|
|
This is only a preview of the May 1996 issue of Silicon Chip. You can view 24 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 "High Voltage Insulation Tester":
Articles in this series:
Articles in this series:
Items relevant to "KnightRider Bi-directional LED Chaser":
|
RADIO CONTROL
BY BOB YOUNG
Multi-channel radio control
transmitter; Pt.4
In Pt.4 this month, we look at the features of the
transmitter PC board and discuss its assembly. It is
a double-sided board with plated-through holes. It
has conventional components on the ground plane
side and surface mount components on the other.
When I am writing for SILICON CHIP,
I am conscious of the fact that it is an
electronics magazine and not a modelling magazine. I know full well that
SILICON CHIP readers devour all kinds
of articles and the knowledge gained
is often applied in fields other than
originally intended.
This was driven home to me in no
uncertain manner when I presented the
Speed1B motor control in November
& December 1992. More of those units
went into non-modelling applications
than into models. They found their way
into electric powered fishing dingys,
full size autogyros, wheelchairs and a
myriad of other items.
The demand was so widespread
that I was forced to design an add-on
pulse generator to allow these units
to be used without a radio receiver.
This was published several months
later and required a new, small circuit
board which is glued to the original
Speed1B PC board.
When I had finally settled on the
circuit for the Mk.22 RF module, I
sat back and contemplated what the
readers might hit me with this time.
Two days after the encoder circuit was
published in the March 1996 issue, I
had a request from a government de
partment for transmitters and receivers. The Mk.22 was of great interest,
Fig.1: this diagram shows the component layout for the
surface mount component side of the board. Crystal X1
and trimmer capacitor are also mounted on this side.
they said, because the RF modules
came out so easily and this would
allow them to use a fibre optic link
without high drama.
So it was obvious that it was going to
be Speed1B all over again. This time I
was determined to be one jump ahead.
One obvious request would be for
an NBFSK (erroneously referred to as
FM in the modelling trade) transmitter.
Another would be for a voice modulated unit. There is a very interesting
band on 30MHz which butts up against
the 29MHz band (it actually starts
where the modelling bands stops) and
allows the use of 100mW unlicensed
transmitters for voice.
We already make an FM simplex
radio link for this band for several
non-modelling customers. A data
link is an almost certain application.
Another was obviously the use on frequencies not approved for modelling,
but for which the potential user was
already licensed.
Thus, the PC board presented this
Fig.2: the component layout for the groundplane
side of the board. Note that the crystal socket is
attached from this side – see photo.
May 1996 53
The groundplane side of the transmitter board carries only a few parts. Note
the upside down crystal socket. Note also that the pins for TB3, adjacent to
transistor Q2, have been clipped off flush with the board surface.
capacitance for the modulation wave
shaping and are best lumped in with
the temperature stable mylar (polyester) capacitors on the encoder PC
board. C11 and C15 have been reduced
to NPO .001µF capacitors.
Production spreads on the FET
have since dictated that R7 should be
56kΩ or less. This plays a part in the
modulation pulse shaping. During
the final PC board layout I was forced
to add a jumper in the form of a 1206
chip resistor. This is shown on the
overlay as R11 and is 1Ω. There was
also a typo on the circuit. Diode D1 is
a BAS16 not BA516.
These circuit additions have resulted in a transmitter which is a delight
to tune and service. One of the big
problems with checking frequency in
a modulated AM transmitter is that
the modulation blanks out some of the
RF and the frequency count is always
low unless you have a gated frequency
counter.
With TB3 in place, no problem. Just
switch off the modulation by shifting
the AM shunt to CW. The result, a
carrier only transmitter on which it
is a snap to check frequency.
Want to check PA current whilst
tuning the PA? No problem, simply
remove the shunt from TB3 and insert
the meter in series with the two CW
pins. Dead easy!
Construction
This photo shows the surface mount side of the board. The only other
conventional components visible are the crystal, trimmer VC2 and the multi-pin
header, TB1. Note also that links across TB2 and TB4 have been installed on
this side of the board rather than on the ground plane side, as depicted in Fig.1.
month is a multipurpose unit and covers all of the above. Frequency range is
from 25-50MHz with suitable coil and
capacitor changes. Accordingly, it has
provision for components which are
not required for this R/C transmitter.
Most of them are of no consequence
for this project, but the programming
pins TB2, TB3 and TB4 need to be
dealt with for circuit continuity.
TB2 is there to program the oscillator for various configurations and
this is hard-wired to the AM position
with a link. TB3 and TB4 are used to
54 Silicon Chip
program the modulator but again TB4
is of no consequence. It is also replaced
with a link. TB3 however is a valuable
asset when testing and servicing the
board as it programs the transmitter
for CW or AM modulation.
Circuit changes
There are some minor component
changes introduced since last month.
R8 has been increased to 22kΩ to
restrict the FET gate bias range and
improve the feel of VR1 when tuning.
C11 and C15 form part of the lumped
The component layout diagrams for
both sides of the PC board are shown
in Fig.1 & Fig.2. I must point out here
that due to the stringent demands
placed upon this module it is best
tuned on a spectrum analyser. For this
reason I strongly recommend that if
you do not have access to a spectrum
analyser, you should buy the module
fully assembled and tuned.
For those not familiar with surface
mount construction techniques, I
would suggest reading the article
“Working With Surface Mount Components”, as featured in the January
1995 issue of SILICON CHIP. You will
need a pair of magnifying spectacles,
a fine-tipped soldering iron and a pair
of tweezers with very fine tips.
Begin by tinning one pad at each
of the surface mount com
ponents
positions, as set out in the above
article. This is a good time to clearly
establish which components are not
mounted by not tinning the pads for
these components.
When all of the surface mount
components are in place, solder the
jumper links as indicated on Fig.1.
These may be made from the tinned
leads of resistors. The longer jumper
between J1 is made from the wire-wrap
wire provided in the kit.
Note that L1 and L3, which appeared on the circuit last month, are
not used. Their positions on the board
are actually bridged by the copper
tracks out so you don’t have to worry
about them.
Coil winding
The coil winding details for L2, L4
& L5 are shown in Fig.3. The direction
of winding is not important but the
number of turns are. However, it is
important that the secondary on L5
is wound in the same direction as the
primary. The enamelled copper wire
provided is easy to solder and a hot
iron will soon burn the enamel away.
Tin one end of the lengths of enamelled wire provided. Only tin about
1mm of the wire to minimise the risk
of a shorted turn on the coil. Due to
the fact that 16 turns just fit on the coil
formers, snip off half of the pin protruding on the winding side of the coil
base on L2 & L5, leaving just enough
pin to solder the wire – see Fig.3.
Solder the end of the wire to the
appropriate coil former terminal and
wind on the correct number of turns
using tight, close spacing. This done,
apply a dab of super glue to the winding to hold it into place, then place the
coil former on the desk to dry. When
you return, remove the desk from the
coil former.
Having gone through the above
ritual you now have three coils with
one end free. Solder this end to the
appropriate terminal and mount L2
and L4. Now wind on the secondary
of L5. Care must be taken here with
the beginning and end terminals (see
Fig.3) and also to ensure that the secondary is wrapped over the eighth and
ninth turns of the primary. The physi
cal location of the secondary plays an
important role in the drive level and
thus harmonic content of the output.
Secure it after it is wound with another
drop of super glue.
Moving the secondary closer to the
base of the coil (collector of Q1) will
increase the drive level and harmonic
content of the oscillator. Mount L5,
taking care to ensure that the primary
and secondary terminals are correctly
Fig.3: coil winding details for L2,
L4 and L5.
Fig.4: depending on how links are
made across TB3, the transmitter
can be set to CW (no modulation)
or AM (normal operation).
aligned with the PC board (primary
terminals closest to the crystal socket).
Finally, solder the shield into place,
making sure that the coil former is
centralised in the top hole.
Crystal socket
At this point, it is wise to deal with
the next messy job which is mounting the crystal socket. The PC board
is designed to allow the crystal to be
removed from the back of the case
and thus the crystal must be mounted
vertically.
However, this dictates that the socket must be glued into the PC board
flush with the top (surface mount side)
of the board. Thus, viewed from the
ground plane side of the PC board, the
crystal socket appears to be upside
down. Do not get this wrong. If you
glue the socket into the wrong side
of the PC board you will have ruined
both items.
Fit the crystal socket into the hole in
the PC board and ensure that it is the
right way up and flush with the surface
mount. Very carefully place a drop of
super glue onto the junction of the
PC board and the crystal socket from
the groundplane side of the PC board.
Once the glue is dry, solder the
two connecting wires into the pads
adjacent to the crystal socket and then
solder them to the socket terminals.
Care must be exercised here for the
plastic used in the socket is easily
melted. Tin both the socket terminals
and the wire ends before soldering
them together with just a quick dab
of the iron.
The rest of the assembly is a snap,
with the only special care needed with
terminal blocks TB1 & TB3 and trimmer capacitor VC2. VC2 is mounted on
the surface mount side of the PC board
for ease of adjustment when tuning.
TB1 is likewise mounted on the SM
side of the board and mates with the
main power connector for the module.
The programming pins for TB3 are
mounted from the groundplane side of
the board with the long side of the pins
projecting through the board and out
onto the SM side of the board. Solder
them to the pads and then remove
the black plastic from the back of the
board. Snip off the pins on the ground
plane side of the PC board as close to
the board surface as possible. This
also applies to the pins on TB1, as the
antenna sits in the channel between
the components and quite close to the
PC board. These pins could short out
the antenna if left too long.
Finally, mount the output FET
using the hardware kit provided. The
mounting of this transistor is designed
to heatsink the transistor firstly into
the groundplane of the PC board and
then from there into the transmitter
case via the mounting brackets. As a
result the transistor runs quite cool,
even with the antenna retracted.
That completes the assembly. Put
the unit aside for a period then come
back and check once more that all
components are correct. Ensure that
the crystal socket is adequately anchored and that the contacts are free
of glue. Plug in the crystal and place
the micro shunt onto the CW position
on TB3 (see Fig.4).
Testing & tuning
This section is a little ahead of
itself as the module really cannot be
completely tested and tuned until
mounted into the transmitter case
with the correct antenna. However,
I will complete the tuning sequence
for those using the module in other
applications. This description will
assume that the module is in the case
and fitted with an antenna 1.5m long
(wire or telescopic).
May 1996 55
I reiterate that unless you have
access to a spectrum ana
lyser, you
really can’t set up this transmitter
module. However, I am presenting
the following details for the sake of
completeness.
First, with a continuity meter test
between the power and ground pins on
TB1 to ensure that there is not a direct
short to ground. Remove the crystal,
hook up the main power connecter or
apply 9.6V to the power and ground
pins of TB1. Set VR1 to mid-range and
screw the tuning slugs into the coil
formers so they are flush with the SM
side of the PC board.
Remove the micro shunt from TB3
and connect a milliammeter (200mA
range) in series with the CW pins then
turn on the power. The PA current will
be somewhere in the order of 15mA.
Set VR1 for a quiescent current of
12.5mA. This should equate to a base
bias voltage of 2.2V approximately.
Remove the meter and replace the
micro shunt on the CW position.
Now quickly go over the board and
check the voltages at the supply rail
(+10.3V), decoupled oscillator supply
rail (+9.54V), base of Q1 (+3.7V), emitter of Q1 (+3.0V), base of Q3 (+2.2V)
and collector of Q3 (+10.37V).
Plug in the crystal and hook up an
oscilloscope to the collector of Q1.
There should be a strong 29MHz signal
present at the collector. Screw the slug
out (anticlockwise) watching for an
increase in amplitude of the 29MHz
signal until it drops abruptly. Screw
the slug in (clockwise) until the oscillator starts and continue on for one half
turn. At this point you should have
about 5V RF signal at the collector of
Q1. The oscillator is now tuned.
Check the frequency with a counter
to ensure that you are within ±1.7kHz
Kit Availability
Kits for the Mk.22 transmitter are available in several different forms, as follows:
Fully assembled module (less crystal) .......................................................... $125.00
Basic kit (less crystal) ...................................................................................... $89.00
PC board ......................................................................................................... $29.50
Crystal (29MHz) ................................................................................................ $8.50
Post and packing of the above kits is $3.00. Payment may be made by Bankcard,
cheque or money order payable to Silvertone Electronics. Send orders to Silvertone
Electronics, PO Box 580, Riverwood, NSW 2210. Phone (02) 533 3517.
of the marked crystal frequency. The
final frequency will depend on the
brand of crystal you have purchased.
The Showa crystal supplied will be
within tolerance. The frequency may
be fine tuned with C2; increasing C2
will decrease the frequency. Do not
exceed 33pF for this capacitor.
Set VC2 to mid-range and, using a
wave meter, field strength meter or
spectrum analyser, tune L2 and L4
for maximum amplitude of the output
signal. At this point I should point
out that the aim here is not to tune
for maximum power but to achieve
a balance between output power on
the fundamental frequency against
harmonic content. This is the problem that arises when tuning without
a spectrum analyser.
To further complicate the tuning
process, VR1 is best set by tuning it
for minimum third order levels. It is
impossible to do this without a spectrum analyser.
Once the transmitter is at maximum
output, take note of the harmonic
levels. VC2 is fitted for suppression
of 60MHz and 90MHz harmonics. Adjust VC2 for the minimum harmonic
levels and then retune L2 and L4 for
the maximum difference between
fundamental and harmonic outputs.
It should be possible to exceed -60dB
on all harmonic levels. At this point,
the PA current should be about 65mA.
Next, set up a second transmitter
at a frequency 60kHz away, with the
modulation removed (CW) and of
approximately equal output to the
Mk.22 transmitter. Place it on a bench
with the antenna fully extended and
switched on. Switch on the Mk.22 and
position it so that a good strong third
order component is clearly visible on
the spectrum analyser display (it will
be the small spike closest to the Mk.22
fundamental spike).
Tune VR1 for the minimum level
of third order intermodula
tion and
move out until the two fundamentals are of equal amplitude. At this
point, the third order intermodulation
component of the Mk.22 should be
approximately 15dB down on that of
the adjacent transmitter.
Go back now and touch up L2,
L4 and VC2 and tuning is complete.
Check the PA current once more to
ensure that it is under 100mA.
Next month, we will discuss the
SC
assembly of the encoder.
Fig.5: here are the full-size etching patterns for the double-sided PC board.
56 Silicon Chip
|
