This is only a preview of the February 1999 issue of Silicon Chip. You can view 34 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 "Low Distortion Audio Signal Generator; Pt.1":
Items relevant to "Command Control Decoder For Model Railways":
Articles in this series:
Items relevant to "Build A Digital Capacitance Meter":
Articles in this series:
Items relevant to "LEDS Have Fun":
|
Command Control
Decoder For
Model Railways
This decoder circuit takes a different approach
to the design featured in our May 1998 issue.
Instead of feeding switched power to the
locomotive motor, it feeds smooth DC which is
better for some motors, including coreless types.
Not only does this circuit use less components
but it ends up on a much smaller PC board.
Design by CAM FLETCHER
Our series on Command Control for
model railways, which was presented
in the January to June 1998 issues of
SILICON CHIP, has created quite a deal
of interest. While there were some
initial problems with the supply of
ZN409CE servo decoder chips, these
have been overcome for the present
40 Silicon Chip
and so quite a few systems have been
built.
As always though, someone can see
a better way or another approach and
so it is with this alternative decoder
design which feeds smooth DC to the
motor and also manages to dispense
with the need for the ZN409CE de-
coder. While achieving this result,
the circuit also manages to use less
components and is accommodated
on a smaller PC board. As a result, it
could be fitted into some N-scale locos
as well as smaller bodied British OO
or HO-scale locos.
Before we describe what this circuit does, we should briefly review
the function of the original decoder
circuit featured in the May 1998 issue
of SILICON CHIP. This was installed
inside a typical HO or larger scale
locomotive and was fed with track
voltage of about 11V DC with a superimposed 5.9V pulse waveform.
The pulse waveform consisted of
blocks of 16 pulses separated by a sync
“pause” and the width of each pulse
contained the speed and direction of
each locomotive on the 16-channel
system. Ergo, a maximum of 16 loco-
motives could be simultaneously
controlled on the system.
The decoder circuitry extracts
the particular pulse from the block
of 16 pulses and then that pulse
is decoded to drive a H-bridge
transistor circuit which drives
the locomotive motor. The locomotive can be driven at any speed
up to its maximum, in forward or
reverse direction. The H-bridge
feeds voltage and current to the
motor in switching mode at a
pulse rate of about 100Hz.
To fully understand the decoder operation and hence the
differences between it and the
circuit described here, you will
need to read the May 1998 article
in detail.
The pulsed mode of operation
is fine for most locomotive motors
and has the big advantage that
the driving transistors stay cool
and do not require any heatsinks.
However some model railway
enthusiasts prefer not to run their
locomotives with pulsed power.
The switchmode operation can
lead to noticeable armature and
gear-train noise and vibration,
especially at low speeds and it
can cause heating problems in
some coreless motors which are
popular with British model railway enthusiasts.
This alternative decoder design uses just three ICs and four
TO-126 power transistors for the
motor drive. The transistors need
to be mounted on the locomotive
body, chassis block, ballast weight
or other suitable heatsink to dissipate the heat produced because
the transistors operate in linear
mode rather than switchmode.
Ideally, the current drawn by the
locomotive will be around 0.1A
or less, to minimise this power
dissipation. If efficient can motors
are used, this small current drain
is certainly achievable.
Fig.1 shows the circuit diagram.
There are few similarities between
it and the circuit of the original
decoder published in May 1998
although the principle of operation is broadly the same, as far as
recovery of pulses is concerned.
From then on, the decoding of the
recovered pulse is quite different.
As already noted, the track
voltage is a 5.9V train of pulses
Fig.1: IC1 and IC2 extract the channel pulse from the 16-channel block while IC3a,
Q2, D5 & D6 produce a DC output which is proportional to the pulse width. IC3c &
IC3d provide the forward/reverse decoding.
February 1999 41
Fig.2: the decoder has the resistors and other small components mounted vertically to save space. The board
for the output transistors is optional as the transistors do require some heatsinking. Note the link under IC1.
Fig.3: this is the artwork for the two PC boards, shown twice actual size. Note that we have used small pads for
the ICs, to allow tracks to run between pins.
superimposed on 11V DC. This is fed
to a bridge rectifier consisting of diodes D1-D4. The bridge rectifier does
not “rectify” the track voltage; it just
allows the circuit to be independent
of the track polarity. The track voltage passes through unchanged, apart
from the small voltage drop across
the diodes.
After the bridge rectifier, we have
10V DC with a superimposed 5V pulse
train. This is fed to the 3-terminal
regulator REG1 to provide +5V for
the ICs. The unregulated DC is also
fed direct to the H-pack transistors,
Q3-Q6. We’ll come back to these later.
The track voltage is also fed via
the 10V zener diode ZD1 and a 470Ω
resistor to pins 6 & 2 of IC1, a 555 timer. The zener diode can be regarded
as a level shifter which effectively
removes the 10V DC, leaving just the
5V pulses to be fed to IC1.
IC1 is a 555 timer but its use in this
circuit is unconven
tional. Its main
function is as a Schmitt trigger to clean
up the pulse waveform after it has
been fed through the bridge rectifier
and zener diode.
42 Silicon Chip
Pin 7 of IC1 is internally switched
to 0V whenever pin 3 is low and so
C3, the 1µF capacitor at pin 7, is
discharged each time pin 3 goes low.
However, at the sync pulse interval,
which is the gap between each block
of 16 pulses, C3 has time to charge
up and turn on transistor Q1 which
then stays on for the duration of the
sync pulse. Q1 pulls pin 11 of IC2 low
and this is the “load” function for the
74C193 up/down counter.
IC2 actually extracts the wanted
pulse for the particular locomotive
from the block of 16 pulses. In effect, it
is loaded with the wanted pulse number by means of the binary data inputs
at pins 1, 9, 10 & 15. The counter then
counts down by 16 from the wanted
number and the recovered pulse appears at the “borrow” output, pin 13.
The magic of this system is that the
wanted pulse with its all-important
width information is recovered intact
and can then be fed to the following
decoder circuitry.
Going back to Q1 for a moment, it is
used to pull pin 11 low for the “load”
function. Normally, Q1 would need a
collector load resistor of, say, 1kΩ, to
make sure that pin 11 is pulled high
when Q3 is off; ie, a pullup resistor.
In this case though, pin 12 is used
to supply the pullup function. This
can only be done with the 74C193 or
40193B ICs. If you use other than 74C
or B series CMOS for this IC, you will
need to isolate pin 12 and provide a
1kΩ pullup resistor from pin 11 to
the +5V rail.
Decoder operation
As already noted, this circuit dispenses with the ZN409CE decoder
chip. Instead, the decoding operation
is performed by IC3a & IC3b in conjunction with Q2, D5 & D6. Pin 5 of
IC3a and the base of Q2 are biased at
+3.3V from pin 5 of IC1. This is not
a normal use for the threshold pin of
a 555 but it works in this application
and saves resistors which would
otherwise be required for a voltage
divider.
The recovered pulse output from
pin 13 of IC2 is applied via capacitor
C5 to the emitter of Q2 and to the
inverting input, pin 6, of IC3a via trim-
The prototype decoder was installed in a Hornby OO scale steam locomotive
and is small enough to fit into some N-scale locomotives. Since the output
transistors are driven in linear mode they need to be mounted on the locomotive
chassis for heatsinking.
pot VR1 and resistor R4. Normally, the
output of IC2 at pin 13 sits at close to
+5V and since pin 5 of IC3 is at +3.3V,
the output at pin 7 will be low (ie,
close to 0V). Diode D5 conducts and
so pin 6 is also held at +3.3V.
When the recovered pulse is delivered from pin 13 of IC2, pin 6 of IC3
is pulled low (ie, it is a “low-going”
pulse) via VR1 and R4 and so pin 7
goes high. D5 is now reverse-biased
and capacitor C4 charges, pulling
pin 6 lower. At the end of the input
pulse, pin 7 goes low again and C4 is
discharged via D5. In effect, IC3a acts
as an integrator of the recovered pulse
and produces a DC voltage which
is proportional to the width of the
recovered pulse.
Diode D6 and capacitor C6 act as a
peak detector or “sample and hold”
circuit. C6 is charged to the peak of
the integrator’s output and again, the
DC voltage across it is proportional to
the width of the input pulse. C6 needs
to be partially discharged each time
a new input pulse appears because
the new pulse may be narrower, corresponding to a new speed and direction setting. This partial discharge is
achieved with Q2 because its emitter
is fed with the input pulse from IC2.
Q2 acts like a grounded base stage,
turning on briefly when its emitter is
pulled low via C5, which enables it
to discharge C6.
Op amp IC3b acts as a unity gain
buffer for the sample-and-hold circuit
which drives the output amplifiers,
IC3c and IC3d. However, even this
part of the circuit is not as simple
as it appears. IC3c is connected as a
non-inverting amplifier and is biased
to +5V from the 3-terminal regulator.
By contrast, IC3d is wired as an inverting amplifier and its pin 3 is also
biased to +5V. Both IC3c & IC3d have
a gain of about 3.8.
Linear drive
Now when the output of IC3b is
around +6.5V no power is delivered
to the motor because the voltage
difference between pins 1 and 14 is
insufficient to bias on the respective
output transistors. Q3-Q6 look like
an H-bridge configuration as used in
the original decoder featured in May
1998 but the circuit function is more
akin to a push-pull complementary
emitter follower setup. When the output of IC3c goes up, IC3d goes down
and motor current flows via Q3 & Q6
while Q4 & Q5 are held off.
Similarly, when IC3c’s output goes
down, IC3d’s output goes up and
motor current flows in the opposite
direction through Q4 & Q5 while Q3
Resistor Colour Codes
No.
1
1
1
1
2
2
1
1
1
Value
150kΩ
39kΩ
27kΩ
22kΩ
10kΩ
8.2kΩ
1kΩ
470Ω
220Ω
4-Band Code (1%)
brown green yellow brown
orange white orange brown
red violet orange brown
red red orange brown
brown black orange brown
grey red red brown
brown black red brown
yellow violet brown brown
red red brown brown
5-Band Code (1%)
brown green black orange brown
orange white black red brown
red violet black red brown
red red black red brown
brown black black red brown
grey red black brown brown
brown black black brown brown
yellow violet black black brown
red red black black brown
February 1999 43
Parts List
1 PC board, 64 x 16mm, code
09102992
1 PC board 15 x 16mm, code
09102991
1 25kΩ top adjust miniature
sealed trimpot (VR1)
The prototype PC board shown
here has been redesigned so that
parts no longer sit on top of the
ICs. The four output transistors
were directly bolted to the
chassis diecasting along with mica
or insulated heatsink washers and
connected to the decoder board
via flying leads.
& Q6 are held off.
Note that while two transistors are
always off, the other pair are driven
in linear mode instead of switch mode
so they will get hot, depending on the
amount of motor current.
The other point to consider is that
the motor does not get pure DC but
a portion of the track voltage. For
example, at full speed, the motor will
get about 9V DC plus the superimposed pulse waveform although its
amplitude is reduced in proportion.
In practice, this does not effect the
motor operation at all and it behaves
as though it is fed with pure DC.
Decoder PC board
The photos in this article show the
prototype decoder built into a Hornby
OO scale steam locomotive. This is a
tender-drive loco (ie, the motor is in
the coal tender) and so the decoder
has to fit in the limited space inside
the boiler. As built, the main decoder
board is mounted on the chassis while
the four output transistors dispense
with a PC board. Instead, they are bolted directly to the chassis diecasting
along with mica or insulated heatsink
washers and with flying wires back to
the decoder board.
We have redesigned the prototype
board so the layout shown in Fig.2 is
somewhat different to that shown in
44 Silicon Chip
the photos. The main decoder board
measures 64 x 16mm (code 09102992)
while the optional output transistor
board measures 15 x 16mm (code
09102991). In addition, the decoder
board may be cut in two and installed
in different parts of the locomotive,
with wires linking the two, if that is
necessary to fit it in.
Because both boards are so small,
you will need to take great care when
assembling them; the risk of solder
shorts bet
ween tracks is high. You
will need to use a temperature-controlled soldering iron with a small tip
and be very carefull when soldering
to the small IC pads. We have used
small pads for the ICs to allow tracks
between pins and for close component
spacing.
Ideally, you should also use an illuminated magnifier for this close and
detailed work otherwise you are asking for trouble. Follow the diagram of
Fig.2 exactly, particularly with regard
to the orientation of the resistors and
other vertically mounted components.
The bridge rectifier is tricky since
the diodes are mounted vertically to
save space. Note that their pigtails
should be kept as short as possible
as well. The anodes of one pair of
diodes connect to the 0V rail while
the cathodes of the other pair connect
to the V+ rail. The two wires from the
Semiconductors
1 555 timer (IC1)
1 74C193, 40193B
programmable up/down
counter (IC2)
1 LM324 quad op amp (IC3)
1 78L05 3-terminal 5V regulator
(REG1)
1 10V 400mW or 1W zener
diode (ZD1)
2 BC548 NPN transistors
(Q1,Q2)
2 BD433 NPN transistors
(Q3,Q4)
2 BD434 PNP transistors
(Q5,Q6)
4 1N4004 silicon diodes (D1-D4)
2 1N914, 1N4148 signal diodes
(D5,D6)
Capacitors
3 1µF 35VW tantalum
electrolytics
3 .01µF monolithics
1 .0022µF greencap (metallised
polyester)
Resistors (0.25W, 1%)
1 150kΩ
2 8.2kΩ
1 39kΩ
1 1.2kΩ
1 27kΩ
1 470Ω
1 22kΩ
1 220Ω
2 10kΩ
track (actually from the locomotive
wheel collectors) to the bridge rectifier
are made as aerial connections to the
paired diodes, in agreement with the
circuit of Fig.1.
The capacitors need to be as small
as possible and that means tantalum
for the 1µF units, monolithic for the
.01µF units and greencap for the
.0022. Other types will not fit.
When the boards are complete you
will need to temporarily connect a
motor and power up the power station. The encoder and decoder must
be set to the same channel. The full
procedure for setup and programming
is the same as described in the May
SC
1998 issue of SILICON CHIP.
|