed loop control systems 2
Servo mechanisms 5
Performance of servomechanisms 7
Improvement of transient response 13
Differentiators 23
Integrators 24
Servomechanisms systems and components 25
Synchronous transmitters 31
Synchros 38
Torque synchros 38
Control synchros 42
Differential synchros 45
Torque differential synchro system 46
The control differential synchro system 48
Resolver synchros 50
Slab synchros 53
The synchrotel 53
Inductance transmitters 55
The E and I bar transducer 55
Linear variable differential transformers 57
Capacitance transmitters 59
Syllabus 63
HOW TO TACKLE THIS BOOK

In general the same applies to this book as to book 1 in this module, in other
words, the category B person should have a sound knowledge of the subject (in
this case servo-mechanisms) and their applications. This means that it is
written to the B2 level, and B1 people should check the How to Tackle This
Book section of book 1 to check what subjects should be learnt and to what
depth. None of the contents applies to the A Line Mechanic.

You may need to read some sections of the book more than once to get a
thorough understanding of the subject, and for Full Student students you can
always contact your tutor if still in doubt.

Some details of scientists have been included for interest only and these need
not be committed to memory.

Note that when dealing with the theory of feed-back we have used electrical
signals as the signal medium and for the mechanical parts we have used,
predominately, rotating shafts. The signal system could be electronic or a light
system and the mechanical element could be motors, shafts, steel cables,
pulleys, hydraulic systems, pneumatic systems, drive belts, etc, or a
combination of any of these.

The signalling system could include computers with pre-programmed

responses to signals to cope with the various frictions in the system and to
keep response time short.

Remember, a feed-back system is not required when the system has to

produce a discrete output. In other words, OPEN or CLOSED (fuel cocks), UP
or DOWN (landing gear). When the system output is variable or continuous
such as flap position (UP, PART DOWN, DOWN), engine throttle controls
(CRUISE THRUST, LANDING THRUST, TAKE-OFF THRUST) then a feed-back
system will be required.

Probably the best examples of a continuously variable output system are

powered flying controls and powered nose wheel steering.

Lines used in drawings are used to represent electrical and mechanical

connections. To distinguish between the two they are drawn thus:
 SERVOMECHANISM PRINCIPLES

A servomechanism is a force amplifier mechanism where the output accurately

follows the input but with greater power. The output movement can be linear
or rotary. If the output is controlled by the input only, with no feed-back, it is
called an Open Loop system. If there is some form of feed-back or follow-up the
system is called a Closed Loop system using the servo-loop to provide part of
the control. Most closed-loop systems use negative feed-back where essentially
the servomechanism constantly compares demand input with output and any
difference generating an error signal which drives the output in the required
direction to reduce the error to zero.

Servomechanisms may be powered:

* Hydraulically – as in servo valves used in powered flying control

systems.
* Electrically – ac or dc powered.

Inputs into servomechanisms may be:

* Mechanical – push/pull rods, pilot operated from the flight deck

for example.
* Hydraulic – from a control valve for example.
* Electrical – ac or dc.

Control within the servo mechanism may be:

* Mechanical – as in most powered flying control units.

* Electrical – using analogue signals.
* Electronic – using position and speed data converted into digital
data.

In the industry, servomechanisms are generally considered to be electrical in

operation, hence their inclusion in the module 3 part of the EASA syllabus.

So we have concentrated here on the electrically operated servomechanisms

while hydraulic servos are covered in module 11/12 (for the mechanical
person) and module 13 (for the avionic person).

Most of the examples in this book use rotary motion but this, of course, can
easily be changed into linear motion by the inclusion of a simple gearbox.

-1-
OPEN-LOOP CONTROL SYSTEMS

Suppose that we wish to control the position of a radar scanner (the load).
Suppose also that we have a motor capable of driving the scanner and some
means of controlling the motor. Such an arrangement is illustrated in figure 1.

Fig. 1 OPEN LOOP CONTROL SYSTEM

The control element controls the magnitude and direction of the input to a
power amplifier, whose output drives the motor at the desired speed in the
required direction. The motor, in turn, moves the load in accordance with the
input demand. The motor is connected to the load mechanically, eg a set of
gears or a rotating drive shaft or moving levers/chains etc.

The control element (in the cockpit) could be calibrated with a scale indicating
the required position of the load. When the control dial is set to the correct
setting for the required position, the load (usually unseen) will move in the
required direction and at a particular speed, and with some systems at a
particular acceleration (rate of change of speed).

In practice, the accuracy of this form of control is limited because there are
several factors, other than the input, that affect the output. For example,
variations in the output load (friction within the support bearings – for
example) would effect the speed of the motor. Any variations in the amplifier
characteristics would also affect the motor as well as the motor circuitry itself.

There are no means of controlling these variations in an open-loop system and,

because of the resulting inaccuracy, open-loop systems are rarely used.

CLOSED-LOOP CONTROL SYSTEMS

These are control systems where the output from the system is feed back to
the input to modify the output. This is called feed-back and can come in many
forms.

The feed-back can be human, mechanical, electrical or electronic.

-2-
If the operator, for example, observes what the load is doing and makes
appropriate corrections at the input, the system is no longer open-loop; it is
now, in effect, a closed-loop system with the human operator completing the
loop between output and input. He/she compares the desired effect with the
actual effect and adjusts the system so as to reduce the error between them.
He/she is thus an 'error detector', and the amount of error which the person
observes determines how adjustments are made to the input to produce the
desired results.

This sort of control is used – using the human error detector – in almost every
activity we do – driving a car – combing ones hair – using hand tools etc.

To measure the error and take the necessary correcting action, we have ‘built
in’ the human operator as an essential control element. A more effective and
efficient control can be obtained by replacing the human operator with an
automatic control system. In addition there is a saving in manpower costs.

The essential features of a closed loop system are:

* The feed-back of output information concerning the behaviour of

the load.
* The comparison of this information with the behaviour demanded
by the input.
* The production of an error signal proportional to the difference
between the desired behaviour and the actual behaviour.
* The amplification of the error signal to control the power into a
servomotor (for electrically controlled systems).
* The movement of the load by the servomotor in such a
direction/speed/acceleration as to reduce the error signal to zero,
at which point the output is the same as that demanded by the
input.

Negative Feedback

With most feed-back control systems the feed-back is negative feed-back. In

other words the feed-back tries to negate or reduce the original input demand
signal so that the output from the system more closely relates to the original
input.

If positive feed-back occurs then the result is usually chaos because the
original demand input is reinforced by the feed-back signal so the input signal
continuously grows. As the input signal grows so does the output signal, with
an increase in the output signal so the feed-back gets bigger, increasing the
input signal – and so on. When this happens it is usually the result of some
malfunction in the system.

All the feed-back systems in this book are negative feed-back systems unless
stated otherwise.

-3-
Figure 2 shows a block diagram of a basic closed-loop control system. In this
arrangement:

* θi is the input demand, which in this case is in the form of a shaft

angle. (θi = theta in).
* θo is the output shaft angle of the load (θo = theta out).
* The control element converts the demand θi into some form
suitable for operation of the error detector, eg produces a voltage
proportional to θi.

Fig. 2 CLOSED LOOP CONTROL SYSTEM

* The feedback element does the same for the output angle θo, eg
produces a voltage proportional to θo.
* The error detector has two inputs applied to it, one due to θi and
the other due to θo. It produces an error signal e proportional to
the difference between the two inputs, ie (θi - θo).
* The error signal operates the power amplifier which, in turn,
causes the motor to rotate until θo equals θi (output equals
demand). At this point the error signal is zero and the drive from
the motor ceases, the output load having taken up the position
demanded by the input.

A Practical Closed-Loop Control System

Figure 3 shows a signal transmitter system consisting of a rotor within a

three coil stator. It is wired to a control transformer, which can be some
distance away, eg at the other end of the aircraft. The system is ac powered
with the rotor of the control transformer (CT) taking up the same position as
the rotor of the transmitter (CX).

-4-
The mechanical input demand θi sets the angle of the CX rotor. The resulting
alternating field in the CT stator coils induces a voltage in the CT rotor and
this voltage is fed as an error or misalignment signal to the amplifier. The
amplifier output is used to drive an ac servomotor that turns the output shaft
to the load and also the shaft to the CT rotor. This shaft turns the CT rotor
through output angle θo.

When the output shaft is turned into alignment with the setting of the input
shaft (θo = θi) the CT rotor is at right angles to the CX rotor and its own stator
field. In this position there is no error signal induced in the CT rotor, there is
no input to the amplifier and the servomotor stops. The output has now taken
up the position demanded by the input.

Fig. 3 PRACTICAL CLOSED LOOP SYSTEM

Deadband

So far we have assumed that how-ever small the input signal there will be
some response at the output. This, however, is not true – or at least depends
very much on the system. With all systems there is a Deadband – a range of
input movement, usually very small, that will not produce any output. This
will occur about each null position. The deadband can be reduced to an
absolute minimum, but this is expensive and may not be necessary.

SERVO-MECHANISMS

To be classed as a servomechanism, an automatic control system must have:

(d) Error actuation.

(d) Power amplification.
(d) Closed loop control.
(d) Continuous operation, or ‘follow-up’ properties, ie if the load is
disturbed from the demanded position, it will return to that
position.

-5-
The system is said to be error actuated because it is the error between the
output demanded by the input and the actual output which starts the action.
The final net input to the amplifier is the error signal and not the input
demand.

There must be torque amplification to be able to drive heavy loads so the servo
contains an amplifier that supplies the necessary driving power to the
servomotor, the motor providing the required torque.

Continuous operation is assured in a servomechanism because any variation

in the output from that demanded by the input automatically produces a
difference signal between output and input, and hence an error signal. The
error signal again starts the correcting action.

A servomechanism has many applications, covering a wide range of power,

position, speed and acceleration requirements.

Types of Servo

There are two main classes of servomechanism - Remote Position Control

(RPC) servos and Speed Control servos.

(a) RPC servos. These are used to control the angular or linear
position of a load.

(b) Speed Control servos. These are used to control the speed of a
load. In this case, the speed of the driving motor is made
proportional to the input demand (usually a voltage input).

Inputs

The two types of input to a servo are:

1. Step Input - created when the input shaft is suddenly rotated from
one angular position to another. Step inputs are considered as a
movement from one position to another taking zero time and
shown on a graph as a straight vertical line. In fact there is always
a finite time taken from position 1 to position 2 but for theoretical
considerations this is ignored.

2. Ramp Input - created when the input shaft is rotated at a constant

angular velocity.

-6-
 Fig. 4 STEP INPUT

Fig. 5 RAMP INPUT

PERFORMANCE OF SERVOMECHANISMS

Response

The response of a servo is the pattern of behaviour of the load when a change
is made to the input condition. It has so far been assumed that if the input
moves to θi the load will simply follow, its response being a reproduction of the
input movement.

The paragraphs that follow will show that matters are not as simple as this.

Step Input - No Friction

With reference to figure 6. Assume that the input and output are aligned at θ0
until the input suddenly changes to θi. An error signal proportional to θ0 - θi
appears at the amplifier input from the error detector and the motor is
energised to move and null, or cancel, the error.

-7-
 Fig. 6 SIMPLE SERVOMECHANISM CONTROL SYSTEM

One important point must be emphasised. The torque delivered by the motor
to the load is directly proportional to the error. It acts only on the inertia of the
load, which therefore accelerates at a rate proportional to the error. As the
error reduces so the acceleration reduces, until it reaches zero with zero error.

But this is not a satisfactory state of affairs, for the load acceleration is in one
sense only and that to increase its velocity. Saying that the acceleration is zero
at zero error simply means that the load has reached a steady speed when we
require it to be stationary. Further, since there is nothing to stop it so it moves
past the required position.

The error signal produced, and, therefore, the torque applied to the load, now
reverses in sense to slow down the load. Since, however, the components
operate symmetrically about the null, the pattern of deceleration is a mirror
image of the original acceleration.

The load stops when it has overshot by the initial error, and from there the
performance is repeated. The resulting load oscillation about the demanded
position is illustrated graphically in figure 7.

Fig. 7 RESPONSE TO A STEP INPUT - NO FRICTION

-8-
Ramp Input - No Friction

The description of the response can be followed in figure 8. In the early stages
of the ramp, while the error signal is small, the load accelerates slowly and
lags behind the input.

The error signal grows as the lag increases, building up the acceleration.
Eventually the load speed equals the input speed but since a substantial
position error exists it continues to accelerate.

When its speed exceeds that of the input the position error starts to decrease;
the acceleration reduces and the load reaches a constant speed at zero
position error with no error signal.

The load speed, however, exceeds the input speed and an overshoot results.
Thus the outcome is a continuous oscillation as shown.

Fig. 8 RESPONSE TO A RAMP INPUT - NO FRICTION

Effect of Restraints

The oscillatory responses are obviously not desirable, and luckily, restraints
on the load have a stabilising effect. Various inherent factors oppose the load
movement and they include static friction, kinetic (sliding) friction, electrical
eddy currents, air resistance, viscous lubricants and others.

Lumping them all together for the moment, the general effect is to reduce the
amplitude of each successive swing until gradually the output becomes steady.
The oscillations are known as transients and they are effective during the
transient response period, or settling time. Once the output has settled it has
reached the steady state.

While restraints are beneficial in stabilising, or damping the response, they do

have certain detrimental effects. One of these is that power is wasted; another
is the introduction of error into the steady state.

-9-
Steady State Errors

Examination of the various restraints present would show that their effect is in
part due to a small constant magnitude force known as Coulomb Friction and
in part to Viscous Friction that increases with speed.

Coulomb friction is that part of the frictional force that is independent of

speed, for example a brake shoe pressed against a brake drum. Its breaking
effect dose not change with speed but is reasonably constant.

The resistance due to coulomb friction tends to degrade the sensitivity of a

servo, for a torque has to be created to overcome it before there is any
movement of the load. To provide this torque the load error must reach some
finite size, and any errors less than this will not produce movement. Figure 9
shows the effect of coulomb friction on the response to a step input.

Fig. 9 RESPONSE TO A STEP INPUT WITH COULUMB FRICTION

The load comes to rest somewhere within a band of error, known as the Dead
Space, the width of which depends on the amount of coulomb friction. For
most modern servos the coulomb friction is very small, and its effect is often
ignored.

Viscous friction does not produce a dead space in the step input case since it
has no value when the speed is zero. It does however produce a similar effect
when a ramp input is considered.

In the steady state the load is moving with constant speed; it is therefore being
resisted by viscous friction. An error signal must be produced to overcome
this, therefore, an error must exist.

The response is illustrated in figure 10 and the error necessary to overcome

the friction is known as Velocity Lag.

- 10 -
The output shaft rotates at the same speed as the input shaft but lags behind
by some constant angle. This positional error is velocity lag.

Coulomb friction may be considered small compared with viscous friction

during a ramp input, but, of course, it also contributes to this error. However,
the greater part is due to viscous friction, and since this increases with speed
the error is generally reckoned to vary in direct proportion to speed.

Fig. 10 RESPONSE TO A RAMP INPUT WITH VISCOUS FRICTION

Response of a RPC Servo to a Step Input (Negligible Coulomb Friction)

With reference to figure 11. Assuming the output shaft is driving a load, and
has taken up a position which agrees with that demanded by the input shaft
(θ0 = θi). The error signal is zero and the servo is stationary.

Fig. 11 BASIC RPC SERVO SYSTEM

- 11 -
 Fig. 12 RESPONSE CURVE

Now suppose a step input is applied. The sequence of events is illustrated in

figure 12.

(a) The input demand θi is suddenly changed to a new value at point

a on the graph. The output shaft cannot immediately follow this
change in demand because of the inertia of the load. Therefore
there is now a difference between θ0 and θi and an error signal is
produced.

(b) The error signal, after amplification, causes the motor to

accelerate in an attempt to bring the output shaft to the new
demanded position. Because of the inertia of the load this takes
time; there is therefore a time lag during which the output angle θ0
is changing in response to the change in demand (b in figure 12).

(c) As the motor turns the load, the output angle θ0 approaches the
demand angle θi. The error signal, which is proportional to (θi - θ0),
therefore decreases but the driving force remains until θ0 equals θi
at point c; this is the required load position.

(d) By the time θ0 has reached the demanded position, the load has
acquired momentum and overshoots to point d.

(e) The error signal now increases in the opposite direction (θ0 greater
than θi) and the motor applies a reverse torque which eventually
stops the load and brings it back to the required position at point
e.

- 12 -
 (f) Once again, however, the momentum of the load carries it past the
required position and another overshoot occurs (to point f).

The load may thus oscillate about its final required position many times before
it comes to rest. A servomechanism that does this is said to be ‘hunting'.

IMPROVEMENT OF TRANSIENT RESPONSE

For many applications the simple servo using its inherent friction for damping
is perfectly adequate. This is usually the case for small position servos, but
where large loads are involved the transient response is unsatisfactory.

Time and energy is wasted during this period, and component wear is
increased. The number of oscillations should be reduced and also the response
time shortened. Two methods are commonly employed – viscous damping and
velocity feedback damping.

Viscous Damping

This method is simply a controlled increase of the inherent viscous damping to

achieve the required response. One device in use is the eddy current damper
as shown in figure 13.

This simple device consists of a thin disc of metal with high electrical
conductivity (usually aluminium), which is attached to the output shaft of the
servo. It spins between the poles of stationary electromagnets mounted around
its periphery.

Fig. 13 EDDY CURRENT DAMPER

- 13 -
Electromagnetic eddy currents are induced of a magnitude proportional to the
field strength and to the disc velocity. These eddy currents set up magnetic
fields that act against the inducing fields and forces opposing the disc rotation
are created.

These forces are close to being proportional to disc velocity, and therefore
provide parallels to the inherent viscous forces. They can be controlled by
adjusting the current flow to the electromagnets, thus varying the amount of
damping that can be applied.

Figure 14 shows the effect of varying levels of viscous damping, coulomb

friction being ignored for simplicity. Using only inherent friction Under
Damping is achieved and the system takes a long time to settle down.

Too much viscous friction will produce a sluggish response and the system is
Over Damped. The degree of damping which just prevents any overshoot is
known as Critical Damping.

Slightly less damping than this, to allow one small overshoot is known as
Optimum Damping, which gives the shortest settling time.

Most designers aim for this condition.

Fig. 14 LEVELS OF DAMPING - STEP INPUT

The effect on the transients for a ramp input can be similarly adjusted to
produce optimum damping. A snag arises, however. This is the fact that for
any increase in viscous friction there is an increase in velocity lag.

Thus to remove the transient oscillations completely a considerable velocity lag

must be expected. Figure 15 illustrates the response for two levels of damping
for a ramp input.

- 14 -
 Fig. 15 LEVELS OF DAMPING - RAMP INPUT

Velocity Feedback Damping

This form of damping is similar to viscous friction damping in one respect; the
compensation produced is proportional to the velocity, or rate of movement, of
the output shaft. Velocity feedback damping has the advantage, however, that
it consumes little power.

In an RPC servo the load is required to be moved from one position to another
as quickly as possible without causing instability such as hunting or wasting
power.

We have already seen that a step input applied to a servo causes the
servomotor to apply a torque, which accelerates the load. As the load gathers
speed and approaches the desired position an arrangement is required that
will 'anticipate' that the load is going to overshoot and so reverse the motor
torque before the desired position is reached.

If the arrangement is adjusted correctly the result is that the load comes to
rest just as it reaches the required position with no overshoot and no hunting.

For a servomechanism, this arrangement is achieved by attaching a tacho-

generator to the output shaft. A tacho-generator is a small ac or dc generator
that produces a voltage proportional to the angular velocity of the output
shaft.

A suitable fraction of this voltage is fed back to the input of the amplifier in
opposition to the error signal (negative feedback) to provide the necessary
compensation. This is known as Velocity Feedback Damping, because the
voltage fed back is proportional to the velocity of the output shaft.

- 15 -
The aim, with velocity feedback, is to reduce the net input to the amplifier to
zero and then to reverse it before the output shaft reaches its required
position.

If the amount of feedback is correctly adjusted - and this can be done fairly
easily by means of a potentiometer - the forward momentum of the load, acting
against the reversed torque, causes the load to come to rest just as it reaches
the required position.

Fig. 16 SERVO SCHEMATIC WITH VELOCITY FEEDBACK DAMPING

The action is illustrated by the graphs in figure 17. Initially, when the step
input is applied, the output shaft does not move and the full amplified error is
applied to the motor (no velocity feedback since the load is at rest).

The motor accelerates the load and, as the input and output shafts come into
alignment, the error falls. At the same time, because the load is now moving
fast, a large velocity feedback voltage is applied in opposition to the error
voltage.

The net input to the amplifier therefore drops rapidly and then increases in the
opposite direction as the velocity feedback voltage becomes larger than the
error signal. A decelerating torque is thus applied to the load before it reaches
the demanded position. As the motor slows down, the velocity feedback voltage
falls; so also does the error signal because the output and input shafts are
coming more into alignment.

- 16 -
 Fig. 17 VELOCITY FEEDBACK DAMPING

When the required position is reached, the motor stops; velocity feedback is
zero and so also is the error signal, the input and output shafts being aligned.

So transient response can be improved in two ways, by applying extra viscous

friction or by velocity feedback. Both increase velocity lag in response to ramp
inputs, but of the two, velocity feedback is preferred since power is not wasted.

Note that in figure 16 ‘theta in’ is compared to ‘theta out’ by a mechanical

error detector. With many systems the unit would be replaced by electrical or
electronic unit with electrical inputs.

Just to re-cap:

* The feedback voltage from the tacho-generator to the amplifier is in

opposition to the error signal.

* In a servo that is being driven at a constant speed there must always be

an error signal to keep the load moving, however this error signal will
have to be larger than required as it has to be greater than the velocity
feedback voltage. Therefore, velocity lag is greater than required, ie
greater positional error between input and output shafts.

We shall now look at the circuits of the more common methods of reducing
velocity lag in step input and ramp input systems.

- 17 -
Transient Velocity Feedback

This type of damping is often referred to as 'acceleration feedback damping'

and is used in an angular velocity system to reduce velocity lag.

With reference to figure 18. Note the Transducers for transducing the
mechanical signal into an electrical signal (or vice-versa) and the summing
unit for calculating E by taking the input ei from the output eo electrically.

Inserted between the tacho-generator and the amplifier in the feed-back circuit
is a differentiating network.

In a speed control servo, the velocity feed-back provides damping when

transient speed changes occur. However, when the speed is constant no
damping is required, but the signal is still being fed to the amplifier and
causing excessive velocity lag.

Fig. 18 TRANSIENT VELOCITY FEED-BACK SYSTEM

The transient velocity feedback system is designed, therefore, to provide

damping during changes in speed, and when the speed is constant, provide no
damping and therefore reduce velocity lag to a minimum.

When the speed of the output shaft is constant the tacho-generator voltage is
constant. After the initial charging current the voltage across R falls to zero
and there is no output from the network, no damping to the system and
velocity lag is reduced. When the speed changes (when damping is required),
the capacitor charges or discharges through R to give an output to the
amplifier to provide the necessary damping.

- 18 -
The tacho-generator output is therefore differentiated by the CR circuit giving
damping only when the load speed is changing.

Phase Advance Damping

Stabilisation of a servomechanism to obtain a good transient response in an

RPC system and a good steady state response in a velocity system and
reducing velocity lag can also be achieved by inserting a suitable network in
the input to the amplifier.

A typical arrangement using a correction network, known as a phase-advance

network is shown in figure 19.

In an RPC system when a step input is applied, there is a large error signal,
the current through R and the charging current through C are in the same
direction. The input to the amplifier is therefore IR + IC (modified error signal).
This causes the servomotor to accelerate rapidly. As the error signal decreases,
C commences to discharge through R and the total current, therefore, into the
servo amplifier is IR - IC.

Fig. 19 PHASE ADVANCE CORRECTION NETWORK

By suitable choices of C and R (controlling the time constant of the network), it

can be arranged that the signal itself becomes zero before the error signal is
zero, providing a retarding torque before the load reaches the required
position. Overshooting is prevented and stability during the transient period
improved. Velocity feed-back damping is omitted, the network is doing the
same job but is not introducing velocity lag.

- 19 -
For a velocity control system (ramp input), the network provides damping on
the initial switch ‘on’ and for any speed change as described for the step input
system.

However, at steady state the error is very small and velocity lag is greatly
reduced. So this system, by removing velocity feedback damping, has reduced
velocity lag, the correction network providing the necessary damping.

Other types of correction networks are used, depending on the response

required. Some RPC servos require only a rapid response with velocity lag
being unimportant; in other RPC servos, velocity lag must be taken into
account. In systems required to rotate at a constant angular velocity, velocity
lag must be reduced to a minimum. Different combinations of C and R in the
network can be used to produce the desired response.

Error Rate Damping

Another method used to reduce velocity lag in an angular velocity control

system is error rate damping. One possible arrangement is shown in figure 20.

One tacho-generator is mounted on the output shaft and produces a voltage,

sent to the amplifier, proportional to the speed of the shaft. A second tacho-
generator is mounted on the input shaft producing a voltage proportional to
the input speed, which is also sent to the amplifier. There are, therefore, three
input signals to the amplifier: output shaft velocity; input shaft velocity and
error signal. The combined input is a voltage proportional to the error plus a
voltage proportional to the input speed minus a voltage proportional to the
output speed.

Fig. 20 ERROR RATE DAMPING

- 20 -
If you remember, in a velocity feed-back system in the steady state, ie input
and output shafts rotating at the same speed, the velocity lag was mainly
caused by the signal from the output tacho-generator.

In this system the input tacho-generator is cancelling this signal in the steady
state, so velocity lag is reduced to a minimum; the position error reducing
further any velocity lag that does exist.

In principle, this arrangement seems fine but it is difficult to get two tacho-
generators to give exactly the same output, especially with time.

However a simplification of this arrangement is possible. The two tacho-

generators are producing a summed voltage output proportional to input
minus output speed, or proportional to the speed of the error signal. The velocity
of this error is equal to rate-of-change of error with respect to time. This can be
achieved by using a differentiator, ie differentiating the error with respect to
time.

Therefore by combining the derivative of the error signal and combining it with
the actual error, the net input to the amplifier is a voltage proportional to the
error plus a voltage proportional to speed (input minus output). So therefore,
in the steady state, the system has low velocity lag. In the transient state the
differentiator will provide the necessary damping signals.

Fig. 21 ERROR RATE DAMPING WITH DIFFERENTIATOR

- 21 -
At start up, rate-of-change of error signal is high and the input to the amplifier
will be high to accelerate the load. As the input and output shaft speeds’ come
into line the output of the differentiator opposes the position error signal to
provide a retarding torque before the load reaches the required speed. At
steady state the differentiator output is zero and position error signal reduces
velocity lag.

Integral Control

The methods so far described reduce velocity lag, but have no effect on lag and
dead space caused by inherent friction. A common method of dealing with
these residual steady state errors is known as Integral Control.

Fig. 22 INTEGRAL CONTROL

The differentiator (figure 22) acts as previously described for error and
damping. The integrator is connected as shown, which integrates the error
signal and feeds it into the amplifier. An integrator is a device that takes an
input signal and slowly builds up that input signal on its output.

When the servo system reaches its steady state with a small residual error, the
integrator output will have built up to provide an additional signal to drive the
load to the alignment position.

Adjustment of the integrator output can be made to ensure that when the
error signal is zero its output is just sufficient to counter the inherent friction.

- 22 -
Summary

Most servos in aircraft systems are damped by inherent friction, extra viscous
friction or velocity feedback. However in some applications, improved
sensitivity and reduction of velocity lag are required. Figure 23 shows a
comparison of damping methods.

Fig. 23 COMPARISON OF DAMPING METHODS

DIFFERENTIATORS

Figure 24 shows a simple differentiator circuit consisting of a resistor (R) and a

capacitor (C). This circuit could be used as a high pass filter, however, with a
square wave input and a short time constant compared to the input pulse, it
can be used as a differentiating circuit.

Fig. 24 DIFFERENTIATOR CIRCUIT

- 23 -
 Fig. 25 INPUT/OUPUT GRAPHS FOR A DIFFERENTIATOR

Figure 25 shows the input pulse (V) and the voltages across C (VC) and R (VR).

Note that:

(a) The voltage output across C is a close approximation to the shape

of the input pulse, the shorter the time constant (CR) the closer
the approximation.
(b) The output across R consists of two very narrow pulses (or spikes)
of opposite sign, one coinciding with the start, the other with the
end of the pulse.
(c) VR = V0.

The output is therefore proportional to how fast the input voltage changes, ie it
is sensitive to rate of change on input voltage.

INTEGRATORS

If the positions of C and R are changed and the time constant is large
compared to the input pulse then an integrating circuit is formed (figure 26).
(Note this basic circuit could be used as a low pass filter). Figure 27 shows the
input pulse to a RC circuit with a long time constant and the resulting output
across the capacitor.

blank

- 24 -
 Fig. 26 INTEGRATOR CIRCUIT

Fig. 27 GRAPH OF INTEGRATOR INPUT/OUTPUT

The integrator therefore takes an input and produces an output across the
capacitor over a period of time depending on the values of C and R, ie the
output is proportional to the time integral of the input.

SERVO MECHANISMS – SYSTEMS & COMPONENTS

There are many different types of servo mechanisms fitted to aircraft and this
section of the book deals with some of the more common systems.

A DC Servo System

The positional feedback signal would usually come from a potentiometer

(variable resistor). The error detector would be potentiometer circuits or
electronic circuits producing an error signal proportional to the difference
between input and output shaft rotations.

The servo amplifiers must produce sufficient power to drive the servomotor
and this power requirement will depend on the system. The power amplifiers
used in conjunction with voltage amplifiers (to provide gain) will thus vary the
power output.

- 25 -
DC servomotors must be capable of being reversed and the torque developed
must be proportional to the error signal input. As torque is proportional to the
product of the armature current and the field current, the armature current is
kept constant so the torque can be controlled by varying the field current
(error signal).

This means that separate supplies are used, the armature current being fed
from a constant current source and the field current being fed from the servo
amplifier. Figure 28 shows a dc operated servomotor. Attached to the motor is
a servo amplifier. Feedback is provided by a potentiometer, the wiper arm of
which is driven by the motor.

Fig. 28 DC OPERATED SERVOMOTOR

The dc tacho-generator mounted on the output shaft of the servomechanism is

a small, separately excited dc generator. It will produce a dc voltage that is
directly proportional to the speed at which it is driven, and the polarity
depends on the direction of rotation.

An AC Servo System

The positional feed-back signal may be from potentiometers, synchros, linear

variable differential transformers (LVDT's) or rotary variable differential
transformers (RVDT's). The servo amplifier is a conventional transistorised
amplifier with voltage and power amplifiers.

AC servomotors capable of producing moderately high torque are usually

induction motors. These run at a constant speed (related to the frequency of
the ac supply), but their torque and their direction of rotation can be
controlled by the amplified error signal, without the use of complex circuitry.

Two-phase and three-phase induction motors are common in ac servo systems

but the two-phase type is more usual. The two-phase induction motor (figure
29) requires two ac supply voltages 90° out of phase with each other. One
phase is fed to the control system and then through the servo amplifier to one
stator winding of the motor; this supply is controlled by the error signal.

- 26 -
The other phase, known as the 'reference phase', is fed directly to the other
phase winding of the motor.

When the reference phase to one stator winding and the error-controlled phase
to the other winding are both present, a rotating magnetic field is produced
and the squirrel-cage rotor follows this field and rotates.

The direction of rotation depends on the sense of the error signal, ie on the
direction of misalignment between input and output shafts of the servo.

As the misalignment signal passes through the zero error signal position, the
phase of the error controlled voltage reverses and this reverses the direction of
rotation of the motor. When the error is zero, the error-controlled phase
voltage is also zero and no rotating magnetic field is produced and the motor
stops.

The torque developed by the motor depends on the magnitude of the error
signal; the greater the error, the larger the torque.

Fig. 29 TWO PHASE INDUCTION MOTOR

Hysteresis motors may also be used as ac servomotors. A typical arrangement

is shown in figure 30. A two-phase supply is required, the reference phase
voltage being applied to phase windings AA1, and the error-controlled voltage
to BB1. At any one instant, A will act as a N-pole and A1 as a S-pole; B and B1
will be neutral.

The rotor (a cobalt steel ring) will therefore have an S-pole induced at X and a
N-pole at Y. A quarter of a cycle later, B will act as a N-pole and B1 as a S-pole,
A and A1 will be neutral. However, the rotor, which is made of a large
hysteresis loop material, will have retained its S-pole at X and its N-pole at Y.
So point X on the rotor is attracted to the B stator winding and Y is attracted
to B1.

- 27 -
Since a torque is being exerted on the rotor it will turn to follow the rotating
magnetic field and will continue to do so until the error signal falls to zero. The
direction of rotation depends on the phase of the error-controlled voltage
relative to the reference voltage and this, in turn, depends on the sense of the
error. The torque developed depends on the magnitude of the error.

Fig. 30 HYSTERESIS SERVOMOTOR

Servomotor Construction

Figure 31 shows a two-phase induction motor with the two phase windings
accommodated in slots in the stator. The rotor is of the squirrel cage
construction type generally with aluminium conductors. The stator and rotor
slots are skewed which prevents the reluctance of the rotor moving away from
one of a number of positions when the slots in the rotor and stator are aligned
and so ensures a smooth torque output.

The ac tacho-generator is used to provide velocity feedback damping in ac

servo systems and is mounted on the output shaft so that it rotates at the
same speed as the load.

Tacho-Generator Construction

The ac tacho-generator (figure 32) is usually a drag-cup generator that

produces an alternating voltage of the same frequency as the ac supply.
However, the amplitude of the voltage depends on the speed of rotation, and
the phase of the voltage leads or lags the ac supply depending on the direction
of rotation.

The drag-cup generator has a two-coil stator whose axes are at right angles, as
in a two-phase induction motor. The rotor, however, is a copper or brass cup
with a stationary iron cylinder fitted inside it to complete the magnetic path for
the stator.

- 28 -
 Fig. 31 SERVOMOTOR CONSTRUCTION

Fig. 32 AC TACHOGENERATOR DETAILS

AC is applied to only one stator coil – the primary coil (the reference winding).
As the output shaft drives the rotor, voltages are induced in it by interaction
with the field produced by the current in the primary stator. At any instant,
maximum voltage is induced in that part of the cup passing through the
primary axis.

Maximum circulating (eddy) currents are therefore on this axis, and these
currents produce a secondary magnetic field that is at right angles to the axis
of the primary field.

This secondary field alternates at the frequency of the supply current but its
magnitude depends on the amplitude of the circulating eddy currents induced
in the rotor by the primary field. This, in turn, depends on the speed of the
drag cup rotation.

- 29 -
The secondary stator winding of the tacho-generator is at right angles to the
primary axis and so has a voltage induced in it by the secondary field only.

This is the output voltage whose amplitude is proportional to the speed of drag
cup rotation and whose phase relative to the reference ac supply depends on
the direction of rotation. A typical ac tacho-generator provides a velocity
feedback voltage of 0.5V per 1,000 rpm of the rotor.

Figure 33 shows the physical relationship between the reference and output
windings and also the relationship between the magnetic flux (phi – shown as
a zero O symbol with a vertical I through the middle) and theta (the angular
movement – shown as a 0 with a slash \ though the middle).

There should be no need to remember these details.

Fig. 33 DRAG CUP ROTATION

If a control system operates on ac but is supplied with dc, then a device known
as a modulator has to be employed. The modulator will convert the dc input
into an ac output of fixed frequency such that the magnitude of the ac output
is proportional to the magnitude of the dc input.

Whenever the dc input goes into reverse polarity, corresponding to a reversal

in the sense of the data being represented by the dc signal, the ac output will
reverse its phase eg, a +ve dc input produces an ac output as at (a) in figure
34 and a -ve dc input produces an ac output as at (b).

- 30 -
 Fig. 34 PHASE ORIENTATION

The modulator is said to be ‘phase sensitive’. Alternatively if the control system

operates on dc but is supplied with ac a demodulator is used. This converts
the ac input signal into a dc output signal.

Whenever the ac signal reverses phase, corresponding to a reversal in the

sense of the data being represented the dc will reverse polarity. So with an ac
input as at (a) in figure 34, this gives a +ve dc output, and an ac input as at (b)
gives a –ve dc output. The demodulator is said to be ‘phase sensitive’. It may
also be called a ‘phase sensitive rectifier’.

SYNCHRONOUS TRANSMITTERS

These are data (normally data on position) information transmitters that

transduce movement information into an electrical signal and sends this data
to remote receivers/indicators in real time (ie synchronously).

The Desynn System

The Desynn is a dc system of transmission that indicates continuous variation

of position. It is a relatively simple system, which, because it produces a low
torque, is useful only for remote indication, such as simple pointer and scale
indicator systems.

Come in a variety of designs which can be used to measure rotational

movement and linear movement. The principle of operation is the same for all
types. Here we will look at the most common type, the rotary Desynn.

As in all electrical remote rotary movement indication systems, the input shaft
is connected to a transmitter element that is connected to the receiver unit by
electrical cables. The receiver unit drives the output shaft that operates the
remote indicator pointer.

In the Desynn system (figure 35), the transmitter is a circular potentiometer

which has three fixed tapings (A, B and C) spaced 120° apart, each connected
to the receiver unit. A rotating, spring loaded, mechanism is mounted on the
input shaft and carries two sliding contacts or wipers that are arranged
opposite one-another.

- 31 -
 Fig. 35 THE DESYNN SYSTEM

The wipers are fed, via slip rings and brushes, from the positive and negative
terminals of a dc supply. The spring in the mechanism (not shown in the
drawing) is to remove any possible backlash in the system.

The receiver has three coils (A, B and C) with axes 120° apart (connected in
star). In the middle is a permanent magnet rotor which is capable of rotation
through 360° and which carries a pointer over a calibrated scale. The three
air-cored coils in the receiver are connected to the tapping points A, B and C
on the transmitter by the three lines shown in figure 35.

A weak circular magnet which plays no part in the normal operation of the
indicator, will return the pointer to an ‘OFF SCALE’ position if the power
supply fails or is switched 'OFF' (see figure 38).

When a dc supply is connected to the transmitter wipers, the voltages at the

tapping points cause currents to flow through the three stator coils in the
receiver. A resultant magnetic field is produced and the rotor magnet aligns
itself with this field. For example, with the input shaft in the position shown in
the top picture (position 1) of figure 36, tapping a is at +24V with respect to
supply and negative, while b and c are both +8V.

(Assuming a supply of 24 volts dc. From the positive wiper arm – at any
position - there is a volts drop to zero clockwise and anti-clockwise around the
toroidal resister to the negative wiper arm. So 2/3rds around the resistor – at
tapping b clockwise and at c anti-clockwise – the voltage is 8 volts.)

- 32 -
 Fig. 36 DESYNN SYSTEM - THEORY

This means that current will flow from tapping a in the transmitter through to
coil a in the receiver; it then divides equally and half the total current goes
through coil b and half through coil c back to the transmitter. The magnetic
fields fa, fb and fc associated with these currents and also the resultant
magnetic field are shown by vectors and the rotor magnet aligns itself with this
field.

The relative strength and direction of the currents are shown by the size of the
dotted arrows.

If the input shaft is rotated 120° clockwise (position 2 figure 36) the voltage
distribution is such that the total current flows through tapping b to coil b, it
then divides equally through a and c and flows back to the transmitter. The
vectors show that the resultant magnetic field has also rotated 120° clockwise
from its initial position and the rotor magnet aligns itself with this new axis.

The magnitude and polarity of the voltage at each tapping point on the
transmitter therefore varies according to the position of the wipers.

Thus, if the input shaft is rotated, the variation of voltage at a, b and c

produces changes in the currents flowing in the stator coils, and a magnetic
field rotating in sympathy with the input shaft is produced.

- 33 -
The rotor magnet remains aligned with this field at all times and so rotates in
synchronism with the input shaft. A pointer, moving over a calibrated scale, is
attached to the rotor so that a remote indication of the position of the input
shaft is immediately available.

Note that we have used a rotation of 120 degrees clockwise as it makes the
sums easier. The angle, of course, could be any angle. Note also that
transmitter and indicator are shown close together in the drawings, they can
be, and usually are, many metres apart connected by a three wire cable.

Fig. 37 A DESYNN TRANSMITTER – EXPLODED VIEW

Fig. 38 INDICATOR – EXPLODED VIEW

- 34 -
The Desynn is said to be self-synchronous because the output shaft
continuously moves in sympathy, or in synchronism, with the input shaft.

Any change in the position of the input shaft results in an immediate and
corresponding change in the position of the output shaft; this synchronous
movement is inherent in the system, ie it is self-synchronous.

The Desynn also gives a continuous indication of change of position through

360°. It is, however, only a low power device, the current being limited by the
resistance of the transmitter potentiometer. The Desynn is therefore only
suitable for operating indicating pointers and similar light loads.

Figures 37 and 38 show exploded views of the transmitter and the receiver
(indicator).

A typical example of the use of the Desynn is remote indication of wing flap
position with the flap operating mechanism connected to the input shaft to the
transmitter.

Figures 39 and 40 show a Desynn system used to indicate flap position on a

small passenger aircraft. Note the linkage between the flap and the transmitter
and the available adjustments. This is adjusted as per the aircraft
maintenance manual (AMM) so that the indicator shows that the flaps are up
when they have reached their upper limit of travel, and fully down when they
have reached their lower limit.

Fig. 39 DESYNN SYSTEM – TYPICAL APPLICATION

- 35 -
 Fig. 40 TYPICAL SYSTEM SET-UP

The table below shows some of the possible faults with Desynn systems and
the subsequent actions.

SYMPTOM CAUSE REMEDY

Pointer ‘OFF SCALE’ Power supply failure/ OFF Check fuse or C/B. Switch
‘ON’

Pointer displaced 180°. Power supply reversed. Check transmitter for correct
Rotation correct. wiring.

Pointer rotation reversed. 2 connections crossed. Check wiring connections

between transmitter and
indicator.

Pointer displaced 120° or 3 connections crossed. Check wiring connections

240° clockwise. Rotation between transmitter and
correct. indicator.

Pointer displaced and Open circuit or short circuit. Carry out insulation
moves in 180° steps. resistance check and
continuity checks.

TABLE 1 DESYNN FAULTS

Figure 41 shows the principle of the adjustment of the extension arm of the
transmitter.

- 36 -
 Fig. 41 PRINCIPLE OF LEVER LENGTH ADJUSTMENT

Maintenance Checks on Adjustment of Transmitter

There are usually two forms of adjustment, one adjusts lever length and the
other lever angle.

Fig. 42 PRINCIPLE OF LEVER ANGLE ADJUSTMENT

- 37 -
Lever length on the transmitter is the distance between the lever pivot point X
and the connection to the link at Y; this arm is adjustable as shown in detail
A-A figure 40. It should be noted that making the extension arm longer would
increase lever length and decrease indicator pointer movement. Making the
extension arm shorter will decrease lever length and increase indicator pointer
movement.

The basic adjustment of this type of system is to set the flaps (or what-ever
system the transmitter is fitted to) to various positions (checking that they
select correctly from the flight-deck and carrying out an actual angular or
linear measurement check of the flaps. Check also that they move smoothly
and all together (symmetrically – port and starboard). Check the flight-deck
pointer indication and adjust the lever length as required ensuring correct
pointer indication at all flap positions.

It should be noted that adjusting lever length generally only counteracts a

regularly increasing or decreasing error, eg for each whole unit of movement it
only moves 0.8 of a unit. In other systems, adjustment of lever angle may be
possible. This is the angle between the adjustable link arm and the extension
arm in figure 40. (Note. It cannot be adjusted on the system shown above).

Unlike lever length adjustment, lever angle adjustment is not a constant value,
ie for each adjustment the pointer deflection becomes progressively greater for
each indication position.

Therefore, on systems with both types of adjustments, it will be necessary to

adjust each in turn to obtain the desired pointer indications – referring to the
AMM of course.

SYNCHROS

Synchros are often used to transmit information from a remote point and
transfer this data to a control unit or instrument.

There are four basic types:

1. Torque synchro.
2. Control synchro.
3. Differential synchro.
4. Resolver synchro.

TORQUE SYNCHROS

These are the simplest of the synchros and used to transmit angular position
information to a remote receiver. The basic construction is shown in figure 43.

- 38 -
 Fig. 43 CUT-AWAY DRAWING OF A TYPICAL SYNCHRO

It consists of a single winding rotor supplied with single phase ac via slip rings
which is supported on a rotor shaft. The stator consists of three windings
connected in star 120º apart.

The rotor shaft is supported at both ends in roller bearings.

The torque synchro circuit is shown in figure 44. It consists of a Torque

Transmitter (TX) and a Torque Receiver (TR). They are shown closely connected
in the drawing but in reality the TX rotor shaft is connected mechanically to
the component position being monitored and the TR can be in the flight-deck
or at any other remote location in the aircraft.

Fig. 44 SYNCHRO CIRCUIT

- 39 -
The TX and TR are similar but the TR has some form of damping mechanism
to prevent oscillation. The TR can be used as a TX but the TX cannot be used
as a TR as it has no damping. Note the TR/TX symbol in figure 45.

Fig. 45 TR/TX SYMBOL

Operation [Figures 44 and 46 (a)]

The single phase ac is fed to both synchro rotors (R1 and R2) and both stators
are connected via S1, S2, and S3.

Assuming the rotors are in identical positions then the ac applied to the rotor
is acting like the primary of a transformer and emf’s will be induced into the
stator windings. These emf’s will depend on the rotor position.

As the rotors are in identical positions the emf’s in the TX and TR stator
windings are identical and no current flows between the two synchro stators.
This is known as the ‘NULL’ position.

The TX is usually mechanically coupled to an input shaft and the rotor of the
TR is coupled to a pointer of an instrument. A typical application on older
aircraft would be flap position indication. The TX being the flap position
transmitter near the flap and the TR being in the indicator on the flight- deck.

When the flap is moved (TX input shaft moves - say 20º) then the emf’s in the
TX stator will change due to change of rotor position. These emf’s are no longer
equal to the TR stator voltages, therefore current flows between the two stators
via S1, S2 and S3. This causes a resultant magnetic field in the TX and TR
stator’s.

The TX rotor cannot move as it is mechanically coupled to the flap. The TR

rotor is free to move and lines up with this resultant field (which is 20º from
the NULL position). As the rotor lines up with this field, the two rotors are
again in the same position, the emf’s induced in the stators are the same, no
current flows and the flap position indicator stops moving and now shows the
new flap position.

The effect of interchanging the rotor and stator connections are shown in
figure 44 (b, c and d). Note particularly that reversal of rotor connections (b)
produces an output that is 180º out from the input rotation. Changing over
any two stator connections gives reversed direction of TR rotation.

- 40 -
Fig. 46 EFFECT OF VARIOUS INTERCONNECTIONS
IN THE TORQUE SYNCHRO SYSTEM

- 41 -
It should be noted that the output of the TR produces a low torque, only
enough to drive a light pointer mechanism.

When the rotors are both aligned to their respective stators by a specified
angle then they are said to be at electrical zero. This standard angle is specified
by the manufacturer and allows replaced synchros to be matched to each
other.

CONTROL SYNCHROS

The control synchro system provides an error voltage signal in the receiving
element as apposed to the previous synchro which produced a rotor torque at
the TR so that is position followed that of the TX.

The error signal of the control synchro is an indication of the displacement of

the two rotor windings relative to each other. The control synchro system
comprises two synchros similar to the torque synchro, but called the Control
Transmitter (CX) and the Control Transformer (CT).

The CX is similar in construction to the CT except the CT windings are of high

impedance. The rotor winding is wound on a laminated cylindrical former and
fitted in slots that are skewed to prevent ‘lock-on’ effect.

The control synchro (figure 47) can be used to indicate control surface
position, flap position etc with the CT in the indicator, or used in a servo
system to provide a signal to a powered flying control surface
servomotor/hydraulic servo valve.

Fig. 47 CONTROL SYNCHRO

- 42 -
One of the two important things to note is that the ac supply is fed only to the
CX rotor and the discriminator amplifier, and the other is that in the NULL (no
error signal) position of the rotors are when they are at 90º to one another.

In the NULL position the ac voltage applied to the CX rotor acts like the
primary of a transformer and voltages are induced in the CX stator windings.
These voltages will be different, and this difference in potential will drive a
current through the CT stator windings. These currents produce magnetic
fields around each coil and the fields combine to form one resultant field. In
this NULL position this resultant field cuts the CT rotor at 90º so no emf is
induced into the CT rotor winding.

If the CX rotor input shaft is moved the voltages in the CT rotor will change,
due to change of rotor position. These new voltages will change the currents
flowing through the CT stator windings and also the position of the resultant
magnetic field.

Whatever angle the CX rotor is turned, the resultant magnetic field across the
CT will also move the same amount. As this resultant field is no longer at 90º
in the rotor but at an angle to it, then an emf is induced in the rotor winding.

This induced emf is fed to the discriminator amplifier where its phase
relationship is compared with that of the supply to provide the correct error
voltage and phase relationship to the two phase induction control phase and
drive the motor in the correct sense.

The motor drives through a gearbox to the indicator (in this example) and also
to the rotor of the CT. The motor drives the CT rotor until it is 90º to the new
field, no voltage will then be induced, the motor stops, and the indicator shows
the movement of the CX input shaft. The two synchro rotors are once again in
the NULL position ie 90º to one another.

Again, in this system if two of the stator windings between the CX and CT are
swapped then, for a clockwise input to the CX rotor, the output signal to the
indicator will be anti-clockwise, ie the indicator will show reverse indication.

QUESTION: What would be the effect of swapping the rotor leads R1 and R2 on
the:

a) CX rotor?
b) CT rotor?

If you cannot work this out contact your tutor (FS students).

‘Stick off’ Voltage

This voltage is applied to a CT to prevent the possibility of the servo system

aligning with a 180° error.

- 43 -
An example of this is described below which is similar to an altimeter fed with
fine and coarse signals from an Air Data Computer (ADC).

When the system is misaligned by a large amount the coarse synchro control
signal is fed into the servo amplifier (figures 48, 49 and 50). When the servo
motor has driven the output shaft sufficiently near to the line-up position, the
fine control signal is switched to the amplifier and the servo system is then
controlled by the fine synchros.

Fig. 48 COARSE/FINE SERVO SYSTEM

Fig. 49 GRAPH OF COARSE & FINE SIGNALS

If the coarse–fine ratio is even and there is an error of 180° the input to the
servo motor will be such that the coarse signal will try to drive the system back
to zero but the fine system will hold the system at 180° and the system will be
unstable at this point (figure 49).

- 44 -
To overcome the possibility of the system aligning with a 180° error a small
alternating voltage known as a ‘stick-off’ voltage is added to the coarse synchro
signal (figure 50), which will cause the coarse synchro signal to be displaced at
the 180° position so ensuring there will be no null point at 180°.

There will be no displacement at the 0° position so the system should align

correctly.

The coarse-fine system can be operated by a relay. The amplifier is connected

in the coarse synchro rotor line so that if the signal exceeds a pre-set value the
amplifier causes the relay to operate and so changes the signal to the servo
amplifier from the fine to the coarse control transformer.

Fig. 50 COARSE – FINE SERVO WITH ‘STICK OFF’ VOLTAGE

DIFFERENTIAL SYNCHROS

Used to produce an output that is proportional to the sum or difference of two

inputs. An additional synchro (TDX or CDX) is inserted between a TX and TR
in a torque differential system and between a CX and CT in a control
differential system.

The Torque Differential Transmitter (TDX) is similar in construction to the

Control Differential Transmitter (CDX). The rotors have three windings instead
of the usual single winding, spaced 120º apart wound on a slotted core
connected in the star configuration. One end of each winding is connected to a
slip ring on the rotor shaft.

The stator windings are the normal star connected windings at 120º to one
another. Figure 51 shows the basic construction. Figure 52 shows the wiring
diagram and figure 53 shows the system electrical schematic. The CDX has
higher impedance windings than the TDX.

- 45 -
 Fig. 51 ROTOR DETAILS

Fig. 52 STATOR/ROTOR COILS CIRCUIT & SYMBOL

The Torque Differential Synchro System

With reference to figure 53. Note the input shaft 1 is attached to the TX rotor
and input shaft 2 is connected to the rotor of the TDX. The TR rotor is
connected to a pointer, giving the sum or the difference between the input
shaft angles 1 and 2. Power supply is connected to the TX and TR rotors.

If input shaft 1 is rotated, say 15º clockwise and input shaft 2 is kept still, the
voltage changes in the TX stator windings to cause current to flow in the TDX
stator. This produces a field which cuts the TDX rotor and this induced emf
drives currents through the TR stator windings, producing a field which has
moved 15º clockwise and the rotor, free to move, will line up with this field,
indicating the movement of input shaft 1. So in this condition the TDX is
acting as a transformer.

Assume input shaft 1 is kept still and input shaft 2 is rotated 15º clockwise.
Moving the TDX rotor clockwise is equivalent to moving the field 15º anti-
clockwise. The induced emf in the rotor produces voltages that drive currents
through the TR stator that causes the field produced to rotate 15º anti-
clockwise and the TR rotor will line up with this, indicating the movement of
input shaft 2.

- 46 -
 Fig. 53 TORQUE DIFFERENTIAL SYNCHRO SYSTEM

For example, with reference to figure 54. If input shaft 1 (TX) is rotated 45º
clockwise and input shaft 2 (TDX) is rotated 15º clockwise then the output will
be the difference between the two inputs, ie 30º clockwise output of TR.

If the stator windings between TX and TDX are crossed (S1 to S3 and S3 to S1)
and rotor windings between TDX and TR (R1 to S3 and R3 to S1) are crossed
(figure 55) then 45º clockwise rotation of the TX rotor and 15º clockwise
rotation of the TDX rotor will give the sum of the two inputs at the TR
(60º clockwise).

Fig. 54 SYMMETRICAL CONNECTION

Crossing over (S1 to S3 and S3 to S1) between TX and TDX produces an output
that is the reverse of the sum of the two inputs (figure 56). Just crossing over
(R1 to S3 and R3 to S1) between TDX and TR produces the reverse difference
between the two inputs (figure 57).

- 47 -
 Fig. 55 TWO SETS OF INPUT & OUTPUT LEADS INTERCHANGED

Fig. 56 TWO TDX INPUT LEADS INTERCHANGED

Fig. 57 TWO TDX OUTPUT LEADS INTERCHANGED

The Control Differential Synchro System

This is similar to the torque differential system with the CDX between the CX
and CT, but the ac supply is fed to the CX only and the output is an error
signal proportional to the sum or difference of the two shaft rotations.

The operation is similar to that of the torque synchro system.

- 48 -
Electrical Supply

The supply frequency will be 400Hz and input voltages will be 115V or 26V
giving induced line voltages of 90V and 11.8V respectively and control
transformer maximum outputs of 57.3V and 22.5V respectively.

Fig. 58 CONTROL DIFFERENTIAL SYNCHRO SYSTEM

The following table gives a list of the possible faults and subsequent actions.

SYMPTOM CAUSE REMEDY

No movement of Power failure/OFF. Check appropriate fuse.
receiver. Switch ‘ON’.
Receiver displaced One pair of rotor Check rotors for correct wiring.
180º from transmitter. connections reversed.
Rotation correct.

Reverse rotation of 2 connections between Check stators for correct

receiver. stators reversed. wiring.
Receiver displaced 120º or 3 connections between Check stators for correct
240º clockwise from stators crossed. wiring.
transmitter. Rotation
correct.
Receiver displaced 0º, 60º, Short circuit between 2 Carry out insulation resistance
120º, 180º, 240º, 300º and stator lines. check between stator lines.
moves in 180º steps.

Receiver oscillates between Open circuit on one stator Carry out continuity check on
2 points approximately 75º line. stator lines.
apart.

TABLE 2 FAULT RECTIFICATION

- 49 -
 RESOLVER SYNCHROS

The function of a resolver synchro is to convert alternating voltages, which

represent the Cartesian coordinates of a point into a shaft position and a
voltage, which represent the polar coordinates of that point. From polar to
Cartesian an ac voltage is supplied to rotate the shaft θ degrees and it will give
two outputs r sin θ and r cos θ. It can also convert Cartesian to polar by
feeding in a sine and cosine voltage and it resolves them into a voltage and
shaft rotation.

An alternating voltage can be represented by a phasor line of length r and the

angle it makes with the x axis (figure 59).

Fig. 59 A PHASOR

This is called the POLAR co-ordinates r θ . This phasor can be split into its
two component parts (x and y) as shown in figure 60. Where x = r cos θ and
y = r sin θ.

Fig. 60 COMPONENTS OF A PHASOR

These are called the CARTESIAN co-ordinates of the phasor. (Named after
Rene du Perron Descartes French mathematician 1596 to 1650.)

Figure 61 shows the circuit of a resolver synchro. It is shown with two

windings on the stator at 90º to one another and two windings on the rotor at
90º to one another.

We shall now look at the resolver synchro operation in both modes.

- 50 -
 Fig. 61 RESOLVER SYNCHRO CIRCUIT

Polar to Cartesian

In this mode the rotor winding R3 R4 is shorted and an alternating voltage r is

fed into coil R1 R2 (figure 62). With the system stationary, S1 S2 winding is
directly opposite so maximum voltage is induced into it, while stator winding
S3 S4 is at right angles to this supply and no voltage is induced.

If the rotor is now moved to angle θ then the voltage in S1 S2 will fall and that
in S3 S4 will begin to rise and at 90º rotation the emf induced into S1 S2 will be
zero and that induced into S3 S4 will be maximum. If this rotation continued
then S1 S2 = r cos θ (cosine waveform) and S3 S4 = r sin θ (sine waveform) will
be formed.

Fig. 62 POLAR TO CARTESIAN CONVERSION

So by rotating the rotor to angle θ we have split the input voltage into its two
Cartesian components.

- 51 -
 Fig. 63 COS & SIN OUTPUT SIGNALS

Cartesian to Polar

Note here the cosine voltage (r cos θ) is applied to the stator winding S1 S2
while the sine voltage (r sin θ) is applied to the stator winding S3 S4 (figure 64).
These two voltages produce fields in the stators that combine to form one field
that cuts the rotor windings R1 R2 and R3 R4. The emf induced into R1 R2 is fed
to an amplifier.

Fig. 64 CARTESIAN TO POLAR SYSTEM

The amplifier output is fed to a servomotor that is mechanically coupled to a

load and the rotor of the synchro. So the rotor and load are turned when the
servomotor operates. When the rotor winding R1 R2 moves to 90º to the stator
field there is no emf induced into it and the servomotor stops.

The rotor winding R3 R4 at this time is aligned with the stator field and the
voltage induced will be the resultant of the input voltages Vx and Vy (figure 65).

- 52 -
 Fig. 65 PHASOR DIAGRAM

The voltage output and the angle the rotor has turned through represents the
polar co-ordinates.

Resolver synchros are used in aircraft instruments and navigation systems.

SLAB SYNCHROS

These are synchros consisting of wound stators and rotors (figure 66). Typical
uses are in gyroscopic instruments where the rotor is connected to one
component and the stator to another. For example to measure roll movement
in a vertical gyroscope, where the stator is mounted on the instrument case
and the rotor on the outer gimbal, so any roll movement moves the case
relative to the outer gimbal.

Fig. 66 TYPICAL SLAB SYNCHRO ROTOR & STATOR

THE SYNCHROTEL

This can be used as a low torque control transformer. Figure 67 shows the
construction and figure 68 shows a sectioned view.

The stationary rotor coil fits into the highly permeable case over the central
core and the conventional three-phase stator also fits over the core. The rotor
is an hollow aluminium cylinder with an oblique section that rotates in the air
gap between the stator and the coil. The rotor shaft is supported in jewelled
bearings.
- 53 -
 Fig. 67 SYNCHROTEL TX – EXPLODED VIEW

Because of the low torque required to operate the synchrotel they are ideal for
connection to lightly loaded systems such as capsule operated devices, eg
Pitot-static capsules. When used as a control transformer it will be connected
to a control transmitter as shown in figure 69.

Fig. 68 SYNCHROTEL TX CUT-AWAY

The synchro transmitter (CX) rotor is typically fed with a 26V 400Hz supply. It
will induce voltages in its stator and these are connected to the synchrotel
stator. The currents flowing through the windings (CX to synchrotel) produces
a magnetic field in the synchrotel that cuts the rotor - the rotor position having
been determined by the capsule movement (figure 69).

- 54 -
 Fig. 69 SYNCHROTEL PRINCIPLE

As the rotor of the synchrotel is around the core then a component of the flux
cutting the rotor will flow along the core and cut the rotor winding R1 R2. The
output from this winding is fed to the servo-amplifier where it feeds the control
winding of a two phase servomotor that drives the indicator and also the CX
rotor.

This changes the CX stator voltages and therefore the synchrotel stator
current and flux. This will continue until the synchrotel stator flux is no longer
producing an output from the rotor winding, ie the NULL position. The
indicator now shows a position that corresponds to the synchrotel input
position pressure applied at that instant.

When used as a transmitter the rotor winding is energised producing an axial

flux along the core which causes a current flow in the rotor. This rotor current
produces a radial flux that cuts the stator coils and induces an emf into them.
The emf induced is a function of rotor position determined by input movement
(eg pressure measurement).

INDUCTANCE TRANSMITTERS

These are operated by single phase ac and rely on induced emf’s being applied
to the transducer which can be varied depending on the transducer moving
element position. There are various ways of changing the induced emf.

THE ‘E’ & ‘I’ BAR TRANSDUCER

This type of transducer is used in older servo instruments, acceleration

sensors, air data computers, cabin pressure transducers etc.

- 55 -
It relies on varying the air gap between two coils with a central coil being fed
an ac single phase supply. The two outer coils are wound in series opposition
in the same circuit and the sum of their currents is used as the output.

The fixed laminated E shaped core has an ac supply connected to its central
limb (coil A), the outer two limbs (coils B and C) are wound with the coils
connected in series opposition. The laminated I shaped core (I bar) sits parallel
to the E core ( E bar) and is attached to whatever we are trying to measure the
movement of, eg in a servo altimeter the I bar is connected to the capsules.

Fig. 70 I BAR NEUTRAL

In figure 70 the I bar is parallel to the E bar and the ac is fed to the central
limb (coil A) creating a magnetic flux which will flow as shown. The flux in the
top and bottom limbs (coils B and C) will be the same, as the air gaps between
the I bar and E bar core are equal. Thus the emf induced into the two coils B
and C will be the same but of opposite phase, so the output will be zero.

Fig. 71 I BAR - POSITION 1

- 56 -
When the I bar is moved by the sensing element, as shown in figure 71, the air
gaps are now unequal. There is a higher flux density cutting coil B (because
the air gap is small) and less flux density cutting coil C (the air gap is larger).
The emf induced in coil B is greater than that in coil C. Because the coils are
wound in opposition the output is the difference between these two but ‘in
phase’ with the input. The amplitude of the output will depend on the amount
of movement of the I bar (refer to the graphs in figure 71).

Figure 72 shows the I bar moved in the opposite direction. In this case the emf
induced into coil C is greater than coil B and the output will be ‘anti-phase’ to
the input, again the amplitude will depend on the amount of movement of the I
bar.

Fig. 72 I BAR – POSITION 2

THE LINEAR VARIABLE DIFFERENTIAL TRANSFORMER (LVDT)

An LVDT consists of a primary coil of a transformer with the two secondary

coils connected in series opposition. In between the two coils is an iron core
that can be moved linearly by the system whose displacement is to be
measured.

Fig. 73 LVDT – CORE CENRAL POSITION

- 57 -
Figure 73 shows the LVDT with the iron core central, the outputs of the two
coils A and B will be equal but of opposite phase, the output of the LVDT is
therefore zero.

If the core is displaced as shown in the next diagram, the output of coil A
increases (iron core in-line with A and air gap small) and that of coil B
decreases (large air gap). The output is therefore the difference and is ‘in
phase’ with the input. The amplitude depends on the amount of displacement.

Fig. 74 LVDT – CORE POSITION 1

If the core is displaced in the opposite direction, as shown in the next diagram,
then coil B output increases and that of coil A decreases. Once again the
output is the difference between the two and is ‘anti-phase’ to the input. Again
the amplitude depends on the amount of displacement.

Fig. 75 LVDT – CORE POSITION 2

LVDTs are used for transducing linear movement such as pressure capsules,
Bourdon tubes, control surface position (attached to the servo actuator). They
are also used in Engine Pressure Ratio (EPR) transmitters, accelerometers and
some older fuel flow transmitters.

In some text books the Linear Variable Differential Transformer may be called
the Linear Variable Differential Transducer. There is some discussion as to the
correct title.

- 58 -
There is another version of this transformer called the Rotary Variable
Differential Transformer (RVDT) that works on the same principle, but as the
name implies the central core moves in a circular movement.

One of the uses is to measure control column movement.

Figure 76 shows the control column position RVDT of the B777. It is a Fly By
Wire system so the position of the control column must be accurately
measured for the data to be sent to computers.

A cable assembly connects the control column to the RVDTs so movement

rotates the central core of the RVDT’s to give an output to the roll control
computers and the spoiler control module.

Fig. 76 B777 FBW SYSTEM – CONTROL COLUMN RVDTs

CAPACITANCE TRANSMITTERS

A capacitor is basically two conducting plates in close proximity to each other

with a dc potential between them (revision from module 3). The capacitance (a
charge between the plates) of a capacitor depends on:

1. The dielectric constant of the material between the plates.

2. The distance between the two plates.
3. The area of the plates.

- 59 -
The relationship between the factors affecting capacitance and the capacitance
of a capacitor can be expressed as follows:

kA
C=
d

Where C = capacitance
k = dielectric constant
A = area of plates opposite each other
d = distance the plates are apart

Dielectric Change Method

In general the constant for a substance cannot be changed but if the dielectric
it-self was moved so that less (or more) of it was between the plates then we
have a means of varying the output from the system.

The most common use of capacitance transmitters using this system is in fuel
quantity indication systems.

The fuel tank sensor has two concentric metal tubes each connected
electrically to a circuit with the fuel being allowed between them. The fuel
being the dielectric. As the distance between the plates and the area of the
plates is set the only variable is the dielectric constant of the material (fuel
and/or fuel/air depending on fuel level) between the plates.

Fig. 77 SIMPLE CAPACITIVE TANK UNIT

As the fuel level rises in the tank, air is displaced by fuel and the dielectric
changes to increase the capacitance of the unit. Thus the change in the
capacitance is related to fuel quantity. As the fuel level goes down so the
capacitance also goes down.

Capacitive tank units are normally calibrated at manufacture and as such

require no on-wing calibration.

- 60 -
Air has a dielectric constant of 1 and aircraft fuel has a dielectric constant of
approximately 2. This means whatever the capacitance was with air as the
dielectric, when fuel replaces the air the capacitance will approximately double
from, say, 100pF to 200pF. This means that when the tank is being filled with
fuel the dielectric changes gradually from air to fuel increasing the capacitance
as the fuel level increases.

This same principle is used for measurement of hydraulic fluid level in the
hydraulic reservoir.

Fig. 78 SIMPLIFIED FUEL LEVEL MEASURING SYSTEM

Air Gap Method

A type of transducer that changes the gap (instead of the dielectric) between
the plates is shown in figure 79. This particular unit is used in digital air data
systems but could be used in any appropriate system. It is made by depositing
a small metallic film on the fixed part of the unit and also on the centre of the
diaphragm. The two ‘plates’ form a capacitor and are connected to a low
voltage circuit.

Fig. 79 CAPACITANCE TRANSMITTER

- 61 -
As pressure is applied to the diaphragm the two metallic films come closer
together changing the distance between the ‘plates’ and changing the
capacitance. This forms part of a capacitive bridge network that will change its
output depending on the pressure applied. These are used as pressure
transducers in some digital air data computers.

Plate Area Method

These consist of a set of fixed (stator) and a set of movable (rotor) vanes
usually made of aluminium. The set of rotors are caused to move in and out
between the stators, without touching them. This changes the effective area
between the rotors and stators (A in the equation) and therefore changes the
capacitance.

They usually have air as the dielectric and the range is typical from 50ρF to
500ρF.

Figure 80 shows a manually operated moving plate capacitor. When the rotor
control is moved so the rotors rotate between the stators to increase, or
decrease, their surface area opposite each other.

Fig. 80 PLATE VARIABLE AREA CAPACITOR

””””””””

- 62 -
- 63 -
- 64 -
- 65 -