This is only a preview of the April 1996 issue of Silicon Chip. You can view 26 of the 96 pages in the full issue, including the advertisments. For full access, purchase the issue for $10.00 or subscribe for access to the latest issues. Articles in this series:
Items relevant to "A High-Power HiFi Amplifier Module":
Items relevant to "Replacement Module For The SL486 & MV601":
Articles in this series:
Articles in this series:
Items relevant to "Build A Knock Indicator For Leaded-Petrol Engines":
|
RADIO CONTROL
BY BOB YOUNG
Multi-channel radio control
transmitter; Pt.3
Following the description of the encoder module
last month, we present the long-awaited AM
transmitter circuit. This has been carefully
designed to keep harmonic content and third
order intermodulation to an absolute minimum.
Modern radio control transmitters
place enormous demands on their designers due to the wide range of (often
conflicting) features expected by the
users and the standards required by
the various watchdogs responsible for
the safe and harmonious application
of technology.
This is particularly true of the
transmitter module. Here the operator
can cause a third (innocent) party to
bore neat little holes in the ground.
We treated this subject in some detail
in the July 1995 issue of SILICON CHIP.
Thus the designer of a modern transmitter module is charged with serious
responsibilities. With this in mind the
design of the RF module presented has
proceeded slowly and cautiously. This
has been far too slow for some, judging
Interference takes on a very serious meaning
for model fliers. If they allow their transmitter
antennas to come in close proximity with their
neighbours, they can cause a third (innocent)
party to bore neat little holes in the ground.
can intrude into the domain of his
neighbour in a very big way. We are all
familiar with broadcast and television
interference but a new dimension has
been added recently in R/C circles, at
least in the form of 3rd order inter
modulation interference.
This aspect of the interference spectrum takes on a very serious meaning
for model fliers, for if they allow their
transmitter antennas to come in close
proximity with their neighbours, they
by some of the letters and comments
we have received in the period since
the publication of the AM receiver.
However, as they say, all good things
come to those who wait, and so here
at last is the long awaited transmitter
module.
Design philosophy
Those who remember the discussion in the June 1995 issue may recall
that at the time, I concluded that the
best approach for an RF module with
reduced third-order intermodulation
would be a class-B push-pull unit.
Initially, I proceeded to design a transmitter along those lines.
I quickly discovered several important aspects of third order intermod
ulation. First, direct injection can play
an important part in the process. Direct
injection occurs where the interfering
RF gets directly into the coils and
PC board tracks as opposed to being
picked up by the transmitter antenna.
This form of injection has been
minimised by the use of an aluminium
transmitter case, a ground plane on the
PC board, shielded coils and most important of all, by having the minimum
number of stages in the transmitter.
Secondly, I discovered that yes, the
push-pull circuit did give good results
but it had to be very carefully designed
and was very tedious and expensive
to build. What finally sunk this very
promising development was the discovery that if the bias was set incorrectly the third order intermodulation
was much worse than a class C output
stage. This was a great disappointment
since the class B stage proved to be
extremely efficient and one module
that we had out flying drew only 18mA
and gave excellent results.
I cried tears of blood over losing that
18mA output stage, especially when
I had to wrestle with this new design
to bring the current drain down to
reasonable limits. More on that later
but I still weep when I think of a transmitter with 12 hours of flying time on
a 600mA.h battery pack, especially
when you listen carefully to one of the
modern computer radios and you can
hear virtually the electrons roaring as
April 1996 65
Fig.1: the circuit consists of a Hartley oscillator, Q1, driving Q3, a VMOS Mosfet
critically biased by trimpot VR1. Modulation is applied to the output stage by
transistor Q2 which varies the supply to Q3.
they are sucked through the wires to
keep up with the demand for current.
At this point my attention was focused on the encoder design which
took many months, leading to its
presentation last month. In the intervening period I was able to formulate
the approach presented in this article.
The key aspect is the oscillator which
delivers a very high drive level with
good stability. In fact, you could almost
hang an antenna off this oscillator and
fly with it but it wouldn't really be
practical. You would need to amplitude modulate the oscillator and the
subsequent frequency modulation and
pulling that with AM would cause all
sorts of problems – not a good practice.
So that meant an RF power amplifier (PA) with modulation. Here I ran
into serious problems as the isolation
between stages was poor – there was
oscillator breakthrough and only 90%
modulation. At this stage the project
looked to be in serious jeopardy.
The standard cure is to use a diode
to set the bias threshold but this meant
more non-lineari
ty in the PA. This
was completely contrary to the design
philosophy which called for the output stage to be biased to the point of
acting as a perfect transistor in order
to reduce third order intermodulation.
66 Silicon Chip
I could have used a buffer stage but
again I ran foul of my own design
requirements, set out above.
At this point I realised that the emitter resistor was the main culprit in the
third order intermodulation process
and I set out on a search for transistors
with diffused emitter resistors. The data
books are full of them but you try buying one in this country. Up until then
I had been concentrating on bipolar
transistors. I then had an inspiration
and decided to use one of the VN series V-MOS FETs and lo and behold all
problems vanished; well, almost.
These FETs make ideal output transistors for transmitters, being almost
indestructible and with good gain at
30MHz. Once the change was made
to a FET PA, the problem of oscillator
break
through was minimised but it
still remains in a very mild form, so
care is need in this area during setup.
The circuit presented also features
some degree of latitude to make it
useful in non-modelling applications.
To this end I have indicated which
components are not used for R/C work
and those needed for matching into a
50Ω coax cable.
As presented, the transmitter delivers close to 500mW into a 1.5 metre
(60-inch) telescopic antenna with a
total current of approximately 120mA.
This includes the oscillator, PA and
encoder current. Useful operating time
from a 600mAh battery pack should be
in the order of four hours.
Circuit description
This photo shows the transmitter
in early prototype form. The
construction starts next month.
Transistor Q1, coil L5, crystal X1
and associated components comprise a
Hartley oscillator which is transformer
coupled into the PA transistor, Q3. R6
and C5 are for decoupling and C4 is
used to shunt any inductance in C5.
This type of oscillator provides a high
level of drive combined with good
depending on the coupling between the oscillator and PA,
too much bias can drive the FET into a very high current
mode. Capacitor C7 provides a ground return for the RF
flowing in the secondary of L5.
In the early stages of development of this circuit, I
had terrible problems with strong harmonics on 90MHz
coupled with very high levels of current in the FET. This
resulted in the FET almost steaming. Yet despite this maltreatment the five original FETs used in the prototypes
are all still working very happily and I have yet to see
one fail. As an added precaution, I have designed the PC
board so that the ground plane and the transmitter case
form a substantial heatsink.
More power possible
This spectrum sweep tells the story of how this new circuit
is successful in suppressing third order intermodulation.
The two large spikes represent the transmitter
fundamentals of the Mk.22 at 29.745MHz and a standard
imported Tx at 29.805MHz. The subsidiary spike at right
shows how the imported unit has substantial third order
intermodulation at 29.865MHz but the intermodulation
product of the Mk.22 Tx is well down, almost in the noise.
Reproduced from the July 1995 issue, this spectrum
sweep shows two conventional class C transmitters
spaced 20kHz apart at 27.175MHz and 27.195MHz. The
interfering signals, spaced 20kHz away at 27.155MHz and
27.215MHz, are only 30dB down on the wanted signals.
stability. The 22pF capacitor C2 is used for fine tuning
the crystal, if required. Increasing C2 will pull the crystal
lower in frequency although there is a limit to this.
Bias for Q3 is provided by trimpot VR1, resistor R8 and
diode D1 and is the core of the intermodulation solution.
The setting of VR1 is fairly critical and the third order
products can actually be tuned out when setting this
trimpot. By watching the spectrum analyser and tuning
VR1, the third order can be reduced to its absolute minimum. As this point is theoretically the point at which
the FET is behaving as a perfect transistor, this point also
corresponds closely to the point which gives the best
harmonic suppression results. One word of warning here:
This circuit is capable of further development and
could eventually deliver up to 1W with care in regard to
harmonic output. Coil L6 and capacitor C8 form a trap
for 90MHz which can prove troublesome at high drive
levels. These are not mounted in the R/C system but the
PC board does provide for them. 1W is far too much power
for R/C work but readers with non-R/C applications may
find this of interest.
The 10Ω resistor R5 is a “stopper” to prevent high
frequency parasitics while resistor R7 is there to discharge
the gate. Q3 is loaded in the R/C circuit with L4. While
provision is made for L3, it is not used in this circuit.
Capacitor C10 swamps the Mosfet capacitance and provides some stability to the output stage. It also provides
production repeatability and tunes L4 to 29MHz. The
amplified RF is then matched to the antenna by an LC
network consisting of capacitor C13 and coil L2.
For those wishing to use a 50Ω coax output, C6 will
provide adequate matching. This capacitor is quite critical
and would probably be best made up of a fixed capacitor
in parallel with a smaller variable type. Provision is also
provided on the PC board for an additional base loading
coil should the application require it. These components
are not used in the R/C system. This coil would be required, for example, if a short antenna was to be used.
TB1 is the transmitter module connector and provides
power, antenna and modulation connections.
Transistor Q2 is the modulation transistor and is config
ured as an emitter follower. Capacitors C11 and C15
provide RF bypassing and assist in the final shaping of
the modulation waveform.
This shaping is absolutely critical if the system bandwidth is to be held inside the ±20kHz allowed under current MAAA guidelines. At this stage of development, the
Mk.22 Tx is rated at ±15kHz at 60dB. This is a little higher
than I would have liked but well within the guidelines.
Capacitors C16 and C14 are DC filters for the supply rail.
This module will tune across the range of frequencies
allowed for R/C work and should tune to 50MHz for
non- modelling applications. The table presented in the
circuit diagram gives some idea of the capacitor changes
required for different operating frequencies. The coils do
not need to be changed.
So there you have it. I promised you a module with reduced third intermodulation and if you look at the spectrum
sweep in the accompanying photo you will see that this aim
has been met. Next month, we will discuss construction
SC
of the transmitter module. See you then.
April 1996 67
|