This is a very large
article on 27MHz transmitters and receivers.
3 kits are available:
- 27MHz Field Strength Meter
- to test
your 27MHz transmitter
-
27MHz Transmitter - 2
Channel - transmits two different
tones on a 27MHz carrier
- 27MHz Receiver - 2 Channel -
receives and decodes 2 different tones
and we have 303MHz and 315MHz links. email
Colin Mitchell for details.
-
FOX HUNT
Where do you
start?
Read all this this article then
build the
27MHz Field Strength Meter kit: $6.50 plus $4.00 post
and 27MHz Transmitter - 2 Channel.
kit: $6.50 plus $4.00 post
then 27MHz
Receiver - 2 Channel.
kit $10.50 plus $4.00 post
What will I learn?
You will learn about designing 27MHz stages -
transmitters and receivers
You will learn about 0SCILLATORS - low frequency and 27MHz
You will learn about TUNED CIRCUITS - how they produce an
enormous waveform when operating at a special frequency.
|
CONTENTS
27MHz transmitter with crystal
- 1-transistor with crystal
Receiver for transmitter
27 MHz transmitter without Xtal
- Very simple 2-transistor circuit - produces tone.
Receiver for Transmitter
27 MHz transmitter without Xtal
- with multivibrator to produce tone.
27 MHz transmitter without Xtal
- with multivibrator to produce tone - improved circuit
Receiver for transmitter
2 Channel 27MHz transmitter without Xtal
Receiver for 2 Channel transmitter
4 Channel 27MHz transmitter with crystal
4 Channel
29MHz Transmitter with Xtal
4 Channel
Receiver
Field Strength MetersIn this
discussion we cover 27MHz transmitters and receivers
as found in remote control cars, aeroplanes, walkie talkies and some of the
older-style garage door openers.
We have provided a number of circuits so you can work out the best type
for your application and these circuits will also help you understand
which components are critical and which components can be changed.
It's a matter of looking at each circuit and seeing the general layout,
and comparing it to the other circuits. In this way you are building up
a concept of "building blocks" and this is the basis to learning
electronics.
Talking Electronics does not provide any kits for these circuits as the
products (toy cars, wireless doorbells etc) are readily available in toy
shops, hobby shops and many of the $2.00 "junk Shops." You cannot buy
many of the special components and the cost of the completed item is
less than buying the components!
WHERE DO YOU START?
Surprisingly, the place to start is to build a piece of test equipment
called a Field Strength Meter. We will describe 3 different Field
Strength Meter circuits. The simplest circuit detects RF from the
antenna of a transmitter and provides a reading on a multimeter.
The other two circuits have a tuned front end and as the capacitor is
adjusted, the meter increases to a peak then drops as the frequency is
passed.
For experimental purposes, it is not necessary to be on the 27MHz band,
but it is handy to know if your transmitter is producing an output.
You need a starting point and you either need a Field Strength Meter
or Radio Controller car or Walkie Talkie.
You need something to make sure your transmitter is transmitting. It's a
Field Strength Meter.
MAKING A TRANSMITTER
This discussion covers a number of transmitters and
receivers. The different circuits show you what is absolutely needed and
how different engineers think. Some circuits will work better than
others and some have unnecessary components.
Then you will need a receiver.
Some circuits are 1-channel, some are 2-channel and some are 4-channel.
Choose the tone that suits your requirement.
First we will go over some background.
Let's start:
6 bands (or frequencies) were allocated for the 27MHz band,
Channel |
Frequency
|
1
2
3
4
5
6 |
26.995
27.045 27.095 27.145 27.195
27.255 |
and these were very popular for
transmission - especially in countries where transmitting was strictly
controlled.
Both
27MHz and 49MHz circuitry produced very low cost devices and they are
still available. But you must be careful as some of the latest types are
much more sophisticated (and sometimes cost less than the older types).
We will
investigate how they work and how they can be modified.
Very little is available on how these circuits work and this article
will cover the "building blocks."
When we use the term "building block" we mean a group of components
making up a circuit that carries out a particular function and
can be connected to another circuit to achieve a final result. In this
way you can create your own project without having to
design each of the sections. A typical example is the 5-channel remote
control circuit we have modified to produce an on-off action from
two of the outputs. You can build these circuits from scratch, but why re-invent the wheel? If you want a 27MHz
or 49MHzlink, the best idea is
to buy a toy and modify it.
If you want voice communication, get a walkie talkie. If you want a
single on-off operation, get a remote control car.
Some remote control cars have up to 5 channels and sell for less than $20.00.
You can get everything you need on 2 printed circuit boards, ready for
modification, without having to source the components.
Look for 4 function models that require 3v operation for both remote
and receiver. The fifth function is "turbo" and is not used in some of
the designs. The photo below is the 4(5) function 27MHz remote
control car we discussed above:

5 channel remote
control car as discussed
in the text. It uses only 4 of the 5 channels.The first two circuits (figs 1& 2) form a single-channel
transmitter-receiver
link.
The second receiver (fig 7) uses a split supply to power
a motor in the forward and reverse direction (it uses the same
transmitter
as shown in fig 1).
The third transmitter & receiver, (figs 12 & 22) is
a multi-channel design, with a chip in the receiver. Then
we cover a 27MHz walkie talkie. This is a 4 transistor model. It uses
the same type of super-regenerative front-end as our receiver circuits
and injects Amplitude Modulated (AM) audio onto the signal. The result
is a very noisy transmission but a very effective way to achieve both
transmission and reception with the minimum of components. Most of
the parts have a dual function, operating in both transmit and
receive mode. This makes the circuit very efficient, component-wise.
Before we start, some of the Japanese transistors have
either a very high frequency capability or a very high collector
current. These transistors need to have an equivalent for the circuit to
work successfully. Here is a list of some of the type you will come
across and some equivalents:
Type: |
|
Gain:
|
Vbe
|
Vce
|
Current |
Case |
2SC3279 |
NPN |
140 to
600
@0.5A
|
0.75v |
10v |
2amp |

|
BC327
BC328 |
PNP |
60
@300mA
|
0.7v |
45v
25v
|
800mA |
|
BC337
BC338 |
NPN
|
60
@300mA
|
0.7v |
45v
25v
|
800mA |
BC547
BC548
BC549 |
NPN |
70 @100mA |
0.7v |
45v
30v
30v
|
100mA |
|
BC557 |
PNP |
|
|
|
100mA |
C945 |
NPN |
|
|
50v |
150mA |
 |
1815 |
NPN |
|
|
50v |
150mA |
|
|
|
|
|
|
|
|
|
|
|
|
 |
8050 |
NPN |
|
|
10v |
1.5A |
8550 |
PNP |
|
|
10v |
1.5A |
9012 |
PNP |
|
|
|
500mA |
9013 |
NPN |
|
|
|
500mA |
9014 |
NPN |
|
|
|
100mA |
9015 |
PNP |
|
|
|
100mA |
9018 |
NPN |
700MHz |
|
100mA |
|
27MHz TRANSMITTER with
Crystal
Fig 1 shows a simple 27MHz transmitter producing a carrier.

(receiver for this circuit
HERE)

The 27MHz transmitter PC
board

Here is the circuit made by Lucian Papadopol
iz6nnh@gmail.com
He has created capacitor values by paralleling two values.
This
means it produces an unmodulated 27MHz signal and when picked up by
a receiver, such as shown in fig 2, the result is a clean, noise-free
reception. To increase the output of the transmitter, the 390R resistor
is replaced by a 220R. This increases the current from 7mA to 12mA. The
resistor could be decreased to 150R for more output.
Page 2 of this article covers test
equipment that can be used to detect the output and the frequency of
transmission.
MAKING A 5-CHANNEL
TRANSMITTER/RECEIVER
This 5-Channel 27MHz link has 5 outputs. The
output goes HIGH when the corresponding button on the transmitter is
pressed. You can use 2,3 4 or 5 of the channels. We chose 5-Channels
as it uses the same number of components as a 2, 3 or 4 channel
design.
We will cover the design in detail:
1. The
Transmitter
2-Ttransistor, 5-tone (200Hz
/ 1kHz) transmitter
2. The Receiver "front-end"
3. The receiver
But first:
THE FIELD STRENGTH METER
Before building the
Transmitter/Receiver you need to build a Field Strength Meter
so you can test the transmitter.
Here is the circuit for Field Strength Meter MkV.
It can be constructed on a small piece of matrix board or simply soldered together
and connected to a 0-500uA movement. A "Movement" is similar to a
"Panel Meter." You can use 0-1mA movement or the 0.5mA current-range
on an analogue multimeter.
This is simply an "untuned" RF detector to prove the transmitter is
producing a signal.

Field Strength Meter MkV
The Field Strength Meter circuit is simple but it is designed
to detect the 27MHz transmitter in this project. The "lead" is flex
or enamelled wire 10cm long and wrapped around the antenna of the
transmitter for 2cm to get RFenergy-transfer. Our prototype
transmitter produced about 25% scale deflection.
1. THE TRANSMITTER
The transmitter consists of two transistors. The first transistor
produces the tone in conjunction with the second transistor and the
second transistor produces the 27MHz signal. Button "A" produces a
200Hz tone button "B" produces 1kHz tone, button "C" produces
1kHz tone
button "d" produces 1kHz tone
button "E" produces 1kHz tone

5-tone 200Hz to 3kHz
transmitter
The second
transistor is a self-contained oscillator and it gets its feedback (to
oscillate at 27MHz) from the transformer. The main coil is the 9t
section and the feedback to the base is 4 turns.
Nothing happens until one of the buttons is pressed as the first
transistor is held in a "turned-off" state by the 3M3 and the second
transistor is not turned on as the base and 2k2 are not connected to
anything.
When a button is pressed, the 4n7 starts to charge via the resistor
connected to the button and the first transistor starts to turn on.
The 4n7 gets
charged to a voltage that allow the first transistor to turn on
slightly. This allows current and voltage to flow through the 2k2 to
turn on the second transistor and produce 27MHz. A
voltage-drop is produced across the 100R load resistor and this
pushes the left-side of the 4n7 DOWN. The right side moves down and
it feeds some of its energy into the
base of the first transistor to turn it on MORE.
This causes the second transistor to turn on more and create a higher
amplitude. After a very short period of time the energy from the 4n7
(only a small amount of energy is delivered and the amount can be
worked out by knowing how many millivolts gets generated across the
100R between normal operation of the second transistor and its
higher turned-on state) has been fully delivered and and the first
transistor starts to turn off. This causes the second
transistor to turn off slightly and the 4n7 "rises in the circuit."
At the same time it gets charged again by the resistor connected to
the switch and the
cycle repeats.
The end result is a fairly brief pulse that causes the second
transistor to create a larger amplitude. This is detected by a
receiving circuit as a higher amplitude 27MHz
signal produced 200 times or 1,000 times per second. This is called
an AMPLITUDE MODULATED SIGNAL and in this case it shown on a CRO as
peaks or spikes and in a speaker as a buzz or tone.
The frequency of the
tone is determined by the value of the 4n7 and the resistor that charges
it.
2. THE RECEIVER FRONT
END
The front end of
a 27MHz receiver is slightly different to the front end of a 27MHz
Walkie Talkie. The circuit is very lightly loaded so that it will
detect the slightest signal and this makes it very sensitive.
The component that does this is the 3k9 in the power rail.
The transistor has only about 2v across it and takes less than 1mA.
The circuit is a common-base amplifier and under certain circumstances,
a single transistor in this configuration will oscillate.


The receiver "front-end" put together for experimenting
First, let's describe how a common-base stage works.
The 220k base-bias resistor turns the stage on. The coil in the
collector is the load but since it is such a low resistance, a
voltage-dropping resistor has to be placed in the emitter. This resistor
only limits the current through the circuit.
It is by-passed with a 3n3 capacitor and does not cone into the
operation of the operation of the circuit.
There are two ways to turn on a transistor.
one is to hold the emitter fixed and deliver a voltage and current to
the base. The other is to hold the base rigid and reduce the voltage on
the emitter.
that is what we have done in this circuit.
The 4u7 electrolytic on the base holds it rigid and the 39p reduces the
voltage on the emitter to turn the transistor ON and during the other
part of the 27MHz cycle, it increases the voltage on the emitter to turn
it OFF.
This concept must be
understood before we can advance further.
The circuit starts by
receiving a pulse of current through the 6t coil and 47p capacitor, when
it is turned on. These two components form a TUNED CIRCUIT and
when they receive energy, they produce a sine-wave (sinusoidal) waveform
that appears on the lower part of the tank circuit and this is passed to
the emitter via the 39p. It's an amazing fact that two simple components
can produce a sinewave that has an amplitude larger than the voltage
applied to the pair, but this is what happens and to fully explain it,
would require another chapter.
However the end result is the emitter is pushed down to turn the transistor
ON more and pulled up to turn it off. The 70t inductor on the emitter
simply keeps it away from the 0v rail so that the pulses on the
39p can have an effect on pushing and pulling the emitter.
To explain how the inductor works is very difficult, however it has the
effect of allowing the 39p to push the voltage lower on the emitter,
than if it were not included. If
the inductor is removed, the 39p will have a very hard job to push
against the 3n3. But if the 3n3 is removed the 39p will have little
difficulty pushing against the effect of the 680R.
However when the inductor is added, the 39p has a much easier job of
pushing and pulling the emitter.
Even though this is the front end of a receiver, it is actually
oscillating at 27MHz.
The circuit has been built for experiential purposes to make sure the
70t inductor and 6t coil work perfectly to price a sensitive front-end.
Never put a circuit on the web without building and testing it with
actual components. Simulation programs cannot predict the outcome of
hand-made components.
The 70t inductor is about 7uH and is wound on a 1M resistor with
very fine wire. The resistor can be any high value and does not play a
part in the functioning of the circuit. It is just a convenient core for
the winding. The carbon content of the resistor does not have any effect
on the inductance.
3. THE RECEIVER
HOW THE RECEIVER WORKS
The receiver is a super-regenerative design and the output is very
noisy. However when a
signal of the same frequency as the super-regenerative circuit surrounds the antenna, the circuit has difficulty radiating a signal
and it takes more current and less current. These variations
appear across the 3k9 load resistor as a change in voltage (a waveform) and the
signal is picked off via a 100n capacitor and passed to a filter stage
that removes most of the background noise and amplifies the signal
(tone).
The following diagram shows just some of the stages needed to decode 5
different tones and deliver the signal to 5 separate outputs:

Block diagram of the 5-Channel Receiver
It would require a minimum of 20 transistors to carry out this
requirement and all this has been done in a single 8-pin chip by the
author, as shown in the following circuit:

5 CHANNEL RECEIVER using H-bridge transistors
The 5 CHANNEL RECEIVER uses a
TE5CHRx chip from Talking Electronics. Click
HERE to buy the chip.

Outputs are normally NOT HIGH or LOW but HIGH IMPEDANCE when not
activated and this
effectively makes them "out of circuit" as far as the other
components are concerned.
Outputs A and B produce FORWARD and REVERSE.
When button "A" on the transmitter is pressed, output A
goes HIGH and B goes LOW.
When button "B" on the transmitter is pressed, output A goes LOW and B
goes HIGH.
Outputs C and D produce LEFT and RIGHT via an actuator (or a motor).
When button "C" on the transmitter is pressed, output C goes HIGH and D
goes LOW.
When button "D" on the transmitter is pressed, output C goes LOW and D
goes HIGH.
When button "E" on the transmitter is pressed, output E goes HIGH.
The outputs from the chip drive an "H-Bridge" and each transistor is
actually an EMITTER-FOLLOWER.
This means the bridge cannot be connected to a supply higher than the
driving circuit.
The voltage drops across the driving components is nearly 2v. This means
the motor will see a maximum of 4v and that's why a 3v motor has been
suggested.
The chip delivers about 25mA max into the base of each transistor and
depending on the gain of the transistor, it will supply about 500mA max.
Output 5 is limited to 25mA.
The H-Bridge section can be designed with an L2930, to take the place of
8 H-bridge driver transistors. The chip has built-in diodes and will
deliver up to 600mA per output.

5 CHANNEL RECEIVER using
L293D driver IC |

(transmitter
for this circuit HERE)
In the circuit above, when the transmitter is off, the car moves forward. When the transmitter
is on, the car reverses and moves in a circular pattern due to the fact
that the front wheels steer straight when the car is moving forward but
turn left when the car moves backward. This allows
the operator to guide the car around obstacles.
It's a very awkward way to control a car and although it is very simple
and clever, it is not really successful in practice.
We will not be going into the mechanics of how the car steers, only
the fact that the transmitter causes the motor to reverse direction.
In place of the motor you could use a relay or two separate motors
to carry out a number of functions and we will show how the circuit
can be modified to do this.
The receiver works on a "tone," "no-tone" principle but the transmitter
doesn't actually send a tone as this would require additional circuitry.
What happens is the receiver picks up random noise from the airwaves
when the transmitter is not operating and this functions as the tone
part of the reception.
This random noise is amplified by the second transistor and passed
to a 0.47u electrolytic that keeps the third transistor in conduction
for the majority of the time. The operation of this will be discussed
later.
The 10u on the output of the third transistor keeps the output low
for the short periods when the third transistor is not low.
The motor is connected in a bridge formation via four transistors
and these change the polarity of the supply to the motor.
When the transmitter is operating, and the receiver is within range,
it picks up a 27MHz carrier that over-rides the random noise and
produces
a CARRIER.
This means the second transistor will not see any noise and thus the
0.47u electrolytic will charge and turn off the third transistor.
The 10u will charge via the 2k2 and the input to the bridge will change
from a LOW to 0 HIGH.
This will turn on the opposite half of the bridge to supply current
to the motor in the reverse direction.
Now we will cover the circuit in detail.
HOW THE TRANSMITTER WORKS
The transmitter is a very simple crystal oscillator. The heart of
the circuit is the tuned circuit consisting of the primary of the
transformer and a 10p capacitor. These two components oscillate when
a voltage is applied to them. The frequency is adjusted by a ferrite
slug in the centre of the coil until it is exactly the same as the
crystal. The crystal will then maintain the frequency over a wide
range of temperature and supply voltage fluctuations.
The transistor is configured as a common emitter amplifier. It has
a resistor on the emitter for biasing purposes but the 82p across
the 390R effectively takes the emitter to the negative rail as far
as the signal is concerned.
The 390R resistor prevents a high current passing through the transistor
as the resistance of the transformer is very low.
The tuned circuit operates at exactly the third harmonic (also called
the third overtone - an overtone is a multiple of a fundamental
frequency)
of the crystal so that the crystal will oscillate at its third overtone
(27MHz) and in-turn, keep the frequency of the circuit stable.
The transformer in the collector of the transistor performs two
functions.
1. It matches the impedance of the transistor to the impedance of
the antenna, and
2. Creates a resonant circuit at 27MHz to make sure the crystal
oscillates
at this frequency.
You can see the transformer creates a resonant circuit by the fact that
it has a capacitor across the primary winding. These two components
create a "resonant" or "tuned " circuit and this is where the circuit
"gets its frequency."
The crystal has a fundamental of about 9MHz and it will oscillate
at this frequency unless assisted to oscillate at a higher frequency.
This is done by the tuned circuit oscillating at 27MHz.
Now we will look at the impedance-matching feature of the transformer.
The impedance of the output of the transistor is about 1k to 5k and
this means it is the impedance (resistance) "it works at." In other
words, it is the characteristic impedance of the transistor in this type
of stage. The impedance of
a whip antenna is about 50 ohms and the transformer matches these
two by having a TURNS RATIO.
The primary has about 12 turns and the
secondary about 3 turns.
This provides part of the matching requirement. The `pi' network,
made up of the 150p, 15 turn air-cored coil and 100p capacitor assists
further in matching the output of the transformer to the antenna.
When the power is applied, the transistor turns on fairly hard due
to the 82p in the emitter being uncharged.
This puts a pulse of energy
through the 10p and as the transistor turns off slightly due to the
82p charging, the energy in the 10p capacitor is passed to the primary
of the transformer to start the 27MHz cycle.
The action of the emitter rising and falling during start-up, allows
the base to rise and fall and this puts a pulse on the crystal to
start it oscillating.
The frequency of oscillation OF THE CIRCUIT is generated by the tuned
circuit in the primary of the transformer and the crystal merely keeps
the circuit operating at exactly 27.145MHz (or 27.240MHz, depending
on the frequency of the crystal).
The turns ratio of the transformer converts a high voltage waveform
(that has little current) from the transistor, into a low voltage
waveform with a higher current.
This is exactly what the antenna requires. But before the signal passes
into the antenna it goes through the pi network, then an 8 turn Radio
Frequency Choke. This is 8 turns of enamelled wire wound on a ferrite
core and is called a base-load for the antenna.
The result is a 27MHz frequency called a CARRIER.
The carrier produces a clean spot on the band that is free from
background
noise.
The 22n across the first stage is designed to remove the high-frequency
component from the waveform. If this were not present, the circuit would never change state.
The receiver is tuned to the frequency of the crystal in the transmitter
via a slug-tuned coil in the collector.
When the transmitter is off, the receiver picks up background noise
and amplifies it to produce random-noise. This is amplified by the
second transistor and passed to the third via a 0.47u electrolytic.
This electrolytic is designed to keep the third transistor ON for
the major part of the time and it does this in a very clever way.
We will assume the supply has just been turned on and the second
transistor
is not receiving a signal. The 0.47u will be uncharged and it will
charge via the 10k collector resistor and the base-emitter junction
of the third transistor.
The action of the current flowing through the base of the third
transistor
will turn it ON but after a short time the electrolytic will be fully
charged and the current will cease and the transistor will turn off.
A 10u on the collector of the third transistor will then begin to
charge via the 2k2 resistor and after a period of time called the
DELAY TIME, the output will be HIGH and change the state of the bridge.
But if a signal is present on the collector of the second transistor,
(in our case this will be background hash), the voltage on the collector
will be rising and falling. When the voltage goes low, it takes the
positive end of the 0.47u low and the other end must follow.
The voltage
on the negative end will go below the negative rail and at -0.7v it
gets clamped by the diode. This means the electrolytic gets discharged
very rapidly when the second transistor turns on.
The result is the electrolytic takes a long time to charge and a short
time to discharge, even when random noise (hash) is being processed.
The action of the 0.47u is amazing and will be explained in more detail
in a moment.
During the short periods of time when the third transistor is not
turned on, the 10u on the collector will take over and hold the signal
low.
It's only when a long duration of silence is encountered, that the
circuit will change state. This period of silence is when the
transmitter
turns ON and the time is very short in real terms.
Transistor Q3 is called the switching transistor. It changes
between HIGH and LOW to create the forward and reverse direction.
The switching transistor feeds two driver transistors, Q4 and
Q9. Each of these drives two output transistors. Q4
drives Q6 and Q7. Q9 drives Q5 and
Q8.
Follow these transistors on the circuit and you will see how the supply
is directed to the motor, firstly in one direction and then the other.
The printed circuit board is quite complex because of the number of
driver transistors. But since these cost less than 2 cents when bought
in the million, it is not cheaper to use a chip.
HOW THE
0.47u WORKS
The 0.47u electrolytic on the base of the third transistor needs
explaining
as its operation is very clever.

Charging the 0.47u
electrolytic
is represented as a battery.
The electrolytic is simply a tiny re-chargeable battery
and when the circuit first turns on, it is uncharged.
The charging current passes through the base-emitter junction of
the third transistor and keeps it ON as shown in fig: 3.
If the electrolytic is allowed to fully charge, the current will fall
to zero and the third transistor will turn off.
But the second transistor discharges the electrolytic quickly before
it has time to fully charge.
It does this by turning ON.
How the electrolytic discharges is shown in fig: 4. The only components
involved in the discharge are Q2 and the diode.
Transistor Q2 is turned on and it will have zero volts (0.3v)
on the collector.

Discharging the 0.47u
electrolytic.
This means the positive lead of the electrolytic
(equivalent to the positive terminal of the battery) will drop from
say nearly 3v, to 0.3v. The negative lead must follow and normally
it would be at -2.7v. Yes, the negative lead would have a negative
voltage on it relative to the 0v rail, if the diode was not present.
BUT the diode on the negative lead gets turned on as soon as the voltage
on the negative lead falls to -0.7v and prevents it going below -0.7v.
As the positive lead falls, the energy in the electrolytic is quickly
discharged through the diode and when the second transistor turns
OFF, the electrolytic is ready for charging, through the 10k resistor.
LOW RAIL VOLTAGE
One of the problems of a low rail voltage is the voltage lost across
each of the output transistors. Each drops about 0.5v across the
collector-emitter junction and this leaves only about 2v for the motor.
However the supply voltage must not be increased above 3v as there
is a very short period of time when the circuit is changing from LOW
to HIGH and both halves of the bridge are ON. This is at the mid-point
of the change-over and if you work out the various voltage drops across
the base-emitter junctions, it leaves about 0.2v for the two 1k
resistors.
With a 3v supply, the base current is limited to 0.1mA by the inclusion
of the two 1k resistors and 10mA for the collector-emitter current.
But if the voltage is increased above 3v, the current will increase
dramatically and the transistors will be damaged.
CONNECTING A RELAY
Fig: 5 shows how a relay can be connected to the driver transistor
to operate when the transmitter is switched on. The change-over contacts
on the relay can be used to power any device when the transmitter
is off or when it is on.

Connecting a relay to the
driver
transistor. The supply for the relay
can be 6v - 12v.
CONNECTING TWO MOTORS
Fig: 6. shows how to connect two separate motors to the circuit. The
motors can be connected to any voltage from 3v to 12v and the direction
of rotation will depend on which way around they are connected, but
transistors Q4and Q7 should be kept at 3v - especially
Q9, as it cannot be taken to a voltage higher than 3v, due
to the way it is connected in the circuit.

Connecting two motors to
the outputs.
A SPLIT-SUPPLY RECEIVER
The second receiver circuit we will study uses more components to
do exactly the same job but it may have better sensitivity due to
the inclusion of one extra stage of amplification and the use of a
higher rail voltage.
The higher rail voltage gives some stages a higher gain due to the
higher amplitude of the signal.
But some of the gain has been lost in the diode pump as this type
of pump requires more energy to charge the 10u than a 0.47u.
The use of a center-tapped voltage source saves two transistors in
the bridge network but necessitates the use of a double-pole switch
to disconnect both halves of the supply.

A 27MHz receiver using a
split supply

The 27MHz receiver PC
board
HOW THE SPLIT-SUPPLY RECEIVER WORKS
The operation of the front end of the split-supply receiver in fig:
7 is identical to the receiver shown in fig: 2.
The use of a PNP transistor for Q1has simply turned the circuit
up-side-down however the antenna is still connected to the collector
and the parallel tuned circuit is also on the collector.
The circuit is turned on by the 33k on the base and the 47n keeps
it rigid and turns the stage into a common base configuration.
The parallel resonant circuit made up of the 8-turn inductor and 15p,
starts the circuit oscillating and the 39p between collector and emitter
provides feedback for the transistor to supply pulses of energy to
the tuned circuit to keep it oscillating.
The 220R and 39p are the emitter biasing components, as well as the
390R, 10n and 47n.
The 100R and 47u are stage-separating components to remove low-frequency
noise from the power rails and the 22n across the first stage tightens
up the power rails as far as the high frequency is concerned and
allows the low-frequency component to appear across the 3k3.
The signal across this resistor is picked off via the 10k/39n
combination
and passed to two stages of amplification.
The 10k and 4n7 form a filter to remove high frequency pulses.
A high frequency pulse will try to charge the 4n7 and most of the
amplitude of the pulse will be lost (attenuated) in the 10k resistor.
Exactly how this works is as follows:
The high-frequency pulse will rise and fall before the 4n7 has time
to charge. But a low-frequency will charge the 4n7 and enter the 39n
for amplification by the rest of the circuit.
Going back to the first stage, we have already mentioned that it is
oscillating at 27MHz and the MOST ACTIVE lead of the circuit is the
collector and this is where the antenna is connected.
The waveform produced by the circuit is passed to the antenna and
radiated to the surroundings.
Any other signals of the same frequency will interfere with the
circuit's
ability to radiate energy and this is reflected down the antenna to
the first stage.
The result is it takes slightly more and less current according to
the intelligence on the signal.
The word intelligence means the information that has been added to
the carrier.
For a transmitted signal this means voice or music etc.
When no transmitted signal is present this is background hash or
"noise."
The changes in current will see a waveform develop across the 3k3
feed resistor. The 10k will detect this and pass it to Q2 for
amplification.
Q2 and Q3 amplify the low frequency (audio) or "hash"
component.
Any high frequency signals will be removed by the 270p capacitors.
They act as negative feedback devices and operates as follows:
A rising signal on the base of the transistor turns it ON and the
collector voltage falls.
The fall in voltage is passed through the 270p (because it does not
have time to charge) to the base where it counteracts the original
signal.
The capacitor ONLY has an effect on high frequency signals and the
low frequency signals are amplified without attenuation. A low-frequency
signal will charge the 270p and get lost in the 270p.
After two stages of amplification, the signal appears at a diode pump
made up of a 15n capacitor, two diodes and a 10u electrolytic.
The charging of the 10u takes quite a number of cycles as the 15n
is like a teaspoon filling a glass with water.
When Q3 turns off, the 15n is charged via the 4k7, D2
and the 10u.
The 15n doesn't take very long to charge and the current flowing through
it puts a tiny amount of charge into the 10u.
Transistor Q3 turns on and discharges the 15n through diode
D1 in exactly the same manner as explained previously.
When Q3 turns off, the 15n is ready to charge up again.
This keeps happening for hundreds of cycles, each time the voltage
on the 10u gets slightly higher.
At a voltage of 0.65v, the base of Q4 begins to turn on.
Below this value the base does not see anything, and does not have
any loading effect on the electrolytic. But at exactly 0.65v a tiny
amount of current begins to flow into the transistor to turn it on.
The electrolytic keeps charging and as the voltage rises to 0.66v,
0.67v, 0.68v, 0.69v, the transistor turns on more and more.
At 0.7v, the transistor is fully turned on and any voltage over this
simply spills into the base and is passed to the negative rail via
the base-emitter junction.
This means the voltage on the 10u does not rise above 0.7v.
To keep the transistor turned on requires a small amount of current
into the base and the electrolytic supplies this current.
In doing so, the energy in the electrolytic gets used up and the
voltage across it reduces.
As the voltage falls, the transistor gets
turned off. When the voltage drops below 0.65v, the transistor is
fully turned off and does not see any voltage below this.
This means the operating voltage for the electrolytic is between 0.7v
and 0.65v.
Q4 feeds Q5 and when Q4 is turned on, the voltage
on the base of Q5 is below 0.65v and it is turned off.
The 10u on the collector of Q5 charges via the 1k5 and when
it is above 3.7v, driver transistor Q6 turns on and output
transistor Q8 operates the motor.
There are two outputs. One drives the motor in the forward direction
and the other drives it in reverse.
THE TRANSISTORS IN THE FORWARD
DIRECTION
There are two transistors for the motor in the clockwise (forward)
direction, as shown in fig: 8.

You will notice the turn-on resistor(s) on the base of the driver
transistor is lower than for the reverse direction and this will allow
a greater current to be delivered to the motor to give it full speed
in the forward direction.
THE TRANSISTORS IN THE REVERSE
DIRECTION
There are 3 transistors driving the motor in the reverse direction,
as shown in fig: 9.

These are the switching transistor Q5,
the driver transistor Q7, and the output transistor Q9.
The reason why a driver and output transistor are need is to provide
a high current for the motor as it needs a high current at start-up
or when under load.
A motor may take only 50-150mA when not loaded but the current will
rise to 300-500mA when loaded.
It the motor does not receive this high current, it will appear the
car has no power.
For the output transistor to deliver this high current, the base must
receive a current according to the gain of the transistor.
The gain of a transistor varies enormously, depending on the current
flowing through the collector-emitter circuit. The DC gain of a
transistor
is generally specified as between 100 - 450, but this is under ideal
conditions and is determined at a collector current of about 1mA!
When the current is increased, the ability of the transistor to amplify
decreases. For a small signal transistor, this may decrease to a gain
of 75 for 50mA or as low as 10 or 20 for 250 - 500mA.
That's right,
the transistor may only have a gain or 10 or 20 when passing a heavy
current.
This means the base must receive a current of 25mA to 50mA to make
certain the transistor will deliver 500mA.
When the transistor turns on fully, the voltage between the collector
and emitter is only about 0.2v to 0.5v.
If the base is not supplied with sufficient current, the transistor
will not turn on fully and the voltage across the collector-emitter
leads may be 0.6v or higher.
This is how the transistor limits the current to the device it is
powering.
For our application we do not want any extra voltage to be lost across
the transistor and so it must be fully turned ON.
So we want the driver transistor to deliver 50mA.
This will be a low-current device and 50mA will be its maximum rating.
We can allow a gain of 100 for this device so that it requires a current
of 0.5mA into the base to turn it on fully.
The turn-on resistor is the 4k7 and when you take off the voltage
drop across the collector-emitter of the switching transistor and
the base-emitter junction of the driver transistor you have about
2v remaining from the 3v supply. This gives a base current of 0.4mA.
This is not enough to supply the motor with full current and thus
the motor goes slightly slower in the reverse direction.
THE ADVANTAGE OF A SPLIT SUPPLY
With the split-supply design there is no part of the cycle when both
outputs are on at the same time. This makes it a much safer design
than the receiver in fig: 2. The section of the circuit we are looking
at, to see if both outputs are on at the same time, is shown in fig:
10.

Determining if both
outputs are on
at the same time.
When the switching transistor (Q5), in fig: 7, is changing
from high to low, there is a gap of about 1.2v where both outputs
are off. Driver Q6 is tuned on when the input line is above
3.6v, and driver Q7 is turned on when the input line is below
2.4v.
SIMPLIFYING
THE SPLIT-SUPPLY
CIRCUIT
There are some unnecessary components in the circuit of fig: 7 and
by clever re-designing, these can be eliminated. This seems surprising
for a mass-produced item but sometimes the designer has not carried
out the final step of a design.
This is to look at each component and say "Is this part necessary?"
If you are not sure, remove it and check the operation of the circuit.
If the circuit operates ok, the component may not be necessary.
There are 10 components in the circuit of fig: 7 that can be removed
and a further 5 can be changed in value when a re-design is carried
out. The result is shown in fig: 11.

The 27MHz single-channel
receiver with the author's modifications.
The first two components to be removed are the 390R and 10n on the
emitter of the first transistor. The 220R is increased to 680R as
shown in fig: 11 to produce the same biasing.
The reason why the 10n can be removed is because it is effectively
across the 390R (via a 47n) so that the join of the 220R and 390R
is effectively at rail impedance to high frequencies.
This means the
39p can be connected to the positive rail and the 390R can be
incorporated
with the 220R.
By using 470p as the high frequency filtering component in each of
the two audio amplifier stages, the 10k and 4n7 filtering components
can be eliminated.
It may also be possible to remove one of the audio amplifier stages
when the 0.47u electrolytic is used as it is much more effective than
the 15n charging the 10u.
The 15n and one of the diodes is not needed when the charging
electrolytic
is 0.47u.
The switching transistor Q5 is not required, however it does
invert the signal so that when it is removed, the resistors to each
of the driver transistors must be changed so that the output driving
the car in the forward direction delivers full power and the reverse
output delivers about 80%.
2-CHANNEL LINK
$10.00 from Talking Electronics
A number of
2-Channel 27MHz links are available on the web as pre-built modules at
very low cost.
They are intended for remote control cars but the trend is now for 4 or
5 channel links, so these are available for $10.00. Email Colin
Mitchell to buy a Tx/Rx link
And get a
27MHz FSM to test the output for $6.50
You may want to control something at the far-end
of a model railway layout and running cables
may be practically impossible.
Or you may want to control something that moves around your layout.
This is an ideal way to solve the problem.
The range is about 10 metres (tested with 3v on the transmitter).
The modules come with whip antennas - spring-steel wires.


The resistors are
different to the markings on the board
The first surface mount transistor is "CR" (2SC945)
The second surface mount is an oscillator chip - not a transistor!!!
This saves about 8 components.

The diode simply prevents the 1kHz switch delivering
a voltage to the output of the IC. It does not have any effect on producing the
250Hz signal. It is just a "gating diode."
The 3-pin chip could be a very small microcontroller that has been
programmed to produce 250Hz when it detects a voltage on the output
pin. The appearance of the 250Hz indicates the output is generated
by a higher frequency and is divided-down.
With 15k, the output has some 1kHz segments in the high part of the
waveform when 250Hz is being delivered. With 1k5, the top of the
250Hz waveform is very smooth.
This IC replaces at least 10 components in a two-transistor
square-wave arrangement and this saves time, space and money.
The surface-mount transistor produces the 27MHz carrier when the 1kHz from the IC
is HIGH. This is because the 1kHz turns the transistor ON during the
HIGH periods of the waveform.
The oscillator produces a sinewave and the feedback is the 10p.
The 39p is effectively across both windings and this forms a tuned
circuit with a frequency of 27MHz.
A tiny amount of this waveform is picked off by the 10p and passed
to the base where the signal delivered to the base turns the
transistor ON more and more until it cannot turn on any more. At
this point in the wave, the signal via the
10p ceases and the transistor turns off. The collapsing magnetic
field delivers its energy to the 39p to charge it and this creates
the second half of the waveform. And the signal from the 10p is
opposite to previously and it has no effect at all on the
transistor. The transistor plays no part in this portion of the
wave. Adjusting the ferrite core changes
the frequency slightly and this is done when the transmitter is a
long way from the receiver so you can detects when the transmitter
is operating at exactly the correct frequency. You use the weak
reception to make the final adjustment.

The receiver board

The receiver is just two IC's and a few surrounding components



The two outputs can be used to reverse a motor or each output can be
used to turn ON a device.
When there is no transmission (reception) both outputs have zero
volts.
For Forward, one output goes high and the other goes low.
The voltage lost across the output FETs is only a few millivolts
(about 3 to 5mV).
The output FETs can handle about 200mA to 300mA.
Each output can be used to turn ON a separate motor:

FOX HUNT $10.00 from Talking Electronics
You can add a flashing LED across the first
push-button on the 2-channel transmitter shown above to produce a
beep-beep-beep transmission. Some 3mm flashing LEDs will work. Some
will not work. There is no way of telling which LED will work and
you have to try them. And they work perfectly.
ooooo00000ooooo
You don't have to buy these modules. You can use the
transmitter/receiver from a toy car that no-one wants any-more.
(some of them are 4 channel).
You can operate sound modules, lights, gates, points and anything up
to 6v and 200mA.
Every module is different with different circuitry and chips. This
discussion is just an example of how the link works.
The LC117A chip, TMX RX 3E chip and TX 3E IC are not available.
A
MULTI-CHANNEL
LINK
A multi-channel link is considerably more complex than a single channel
design but it offers the possibility of designing a project that has
more features.
The multi-channel transmitter shown in fig: 12 has forward, stop,
reverse as well as left, centre, right steering.

A multi-channel 27MHz
transmitter
This represents 6 channels and they are created by changing the
mark-space
ratio of a square wave oscillator as well as its frequency.
The photo shows the components on the PC board:

The 6 (4) channel
transmitter board
When the transmitter is not operating, the receiver picks up hash
(background noise) and no outputs are activated. This represents the
STOP function.
When the forward function is selected on the transmitter, the
square-wave
oscillator operates at its high frequency setting, with an equal
mark-space
ratio.
If left-turn is selected at the same time, the mark-space ratio is
altered to 1:3 while the frequency remains the same.
If right-turn is selected, the mark-space ratio is 3:1, with the same
frequency.
If the reverse function is selected, the frequency of the oscillator
is reduced to half and if the centre steering is selected, the mark
space ratio is 1:1.
If the left steering is selected, the mark-space ratio is 1:3 and
if right steering is selected, the mark-space ratio is 3:1.
To understand how the channels are produced, you need to know how
a multivibrator works.
HOW
A MULTIVIBRATOR WORKS
The multivibrator in the transmitter consists of transistors Q2,
Q3 and the surrounding components. This is shown in fig: 20.

You will notice the symmetry of the circuit and this produces an output
waveform
that is either HIGH or LOW. The circuit changes from one state to
the other very quickly and this produces the fast rise and fall of
the waveform and thus its square nature. The HIGH part of the waveform
is called the Mark and the LOW is the Space, as shown in fig: 15.
A square wave with a 1:1 output has the length of the mark equal to
the space.

For the transmitter in fig: 12, the output of the multivibrator for
the Forward function is shown in fig: 14.
We can take this as the reference waveform as all the other waveforms
will be a multiple of this.
For instance, if the left-turn is selected while in the forward
direction,
the waveform changes to that shown in fig: 15.
Note the short period
of time the waveform is HIGH compared to the LOW time. If this waveform
is passed into an integrating network, the percentage of time it is
high can be determined and an output activated. This is what the chip
does in the receiver.
It determines one of six functions and produces
outputs to steer the car in the left or right direction and/or drives
the car in forward or reverse. It also detects when the transmitter
is not operating and stops the car.
If the forward-and-right controls are selected the waveform is shown
in fig:16.
When reverse is selected, the multivibrator operates at half the
frequency
due to the 82k resistor added to the base of the two transistors in
the multivibrator.
The resulting waveform for reverse is shown in fig: 17.
If reverse-and-left is selected, the waveform is shown in fig: 18.
If reverse and right is selected, the waveform is shown in fig: 19.
THE TURN-ON
CIRCUIT
The transmitter doesn't have an on-off switch. It is turned on when
the forward-reverse control is moved from the stop position. This
switches a diode into circuit. The diode charges the 100u via the 4k7 to turn on the emitter-follower
transistor Q1. The base rises to just below rail voltage and
the emitter is about 0.7v below this.
The emitter becomes the power rail for the rest of the circuit and
while the controls are in the forward or reverse direction, the circuit
is supplied with voltage and current.

Block diagram of
multi-channel transmitter.
The turn-on circuit supplies current to the rest
of the circuit when the controls are activated.
When the control is returned to the stop position (via a spring-return),
the current required by Q1 to keep it turned on is supplied by the
100u on the base and as the energy is delivered from the electrolytic,
the voltage across it reduces.
This reduces the voltage across the circuit but since it is not sending
out a signal, this does not matter.
After a minute, the voltage drops to almost zero and the electrolytic
is finally discharged completely by the 1M (and 4k7 in series with
it).
The stand-by current drops to less than 1 micro-amp, the leakage
through the collector-emitter junction when the transistor is not
turned on.
MORE ON HOW THE MULTIVIBRATOR WORKS
Transistors Q2 and Q3 form a multivibrator and the
operation
of this is fully covered in our books titled Learning Electronics
Book 1 and Book 2.
The circuit is basically regenerative in which one transistor turns
the other off then the second turns the first off.
When the circuit is first turned on, both the bases are pulled high
via the 10n capacitors but one of the transistors turns on before
the other and robs it of turn-on voltage.
But the transistor cannot stay turned on forever as the 10n capacitor
becomes charged and as it turns off, it sends a pulse to the other
transistor.
The second transistor turns on and completely removes
the turn-on voltage from the first transistor.
Eventually the second transistor cannot remain fully turned on due
to the 10n becoming charged, and starts to turn off. This sends a
pulse to the first transistor and it starts to turn on.
Each transistor has a gain or amplification factor of about 100 and
when we say one transistor begins to turn off slightly, this change
is passed to the base of the opposite transistor and the result is
magnified 100 times on the collector.
This is then passed to the base of the first transistor and suddenly
a tiny signal gets passed back as a huge signal.
That's why each transistor reacts so quickly and the result is a very
fast change from one state to the other.
This is shown in the shape of the output waveform.
The rise and fall
times are very short and the sides of the square wave are very steep.
The frequency of the output is determined by the value of the components
on the base. This includes the base resistor and the capacitor
connecting
to the opposite transistor.
In the circuit of fig: 12, the capacitors are fixed at 10n and the
resistors are changed.
An increase in resistance causes the capacitor to take longer to charge
and decreases the frequency of the circuit.
The output of the multivibrator is passed to the base of the RF output
transistor where it controls the on/off time for the transmitter.
When the transmitter is turned on, a 27MHz frequency is injected into
the base of the RF output transistor via a 47p from the crystal
oscillator.
This crystal oscillator is made up of transistor Q4 and its
surrounding components.
The transistor is turned on via the crystal and 22uH inductor. The
crystal is equivalent to about 20p and the resistance of the inductor
is about 1 ohm.
The emitter is held fairly rigid via the 47p and the transistor gets
a very short pulse from the crystal.
This puts a pulse of current through the coil and the current creates
magnetic flux. As soon as the pulse ceases, the magnetic flux collapses
and the inductor produces a voltage in the opposite direction and
passes the waveform through the 47p to the base of the RF output
transistor.
It also passes the waveform through the crystal to turn off the
oscillator
transistor Q4.
When the transistor is turned off, it does not put any load on the
inductor and the amplitude of the waveform is fairly large.
After a short period of time, this waveform ceases and the transistor
gets turned on by the 120k base bias resistor.
This injects another
pulse of current into the inductor and the cycle repeats.
The inductor creates the time delay for the waveform as it takes time
for the current to convert to magnetic flux then back into a voltage
in the opposite direction. This time-delay approximates to about 27MHz
and the crystal locks it on to the frequency of 27.240 by exhibiting
a larger capacitive effect at this exact frequency.
This is how the circuit is pulled into line and kept at an exact
frequency,
even though the supply voltage may decrease or the temperature may
rise.
The 27.240MHz waveform is passed to the RF output transistor and the
transistor is turned on and off at the frequency of the multivibrator.
The transistor is in common emitter mode as evidenced by the 10n on
the emitter.
The impedance of this capacitor at 27MHz is very small
compared to the 100R and the emitter considers it is connected to
the negative rail as far as the high frequency is concerned.
The 27MHz waveform on the base is amplifier by the transistor and
appears on the collector in magnified form.
The 22uH inductor on the collector prevents the signal passing to
the power rail.
It does this by producing a "back-voltage." As the transistor turns
on, the current through the inductor increases and magnetic flux is
produced in the coil that cuts the other turns of the coil and this
induces a voltage and current in them that is in opposition to the
current being delivered. The result is a reverse voltage is produced
that makes it difficult for the forward voltage to enter the coil.
This means the forward voltage gets larger and larger in an attempt
to enter the coil and the result is a large voltage appearing on the
collector of the transistor.
This voltage passes through the 47p to a tuned circuit made up of
a 11 turn inductor and 15p capacitor.
These are designed to match the high impedance of the output of the
transistor to the low impedance of the whip antenna.
Matching is required to get the maximum signal to pass into the antenna.
This completes the coverage of the sections in the transmitter.
THE MULTI-CHANNEL RECEIVER
The signal from the transmitter is picked up by the receiver as bursts
of tone between hash.
Viewing the signal on a CRO (Cathode Ray Oscilloscope) will look
something
like fig: 23.

The signal from the
multi-channel transmitter will consist
of a regular waveform between background hash.
The receiver is required to pick out the signal from the noise and
it does this by a process called integration and differentiation where
the signal is detected due to its regular nature and this is used
to charge a capacitor.
Another circuit determines the length of time
the tone is present and these are combined to determine the nature
of the control signal.
Most of the circuitry for doing this is locked inside the chip in
the receiver and the only components we can see are the external items
on pins 10, 1 and 19.
These determine the frequency detected by the chip and the length
of the "highs," but all the rest of the signal processing is done
inside the chip.
The chip detects the waveforms shown in figs 14 - 19 and turns on
the appropriate outputs.

A multi-channel 2MHz
receiver

The 27MHz receiver PC
board
Two outputs drive the motor in the forward/reverse direction and 4
outputs drive the transistors for the steering motor.
The steering motor is simply a rotary actuator. This is similar to
the armature of a motor, positioned inside a circular magnet.
The armature does not need brushes as it will only turn about 45°
in one direction and 45° in the opposite direction, depending
on the direction of the current.
The output of the shaft will be connected to a lever to steer the
front wheels.
The chip controls the two diagonally opposite transistors for the
clockwise and anticlockwise rotation to get left and right steering.
All the rest of the circuit has been previously discussed and the
only new feature is the tapping at 4.5v for the motor.
A diode on the 4.5v rail drops the voltage to 3.8v and the two output
transistors drop a further 1v, so that motor receives about 2.8 to
3v.
Here are some remote control items, shown
on the web, by a hobbyist who disassembles devices and makes a new
project:

Some of these components
were used to build a project and present it the web.
A 27MHz WALKIE TALKIE
An Overview
Walkie Talkies are the next logical step in this discussion. They
show how a crystal oscillator can be used to transmit voice.
Transmitting a voice via a crystal locked oscillator is not easy.
This is because the crystal is locking the frequency and it is very
difficult to shift it.
The only way to do it is to add the audio as an amplitude component
so that the amplitude of the oscillator rises and falls with the audio
signal but its frequency does not change.
The only problem with this mode of transmission is interference.
Electrical
noise entering the airwaves is also a varying amplitude waveform and
the receiver will pick this up at the same time and produce a very
noisy result. This is one of the reasons why walkie talkies are so
noisy. However it is a starting point for learning about transmission
and the circuit in fig: 24 shows how the audio is added to the carrier.

A 4-Transistor Walkie
Talkie
Nearly all the components in the 4-transistor circuit are used for
both transmitting and receiving. This makes it a very economical design. The
frequency-generating stage only needs the crystal to be removed and it
becomes a receiver.
The operation of this circuit coincides with our discussion on receiver
circuits at the beginning of this article where we said the receiver
was oscillating all the time, similar to a weak transmitter.
A 390R is added to the emitter of the oscillator stage to reduce the
activity and turn it into a receiver.
The next section of the circuit is called a building block.
It consists
of three transistors directly coupled to produce an audio amplifier
with very high gain.
The first transistor is a pre-amplifier and the next two are wired
as a super-alpha pair, commonly called a Darlington pair to drive
the speaker transformer.
The third block is the speaker. This is a separate item because it
is used as a speaker in the receive mode and a dynamic microphone
in the transmit mode.
A speaker can be used in reverse like this and it is called a dynamic
microphone because of the coil and magnet arrangement.
When you talk into the cone, the movement of the voice coil in the
magnetic field produces a few millivolts output. This can be coupled
to a high gain amplifier to get quite good results.
When the walkie talkie is in the receive mode, the first transistor
is configured as a receiver and the audio is picked off the 4k7 load
resistor via a 0.47u electrolytic.
It then passes through a volume control and into the three transistor
amplifier. The speaker transformer couples the amplifier to the speaker
and we hear the result.
When the walkie talkie is in the transmit mode, the speaker is placed
at the input of the audio amplifier.
The audio is then amplified and
the waveform appears as THE SUPPLY VOLTAGE FOR THE TRANSMITTER STAGE.
The crystal is connected to the first stage and the gain of the
transistor
is increased by removing the 330R and only using a 56R for the emitter
resistor.
The speaker transformer is not used as a transformer in this mode
but as an INDUCTOR to couple the output of the audio amplifier to
the power rail and the signal developed across the winding is passed
to the transmitter stage as the supply voltage for the transmitter.
As the waveform rises and falls, it changes the gain of the first
stage and thus the amplitude of the transmitted signal.
This is how the signal becomes an Amplitude Modulated (AM) Radio
Frequency
(RF) signal.
THE WALKIE-TALKIE
CIRCUIT IN DETAIL
In the receive mode, the first transistor is configured as a low-level
oscillator.
The base is tied to earth via a 39n capacitor. This makes it a
common-base
configuration and the gain of the transistor is high. The input (the
collector) is also high, whereas the input (the base) of a
common-emitter
stage is medium to low.
If this type of stage were used, the antenna would not be as sensitive
in detecting up a signal.
The feedback for the transistor is provided by the 33p between collector
and emitter.
The emitter has a 330R and 56R in series to keep the gain low.
The circuit starts up and oscillates due to a tuned circuit on the
output of the RF transformer. The transistor detects this oscillation
on the primary side of the transformer and passes the signal to the
emitter via the 82p, where the gain of the transistor increases the
amplitude of the signal to a medium level. If the amplitude is too
high, the stage will not be responsive to the surrounding signals.
Any nearby signals of the same frequency will increase and decrease
the current taken by this stage and the information on the signal
will appear across the 4k7 load resistor as a varying voltage. The
0.47u picks off the voltage and passes it as an audio signal to the
volume control and finally the 3-transistor amplifier.
The 4n7 between base and ground of the first amplifier transistor
is designed to remove any high frequency signals and the output of
the transistor goes to a super-alpha pair to drive a speaker
transformer.
The speaker transformer matches the output of the transistor to the
8 ohms of the speaker.
Matching is done by the transformer having a turns ratio.
It has 525 turns for the primary and 75 turns for the secondary.
The purpose of the transformer is to convert a high voltage (about
7v), with low current to a low voltage (about 400mV) at high current.
This is what the speaker requires. It needs a high current to pull
the cone into the magnetic field.
The remaining components are biasing components or capacitors to remove
the high-frequency signal.
SETTING THE BIAS FOR THE 3-TRANSISTOR AMPLIFIER
The circuit in fig: 25 shows the components that set the bias for
the three transistors.

The biasing of the
3-transistor amplifier
All the other components have been left out
because they do not determine the DC bias point.
The biasing starts at the base of the first transistor. It is turned
ON, but not fully, by the 1M resistor until the collector voltage
falls to half-rail voltage.
The 1M and 5k6 resistors are chosen so that this occurs.
This is the
ideal set-point so that the pre-amplifier transistor can amplify both
the positive and negative excursions of the signal without distortion.
The super-alpha pair (the second and third transistors in the
3-transistor
amp) drops a total of 1.3v across the base-emitter junctions, leaving
3.2v across the 100 ohm emitter resistor.
By ohms law, this will produce 32mA as the idle current (quiescent
current) for the audio stage.
Here is a hand-cranked walkie talkie:


The walkie talkie in
transmit mode. The signal
passes from the audio amplifier to the RF stage
via the power rail

The walkie talkie in
receive mode. The circuit is
conventional with the first stage feeding the
3-transistor amplifier via a volume control
THE EMITTER BY-PASS CAPACITOR
The 33u electrolytic on the emitter is called the emitter by-pass
capacitor. It connects the emitter to the 0v rail when the stage is
processing a signal and the signal by-passes the 100 ohm resistor.
To see how the electrolytic works we firstly have to remove it and
see why the Darlington pair has NO GAIN.
Refer to fig: 25. When a signal is fed into the base of the
pre-amplifier
transistor it will be amplified about 100 times and appear on the
collector.
Suppose the collector voltage rises 5mV.
This will be passed to the
base of the top transistor of the Darlington pair and since it is
almost fully turned on, the emitter will rise too. The emitter of
this transistor is connected to the base of the lower transistor and
the base will pull the emitter up too.
The collector voltage will not change and this means the transistors
will produce NO gain because the voltage on the emitter is allowed
to rise.
It we hold the voltage on the emitter rigid, the pair will give us
gain.
To do this we connect an electrolytic between 33u and 100u. It has
the feature of taking a long time to charge (and discharge) - compared
with the rise and fall time of the signal.
When the 5mV waveform appears on the base of the Darlington arrangement,
the base tries to rise but it is fixed by the characteristic voltage
of 0.7v developed across each of the base-emitter junctions. The result
is the base rises 0.1v and both transistors get turned on more.
The resistance between the collector and emitter of the output reduces
and the transistor allows more current to flow through the primary
of the speaker transformer.
THE SPEAKER TRANSFORMER
A lot of discussion could be devoted to the operation of the speaker
transformer
as the design of a transformer is very complex.
There are two ways you can design a transformer. One is to calculate
the requirements from scratch and the other is to copy an existing
design and make modifications until the desired result is achieved.
Copying and modifying is the quickest.
If you use the theoretical
approach you will invariably have to modify the design to get it working
perfectly.
The speaker transformer used in fig: 24 is 1k to 8 ohm. These are
the impedance values measured at 1kHz.
The actual DC resistance of the primary is 42 ohms and the secondary
is 1 ohm.
The DC resistance of a transformer is different to the impedance value.
If the transformer was larger, the wire diameter would be larger and
the DC resistance could be as low as 10 ohm and 0.5 ohms.
The impedance is the resistance as seen by the transistor at 1kHz.
It "sees" a 1k load at 1kHz and a higher impedance at a higher
frequency.
Energy is transferred from the primary to the secondary via magnetism.
The primary produces a magnetic flux that passes into the magnetic
core surrounding the windings. This magnetic flux cuts the turns of
the secondary and produces a voltage in it.
The voltage produced is
proportional to the number of turns.
In our case the primary has 525 turns and the secondary has 75 turns.
This is exactly a 7:1 ratio and it means the transformer will
theoretically
convert a 7v waveform at 10mA into a 1v waveform with a current of
70mA.
A small transformer like this has an efficiency of about 50 - 70%
however it is performing a very big task, matching 1k to 8 ohms and
the speaker would not work if it were connected directly to the
transistor.
To directly couple the speaker, the emitter resistor would have to
be lower. The circuit would then take 70mA to get the same result
as with the speaker transformer. And even then the transformer provides
a much better match.
THE TRANSFORMER AS AN INDUCTOR
When the transformer is used as an inductor in the transmit mode,
the speaker is not connected and the secondary does not see a load.
This means the primary does not see a "reflected" load and the impedance
of the transformer is increased considerably.
The effect is the transistor sees a higher impedance and this means
it finds it easier to develop a signal across the primary.
To give a very simple analogy, the transformer (with the speaker
connected)
is like a very stiff spring. When the speaker is removed, the
transformer
is like a very weak spring.
The transistor finds it very easy to pull the bottom end of the spring
down (the top is connected to the positive rail).
When a signal is processed by the Darlington pair in transmit mode,
the emitter is held rigid by the 33u and the only thing that can happen
is the weak spring gets pulled down.
By referring to the circuit diagram in fig: 24, the bottom lead of
the transformer becomes the power rail of the crystal oscillator and
as the voltage on the transformer rises and falls, the supply voltage
to the oscillator increases and decreases. and affects the gain of
the oscillator.
Now we come to the difficult part of explaining how a voltage is
produced
across the primary winding.
During the quiescent (idle) mode, about 1.5v is dropped across the
42 ohm resistance of the primary.
When a signal is processed by the Darlington pair, the resistance
between the collector and emitter is reduced and a higher current
flows.
The action of this current increasing creates an expanding
magnetic flux in the transformer and this flux cuts the adjacent turns
of the primary and induces a voltage in each of the turns in the
opposite
direction.
This means the voltage produced by the transistor has to be greater,
in an attempt to pass current into the inductor.
This voltage is picked
off the inductor and passed to the first stage in the circuit and
becomes the power rail.
The fluctuating power rail alters the gain of the stage and amplitude
modulates the 27MHz signal to produce audio on the carrier.
The result is an Amplitude Modulated (AM) Radio Frequency (RF) signal.
MORE 27MHz WALKIE
TALKIE CIRCUITS
The following 3 circuits have been taken from different Walkie Talkies
and show designers have included a crystal, a call button and an LM386
audio amplifier chip into their designs. You can take the different
features from one circuit and incorporate them into your own design.
Just remember, receiver circuits that are only for receiving a signal
are designed to pass very little current so they are very sensitive.
Walkie Talkie circuits need the fist state to deliver an output as well
as receive a signal so they have to be designed with both transmit and
receive capabilities.

27MHz Walkie Talkie without Crystal

27MHz Walkie Talkie with Call

27MHz Walkie Talkie with LM386 Audio Amplifier
49MHz WALKIE TALKIES
Two bands have been allocated for walkie talkies and remote control
equipment. These are the 27MHz band and 49MHz band.
The 49MHz band has slightly better performance due to the short antenna
being closer to the wavelength of the signal.
The two bands allow more remote control cars to be raced together
without interference between the cars.
FURTHER USES
All of these circuit can be found in remote-control toys from your
local department store.
Simply buy a remote control car and give it to a youngster to play
with.
After a day or two he will lose interest and you will be able to pull
it apart and adapt it to your own use.
To create a private channel, simply replace the crystal with one of
a slightly different frequency and retune both the transmitter and
receiver coil.
The multi-channel receiver has even most possibilities. You can control
four different devices directly and even more by gating the outputs.
The simple 27MHz link will be used with one of our Talking Electronics FM transmitters to
turn it ON and OFF remotely.
The 27MHz transmitter will work up to
60ft (20m) and will allow you to turn off a transmitter to give it
added security from being detected.
The receiver will have to be designed to turn on for 0.5sec every 10 seconds to detect if
a turn-on transmission is being sent and the whole circuit will then
shut down to conserve power if a reception is not detected.
This means you will have to transmit for at least 10 seconds to be sure
the receiver picks up the signal.
On the next page we cover some more 27MHz transmitter circuits, and on
P3 we cover some 303MHz links.
P2
P3
1/10/2016
|