This is only a preview of the July 1995 issue of Silicon Chip. You can view 31 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. Items relevant to "A Low-Power Electric Fence Controller":
Items relevant to "Run Two Trains On A Single Track":
Items relevant to "Satellite TV Receiver; Pt.3: Setting Up A Ground Station":
Articles in this series:
Items relevant to "Build A Reliable Door Minder":
Articles in this series:
Articles in this series:
|
REMOTE CONTROL
BY BOB YOUNG
Minimising transmitter interference
This month, we will examine a startling new
development which has radically altered
the design parameters for the new Mark 22
transmitter.
In past issues of SILICON CHIP I have
often referred to projects taking on a
life of their own, once some heavy
duty effort is directed towards them.
The Mark 22 project is one of the best
examples of this process in action that
I have encountered during my long
career in R/C manufacturing.
It began with me being dragged kicking and screaming by the editor of this
magazine once more to the dusty stool
in front of my old drawing board. He
said he (and his readers) badly needed
this R/C project and that my feelings in
the matter were of no account. As this
argument had raged on for some time,
I finally realised that further argument
was futile. With that, I reluctantly set
to work to modernise my old Mark 14
AM system. “That should get him off
my back”, I mused.
From there I have watched the
Mark 22 develop into one of the most
versatile R/C systems on the market
today. So much so that the model
aircraft fraternity have greeted it with
a degree of enthusiasm that has taken
me completely by surprise. As their
enthusiasm has grown so has my own.
Each step in the development process
A classic scene at an R/C car track. This is much less than the recommended
minimum spacing of three metres. How many of these transmitters are
interfering badly with each other and possibly causing loss of control?
72 Silicon Chip
has led smoothly (and not quite effortlessly) to the next logical step, to the
point that I now find myself once more
at the cutting edge of R/C development
here in Australia.
However, I have got ahead of myself
a little and I must return to the February 1995 issue of SILICON CHIP which
featured the new Silvertone frequency
keyboard. This keyboard is the latest
in a long line which began in 1969 and
now incorporates the new frequencies
on the 36MHz modelling band.
The MAAA (Model Aeronautical
Association of Australia) had released
these frequencies as a result of the latest range of very elaborate R/C equipment arriving from overseas and I was
asked to prepare a new keyboard in
anticipation of the operation of 10kHz
spacing on the 36MHz band.
It was widely believed that this
generation of equipment would allow
safe operation on 10kHz spacing. Well,
the upshot of recent field testing is that
the MAAA has decided that 10kHz
spacing is definitely NOT SAFE. The
new 10kHz frequencies will be used
but only at 20kHz spacing.
That is startling enough but another
problem has come to light because of
this close spacing proposal which is
potentially more serious and it has
been there all along. I am speaking of
direct interference between transmitters when they are physically close
together and operating on adjacent
frequencies.
What happens is that if two standard transmitters are oper
ated close
together they both radiate extra signals
and these extra signals will be on frequencies which might be being used
by other radio control transmitters at
the time.
So here is the scenario. Two transmitters are being operated close
together and they both radiate inter-
Fig.1: this frequency spectrum shows two conventional
class C R/C transmitters spaced 20kHz apart at
27.175MHz and 27.195MHz. Note the interfering signals
spaced 20kHz away at 27.155MHz and 27.215MHz. These
signals are only 30dB down on the wanted signals.
fering signals at the same frequency as
another R/C transmitter on the same
field. The result can easily be that the
third model loses control and has a
crash! No-one has done anything illegal and the poor unwitting victim is
left wondering why it happened. Has
this happened to you?
What we’re talking about is 3rd
order intermodulation. This type of
interference is generated when two
non-linear (class C) transmitters are
operated in close proximity of one
another. The 3rd order products (P)
are generated according to the formula:
P = (2F1 - F2)+ (2F2 - F1)
Let’s put some actual operating
frequencies into this equation. If we
have two transmitters operating at
27.195MHz and 27.175MHz (ie, 20kHz
apart), they will produce interfer
ing fre
quencies at 27.215MHz and
27.155MHz. Note that these interfering
signals are “legitimate” frequencies on
the same 20kHz spacing.
The effect is shown in the frequency spectrum of Fig.1 which is part of
a series of tests I did for this article.
This photo shows the two operating
frequencies as the taller spikes while
the unwanted frequencies on either
side are only 30dB down. (Note that
this and the other spectrograms shown
in this article were taken with unmod
ulated transmitters to give a clearer
result).
The power of these interfering frequencies is inversely proportional to
the square of the distance between the
antennas; so the closer they are, the
Fig.2: this frequency spectrum shows a class B transmitter
at 27.195MHz and a class C unit at 27.215MHz. Note
the lower amplitude unwanted signal at 27.175MHz, the
result of the improved linearity of the class B transmitter.
worse is the interference. This interference can be very powerful and quite
capable of shooting down an aircraft.
And if the two R/C antennas touch,
as they easily can in the excitement
of a race, the power of the unwanted
products can be almost equal to that
of the proper signals.
I knew of the problem but had
no real concept as to its magnitude.
During these tests I generated enough
3rd order product to lock out PCM
receivers and drive them into fail-safe.
That is not the end of the problem
as there will also be 5th order inter
modulation products and these are
demonstrated in one of the spectrum
photos (Fig.4). Hence, as well as the
interfering signals noted above, there
will also be unwanted signals at
27.235MHz. and 27.135MHz, although
their power level will be reduced.
This problem is a well understood
by RF engineers. When working with
multiple transmitters on a single tower, they spend considerable amounts
of time minimising 3rd and 5th order
intermodulation products.
Why does it happen?
When a transmitter is operated close
to a second transmitter, some of the
radiated RF is absorbed by the second
transmitter’s antenna and its tank circuit. This unwanted RF energy finds
its way to the base-emitter junction of
the PA transistor which is operating
in non-linear class C mode. Because
of this, it acts as a mixer and so the
unwanted difference frequencies are
ampli
fied and radiated along with
the transmitter’s proper signal. The
second transmitter affects the first
transmitter in exactly the same way, so
both transmitters radiate the unwanted
frequencies.
I must emphasise that this interference problem has always existed but
it becomes much worse when the frequency spacing between transmitters
is reduced. It is bad enough when a
spacing of 20kHz is used and is quite
capable of causing crashes. But with
10kHz, the problem would be a great
deal worse.
How do you stop it?
So what measures can be undertaken to eliminate this problem, or
at least minimise the risk? First and
foremost, the transmitters should be
far apart; ideally no closer than three
metres between them.
Second, anything that attenuates the
incoming RF will help and so a metal
transmitter case is to be preferred. The
Mark 22 transmitter will (naturally)
feature a metal case.
Third, a good way to minimise the
problem is use a more linear transmitter circuit. So instead of using the
conventional class C transmitter, a
move to class B transmitters is very
worthwhile and this is demonstrated
in the spectrum photos of Figs.2 & 3.
Here, one of the transmitters is a class
B model and you can see that one of the
unwanted signals is greatly reduced. If
two class B transmitters are operated
close together, the overall radiation of
July 1995 73
Fig.3: this test is the same as Fig.2 except that the transmit
ters have been swapped; class C at 27.195MHz and class B
at 27.215MHz. In this case, the lower amplitude unwanted
signal is at 27.235MHz.
interference signals is reduced even
further.
Demonstrate it for yourself
Many people express surprise at
the thought of a transmitter absorbing
power and re-radiating it, but it is acting purely as an absorption wavemeter
and this can be easily demonstrated,
without any need for test gear.
If you have two transmitters with RF
meters, switch on one and move the
other’s antenna in close proximity to
it, you will see the meter of the OFF
transmitter begin to read RF from the
ON transmitter. Move them closer
together and you will see the meter
on the OFF transmitter register a substantial signal. So you can imagine
that when that second transmitter is
turned on, all hell breaks loose and
the interference is rife.
Here then is an explanation for the
completely transient and random
nature of some interference. Over the
years I have spent hundreds of hours
going through sets which have come
in with vague complaints of “interference” and all of the sets have checked
out perfectly normal and few have ever
returned. Was it 3rd order intermodulation? There is no way of knowing
but it is highly probable.
Having said all of this, I must state
also that there is no need for panic.
Safe field procedures will eliminate
the problem completely and these
include separation of each transmitter by a minimum three metres, field
testing of suspect transmitters, placing
adjacent transmitters at the opposite
74 Silicon Chip
Fig.4: taken at a different screen refresh rate, this
spectrogram reveals the presence of 5th and higher
order interference products, as well as the 3rd order
signals.
end of the flight line and finally if
necessary, changing the frequency of
suspect systems.
Finally, if you see three keys in the
keyboard on adjacent channels and
yours is one of them, then be doubly alert as to where the other two
transmitters are located while you
are flying.
Why have 10kHz spacing?
The proposed introduction of the
10kHz spacing system was primarily
to enable large clubs to increase the
amount of activity per hour. Remember
here, it is not that anyone particularly
wants 60 aircraft in the air at once –
nothing is more unpleasant than a
crowded sky.
The idea is to free up channels so
that testing, motor tuning and field
adjustments, all activities that tie up
frequencies for long periods, may be
carried out. Plus, the more frequencies
that are available means less frequency
clashes and fewer accidents.
Proposed transmitter
After these tests, I was faced with a
dilemma. I have just designed a brand
new (class C) RF module and now it
is clear that a class B or better still a
class AB (linear) PA would minimise
the problem and thus make operation
on busy fields that much safer. Hence,
I have no hesitation in delaying the
design to incorporate a vital feature
for safe operation. I want the Mark 22
to be as good as I can make it.
Thinking about it, I cannot understand why this problem has not been
analysed and solved long ago. The
Americans obviously understand it for
they recommend a minimum of three
metres separation and even go so far as
to place boxes on the flight line three
metres apart and each pilot must not
leave his box. Some American clubs
even go to extremes and recommend
10 metres separation. But even then
they can run into problems, as indicated by the landing strip diagram
of Fig.5.
If we adopt the practise of separating
two adjacent transmitters by putting
them at opposite ends of the flight line,
then the aircraft comes much closer to
an interfering transmitter and the problem of signal strength ratios begins
to become a factor. This is a separate
issue to 3rd order interference and is
purely related to system bandwidth.
Fig.6 shows a simple go/no-go test
for determining the safety of operating
two R/C systems simultaneously. Here,
the signal strength ratios are related
to distance and a minimum of 12:1 is
called for. To go closer than two metres
to the receiver distorts the test due to
overload of the receiver.
In this regard I recommend that all
transmitter antennas be telescoped
when entering the pit area. Notice
the similarity of this test to the conditions illustrated in Fig.5. Silvertone
developed this test in 1969 and it has
gained widespread acceptance all over
Australia.
So it is obvious there are advantages
to the linear PA in R/C transmitters. If
the 3rd order is eliminated or at least
minimised, then operators can be
MODEL 1
TRANSMITTER 1
Fig.5: a typical landing field with transmitters spaced three
metres apart. This can place interfering transmitters much
closer than the control transmitter, as the model comes into
land.
INTERFERING
TRANSMITTER
11
11 TRANSMITTERS
SPACED 3 METRES
APART
LANDING
FIELD
30
METRES
TRANSMITTER
1
grouped much closer in a much safer
pattern, taking into account the signal
ratio problem.
Interference test procedure
For those who wish to conduct a
simple field test to determine the safety
of operating two R/C systems simultaneously, the following procedure is
recommended.
(1). Place model 1 on the ground
with the antenna parallel to the ground
45ø
STATIONARY
MODEL
CONTROL
TRANSMITTER
33 METRES
90ø
WALK IN
UNTIL
INTERFERENCE
OBVIOUS
RATIO TO BE
BETTER THAN
12:1
INTERFERING
TRANSMITTER
Fig.6: standard interference test (developed by Silvertone) for two adjacent
transmitters.
and about 30cm above the ground. (If
the antenna is closer to the ground,
ground effect will distort the results.)
If the model features a whip antenna,
be sure that it is vertical.
(2). Take the control transmitter for
the model under test out 33 metres
from the model, switch on the Tx,
fully extend the antenna and hold it
vertical. The angle between the receiver antenna and the transmitter should
be 45 degrees.
(3). Check the operation of the controls in the model to ascertain that all
are working correctly.
(4). Take the interfering transmitter out approximately 10 metres but
on a line at right-angles to the first
transmitter. Fully extend the antenna,
switch on the Tx and hold the antenna
vertical. This ensures that the receiver
antenna is evenly polarised and receiving equal field strengths.
(5). Take note of the operation of
the control surfaces in the model. At
these distances there should be no
noticeable effect on the controls.
(6). Walk towards the model with
the interfering transmitter along the
45° line. Keep moving closer until the
controls begin to exhibit some signs
of interference. Note the distance
from the model at which this occurs.
The ratio of the two distances of the
transmitters from the model should
be greater than 12:1. Thus, with the
control Tx at 33 metres, there should
be no interference with the interfering
Tx 2.5 metres away from the model.
(7). Repeat the test using model 2
and with the original control transSC
mitter as the interfering Tx.
July 1995 75
|