This is only a preview of the March 1996 issue of Silicon Chip. You can view 28 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:
Articles in this series:
Items relevant to "Programmable Electronic Ignition System For Cars":
Items relevant to "Automatic Level Control For PA Systems":
Articles in this series:
Items relevant to "A 20ms Delay For Surround Sound Decoders":
Articles in this series:
|
REMOTE CONTROL
BY BOB YOUNG
Multi-channel radio control
transmitter; Pt 2
This month, after a long hiatus, we continue the
description of the Mk.22 transmitter by presenting details of the encoder module. This is a great
deal more flexible than was envisaged in the
June 1995 issue and is still based on discrete ICs
rather than a microprocessor.
Here it is at last! What can I offer
in my defence for the long delay?
Nothing much except that the size
and complexity of this project has
grown alarmingly and has necessitated
calling in additional help in order to
get it finished. I am indebted to Dean
Herbert (Microherb Electronics) for the
design of the basic encoder module
presented here.
Once again, I set out to design one
construction of the basic 8-channel
transmitter.
This transmitter is intended as a
companion to the Silvertone Mk.22
AM receiver published in the December 1994 & February, March & April
1995 issues of SILICON CHIP. However,
it will also act as a replacement for
almost any AM PPM transmitter currently on the market.
While most transmitters on the
As can be imagined, this transmitter
did not fall out of a tree but came as
a result of months of gruelling proto
typing and missed deadlines.
thing and instead ended up with something completely different; the best
transmitter that Silvertone has ever
produced. The complete transmitter
consists of four PC boards: RF module
(AM), basic 8-channel encoder, an
expansion PC board with an additional 16 channels, and a configuration
module. In the next few articles, we
will be dealing with the design and
54 Silicon Chip
market have a moulded plastic case,
the new Silvertone Mk.22 has a more
rugged powder-coated aluminium
case measuring 170 x 140 x 47mm
which is quite compact. Apart from
being more robust, the metal case is
a desirable feature as it reduces the
possibility of interference (third-order
intermodulation) from other transmitters close by.
Packed into the case is a glittering
array of features. Most are routine,
common to all modern R/C systems
but some are completely unique. As
noted previously, it is expandable
from 8 to 24 channels and has capabilities for channel allocation, servo
reversing, gain control, dual rate, servo travel length adjustment, pseudo
endpoint adjustment and mixing on
every channel.
Also included are four programmable on-board mixers (two inverting and
two non-inverting), two programmable
on-board toggle switch control modules, fully programmable front panel
switches and controls, and an expansion port for a configuration module
of which there are four at the moment:
F3B, Helicopter, Aerobatics and the
very exciting and completely unique
formation flying module. These
modules plug into the configuration
port (mix expand) and configure the
transmitter into a task-oriented unit.
Finally, as a topping for this culinary
delight, we add a dash of exclusive Silvertone relish: frequency interlock and
mixed mode dual control (buddy box).
As can be imagined, this transmitter
did not fall out of a tree but came as
a result of months of gruelling proto
typing and missed deadlines. By far
the most difficult task was keeping
the system user friendly. A great deal
of work has been done on the development of the input networks to each
channel and much care devoted to
maintaining a completely identical
board layout so that mastering one
channel results in the mastery of all
channels.
For example, there is only one type
of front panel switch – a SPDT toggle
fitted with a 3-pin socket. Thus, any
switch may become a retract switch,
dual rate switch, mix in-out switch
etc, depending on which set of header
pins it is connected to. The mix select
switch only calls for a 2-pin socket, so
any pair on a 3-pin socket will suffice.
The sockets on the switches provide
an added advantage in that the action
of the switch (UP-ON) may be quickly
and easily reversed (DOWN-ON).
In other words, the front panel
switches are fully program
mable.
Alternatively, the switches may be
replaced by shorting links on the
header pins for permanently installed
features. Thus, for example, coupled
Aileron/Rudder (CAR) may be installed permanently with a shorting
link or set up to operate in the switch
in/out mode from the front panel. On
the other hand, Flap/Eleva
tor auto
compensation is more usually installed as a permanent feature and thus
a simple shorting link on the header
pins will suffice.
Likewise, any channel may be programmed either to be pro
portional
or switched and there are two toggle
switch modules onboard for this
purpose.
The major compromise in the system came about as a result of reducing
the number of potentiometers to be
adjusted and the number of shorting
link combinations available. This was
achieved by making the gain control
pot programmable and is probably
the most clever feature in the user
interface. Thus, we ended up with
only a single potentiometer to adjust
for each channel, which in turn may
be the servo travel volume adjust
(ATV), endpoint adjustment, dual rate
set pot, mix ratio set pot or whatever,
depending upon how the channel is
programmed.
As a result of this simplification, certain combinations tend to compromise
the action of this pot. A good example
is the simplification that took place in
the dual rate programming. Originally
the NORMAL/DUAL RATE programming pins TB1, TB3 etc consisted of
six pins arranged in two rows of three;
a very cumbersome programming arrangement with many combinations.
Now the way we achieve 120% servo travel is to remove the 33kΩ input
resistor from circuit, in the GAIN VARY
setting. Thus, the NORMAL range
is 1-2ms and with the 33kΩ resistor
removed, 0.9-1.1ms. By placing the
33kΩ input resistor on the control pot
side of TB1, it was possible to reduce
the programming to a simple 3-pin
plug which allowed the use of an
SPST switch for remote operation. The
trade-off is that the dual rate pot will
actually increase instead of decreasing the servo travel as it approaches
full clockwise rotation. This comes
about because VR1 effectively shorts
out R2 as it approaches the clockwise
terminal.
Thus, when setting the dual rate
throw, starting from full clockwise rotation will give 120% servo travel. As
the pot is rotated anticlockwise, this
will drop back to 100% then on down
to 20%. All of this will be explained
in detail in future issue.
Whilst this action is a little unusual,
it is still dual rate even if it does go
higher than normal throw. It is only tra-
encoder circuit.
Circuit operation
The basic encoder follows the design philosophy pioneered in the early
1970s which culminated in the Signet
ics NE5044. It uses a multiplexed ramp
generator IC3b to generate standard
pulse position modulation (PPM) –
see Fig.1. Neutral for all 24 channels
is set by a single pot, VR2, associated
with IC1b. This feature represents a
significant cost saving in transmitters
with more than four channels.
IC4 is the 8-channel multiplexer, a
4051. In a full system, there are three
of these which will allow 24 channels,
via the expansion PC board. IC4 samples each control input sequentially
until all inputs are examined and then
there is a pause (sync pause) before the
process begins again. The rate at which
Packed into the case is a glittering
array of features. Most are routine,
common to all modern R/C systems
but some are completely unique.
dition that states that dual rate must go
down from the normal control throw.
Another good example of this sort
of compromise is the situation that
arises when programming for dual
rate combined with coupled channel
mixing; CAR, for example. In this
case, the Aileron gain set pot becomes
primarily the dual rate set pot and
the mix ratio adjustment is set on the
auxiliary mix pot which is part of the
four on-board mixers. This feature
was a particularly difficult one to
achieve, for in the beginning I could
not switch the mix ratio with the dual
rate switch. By utilising the spare pins
on the MIX EXPAND port, I found the
programming combination I required.
This called for the pins on the MIX
EXPAND port to be double-sided and
we will cover this point in detail in the
following articles. Now we have full
mixing with dual rate on the mix ratio.
Actually, I’ve got a little ahead of
myself in talking about these circuit
details but they are really operational
features so it was hard to avoid. Let’s
now get down to the nitty gritty of the
this sampling takes place is called the
FRAME RATE and is typically 16-24ms
in the 8-channel system.
The eight identical input stages each
contain a 3-pin plug (TB2, TB4, etc), to
which the control stick pots are connected. These provide the channel allocation and servo reversing features.
The second set of 3-pin plugs (TB1,
TB3, etc) are used to select NORMAL
or GAIN VARY modes, depending on
how the associated shorting link is
plugged in to short between the centre
and one outside pin.
Setting the shorting link on the
NORMAL pair gives a fixed 1-2ms
pulse width variation. Setting the
Fig.1 (next page): the circuit uses an
8-input multiplexer (IC4) which is
switched by counter IC5. IC4 samples
all eight inputs in sequence and this
creates a staircase waveform at the
output of IC3a. the two comparators
(IC1a & IC1b) then transform this into
the pulse position modulation (PPM).
March 1996 55
56 Silicon Chip
March 1996 57
each individual channel or input in
the encoder.
Thus, the output of IC3a will be
DC but stepped up and down (ie, a
staircase), according to the settings on
each of the eight inputs. This output
is fed to pin 2 of comparator IC1a
where it is compared with the ramp
generator (IC3b) at pin 3. The output
of the comparator is a series of narrow
pulses whose timing, relative to each
other, is a function of the DC input
and the ramp; so the higher the DC
input, the longer the time between
successive pulses.
Thus, the output of the comparator
is a block of eight pulses which have
times between them proportional to
the gain settings on the inputs. This
is known as pulse position modula
tion (PPM).
Synchronising
This scope photo shows how the encoder produces PPM (pulse position
modulation) from the eight multiplexed input channels. The upper trace is the
staircase waveform at pin 1 of IC3a (following the multiplexer) while the locked
pulse waveform is from pin 7 of IC1b. The long positive pulse the sync pause.
shorting link on the GAIN VARY pair
provides 20-120% servo travel controllable from VR1, VR3 etc. The shorting
links on TB1, TB3, etc may be each
replaced with an SPDT switch which
allows remote programming from the
front panel.
VR1, VR3, etc are gain controls
and can perform multiple functions
depending on how the terminal blocks
TB1, TB3, etc are set up (programmed).
Thus, they can provide the dual rate
adjust, servo gain (ATV) and mix ratio.
Mixing expansion port
TB10 is the mixing expansion port
and this is normally fitted with a
shorting plug for the main 8-channel
input leads. This is removed when
the configuration module is plugged
into this port. The pins for this port
are double-sided and they also act as
pick-up points for the mixers involved
with IC6. We’ll come to those later.
TB11 is the 24-channel expansion
port. TB12, TB13 and TB14 are the
channel number select connectors and
select 8, 16 and 24 channels respectively. For example, if TB13 is shorted,
then 16-channel operation is selected.
The main board comes with a shorting
bar on TB12 (on the PC tracks) which
must be cut if you intend to install
more than eight channels. Likewise,
58 Silicon Chip
R25 must be removed for more than
eight channels.
These connectors may be hard wired
or fitted with header pins if you intend
to swap backwards and forwards from
8 to 16 or 24 channels. These pins
could even be wired to a 3-position
switch on the front panel which would
allow front panel selection of 8, 16 or
24 channels. As pointed out previously, the flexibility of this system is
virtually unlimited.
TB7 is the power input connector
and it also carries the modulation to
the RF module.
Let’s start our analysis with IC5, a
4024 counter which continually feeds
a series of binary numbers to the A, B
& C pins of IC4. Thus, IC4 sequentially switches each of its eight inputs
through to R20, the input resistor for
IC3a which is a DC amplifier.
Because IC4 is an addressable
analog switch, any resistance in series
with its inputs (ie, R3, R6, R9, R12,
R26, etc) must be considered to be in
series with R20. This total resistance
will therefore determine the gain of
each individual input.
From this simple fact derives the
magic of the Mk.22 encoder. The ratio
of this total resistance (including R20)
and IC3a’s feedback resistor R18 will
determine the servo travel (GAIN) for
The 8-pulse block is synchronised
by the sync pause generator, IC3c, an
op amp functioning as a one-shot. Each
time Q4 of IC5 goes high, it charges C9
via diode D2 and R16 and the resultant
high pulse from IC3a resets IC5 via
R21. So IC5 starts again and switches
the first of the eight inputs through to
IC3a and the sequence continues.
The length of the sync pause is
controlled by the RC time constant of
R15 & C9, which set it at 8ms.
This is a very important point,
particularly in 16 and 24-channel
transmitters, as it gives the minimum
frame (repetition) rate and thus helps
to minimise servo slow down. This can
arise in some servos if the servo pulse
stretcher cannot cope with the long
repetition rates used in the high level
transmitters. The alternative system
found in some transmitters is to use
a fixed frame rate which must be long
enough to encompass the maximum
width control pulse (2ms) plus the sync
pause (8ms). Thus, a 24-channel transmitter of this type would use a fixed
frame rate of [(24 x 2) + 8]ms = 56ms.
By contrast, the Mk.22 uses a
swinging frame rate which varies
between [(24 x 1) + 8]ms = 32ms to
56ms, depending upon where each
of the control pots is set. In high level
transmitters, it is a good idea to leave
all unused channels set at 1ms to speed
up the repetition rate.
In the 8-channel transmitter, we use
a frame rate which will vary from 1624ms. The form of frame rate generation used in the Mk.22 also has another
advantage when changing from 8, 16
or 24 channels. As the sync pause is
tacked onto the last pulse, the frame
rate increase is adjusted automatically
to suit the number of channels in use.
For people who are concerned about
achieving the minimum frame rate
(maximum data refresh rate), the sync
pause may be set lower but be aware
that there is no standard for the sync
pause in commercial receivers. Some
will operate comfortably on 4-5ms
but others will fly out of sync even
at 6 or 7ms. From past experience, I
have found 8ms to be a fairly safe time
constant. As the Mk.22 Tx is intended
as a replacement for all commercial
transmitters, we have been a little
conservative here.
Unfortunately, the picture of the
circuit operation presented so far is
a lot more complex in reality. IC2a,
a D-type flipflop, is actually the controller of everything. And to further
confuse things, it is not even used like
a conventional D-type. Instead, it is
used as an RS flipflop which is “set”
by the pulse output from IC1a and
“reset” by IC1b, another comparator.
Neutral comparator
IC1b is the neutral comparator and
is set by VR2. The voltage from VR2 is
compared with the same ramp generator signal from IC3b and this produces
a similar series of narrow pulses with a
1.5ms spacing between them. So each
time IC2a is set by IC1a, it is reset by
IC1b a little later.
IC2a not only clocks counter IC5 but
it also drives the ramp generator, IC3b,
which is actually an RC integrator; it
“integrates” the output of IC2a and so
we have a sawtooth waveform which
is locked to IC2a, to the counter and
to everything else.
So how do the pulses from IC3a
actually get to the modulation output
on plug TB7? The answer is that they
don’t. The pulses generated by IC1b,
since they are locked to everything
else, actually become the modulation.
Comparator IC1b drives transistor
Q1 and thence D-type flipflop IC2b
which functions as a monostable
multivibrator. Its Q output will go high
when ever it is clocked by comparator
2 (IC1b) and stay high for a period set
by the RC network of R11 & C5 which
drives the reset pin. R52 and D3 speed
up the recovery time and eliminate
variations in the modulation pulse
length due to variations in the width
of the control channels. The nominal
length of the modulation pulse is set
at 350µs.
IC3d is the pulse shaper, an op amp
integrator used to adjust the rise and
fall time of the modulation pulse to
the RF modulator, an important point
when we come to the RF module. A
correctly set modulation pulse will result in a bandwidth of around ±10kHz.
IC3 is a TLC 2274, specified to provide
near rail-to-rail switching for the RF
modulator. R55 and C4 also help to
reduce the modulation rise and fall
times to reduce RF harmonics. Whilst
in theory C6 should provide symmetry
on both leading and trailing edges,
in practice we found C6 controls the
leading edge slope and R55 and C4
control the trailing edge slope.
YOU CAN
AFFORD
AN INTERNATIONAL
SATELLITE TV
SYSTEM
SATELLITE ENTHUSIASTS
STARTER KIT
Mixers
Op amp IC6 provides four mixers
with gain set by VR8, VR10, VR12
& VR15, respectively. The two small
modules at bottom right – TB17, 20
and TB21, 24 – are toggle switch control modules. TB27 and TB28 (middle
left) are mixer select programming
pins. These mixers are connected to
the main circuit by small patch cords
to the appropriate pins on the mix
expand port, TB10.
Another very important feature of
the circuit is the voltage reference
rails provided by R22, R23, R58 & R61
which are 1% resistors. These accurate
voltage references are derived from
REG1, an LP2950 low drop-out 5V
regulator This regulator allows reliable
operation down to 5V or less on the
transmitter battery.
These accurate reference rails allow
servo reversing on all channels by simply reversing the control pot polarity.
In the Mk.22 encoder, this is done by
reversing the 3-pin socket associated
with each control pot. This function
could be achieved with switches but
it would mean the loss of channel
allocation.
Channel allocation in the Mk.22 is
achieved by simply connecting any pot
to any channel. Channel allocation is
a vitally important feature when we
come to the F3B module for example,
where two channels are required for
ailerons and another two for flaps;
using only one stick axis for each pair
of channels.
So there we have it, a basic no frills
8-channel encoder with expansion to
SC
24 channels if required.
YOUR OWN INTERNATIONAL
SYSTEM FROM ONLY:
FREE RECEPTION FROM
Asiasat II, Gorizont, Palapa,
Panamsat, Intelsat
HERE'S WHAT YOU GET:
●
●
●
●
●
●
400 channel dual input receiver
preprogrammed for all viewable satellites
1.8m solid ground mount dish
20°K LNBF
25m coaxial cable
easy set up instructions
regular customer newsletters
BEWARE OF IMITATORS
Direct Importer: AV-COMM PTY. LTD.
PO BOX 225, Balgowlah NSW 2093
Tel: (02) 9949 7417 / 9948 2667
Fax: (02) 9949 7095
VISIT OUR INTERNET SITE http://www.avcomm.com.au
YES GARRY, please send me more
information on international band
satellite systems.
Name: __________________________________
Address: ________________________________
____________________P'code:
__________
Phone: (_______) ________________________
ACN 002 174 478
March 1996 59
|