The Smith-Kettlewell Technical File

A Quarterly Publication of
The Smith-Kettlewell Eye Research Institute’s
Rehabilitation Engineering Research Center

William Gerrey, Editor

Issue: [current-page:title]

Original support provided by:
The Smith-Kettlewell Eye Research Institute
and the National Institute on Disability and Rehabilitation Research

Note: This archive is provided as a historical resource. Details regarding products, suppliers, and other contact information are original and may be outdated.

Questions about this archive can be sent to


Timer Chips

A Smith-Kettlewell General-Purpose Auditory Thermometer Circuit

The Archer 277-1012 Radio Control Set

A Simple-to-Construct Burglar Alarm Circuit

A Feeder System for the Japanese Tubular Solder Guide


By Albert Yeo


These chips contain eight dividers (flip-flops), a
built-in clock oscillator, and a control pin for FM'ing (should
your project require this). Certain of them also have a pin
which provides a regulated 5-volt output. These chips are very
versatile, as you will see.

I recently decided I had the need for an instrument which would
charge a number of nicads, and keep them topped up until I was
ready to use them. The obvious choice is the "Universal Nicad
Battery Charger" (SKTF, Winter 1981). However, when I looked
around for the specific Exar chips for the timing circuit
(XR2242's), I had difficulty in locating them. Then I came
across the uA2240 at Radio Shack (their catalogue number being
276-1735); it was just what I needed. It turns out that there is
a whole family of related chips--some originally by Exar, the
Fairchild uA2240 I came across, and a set of CMOS units made by
Intersil. With these chips being so popular and readily
available, you should have these data in your notes. (Note that
the Fairchild number, "uA2240" as listed here, actually has the
Greek letter "mu" in its prefix; literary Braille forces us to
use the lower-case "u" as a substitute.)

The TTL 2240 Series

The Exar XR2240, its little brother, the elusive XR2242, and
the Fairchild uA2240, all contain a TTL 8-bit counter and a
built-in "time base" (a clock oscillator similar to that in the
555). With their counters being TTL, they have need for a 5-volt
supply, a crude version of which is actually included in the

The XR2242 differs only in that not all of its divider outputs
are made available to chip pins; this results in an 8-pin DIP
form, whereas the others are 16-pin. On the 16-pin units
(including the Intersil CMOS version), the outputs of all
dividers are available externally on pins 1 through 8; they are
labelled Q1 through Q128. Furthermore, they are open collector.
This means that you can short any or all of the outputs together
on a common output bus to get the kind of count you want (any
number between 1 and 256).

In other words, each output pin goes to the collector of an NPN
transistor; the emitters of all eight transistors are grounded.
In order to make an output--or combination of
outputs--functional, you will need a pull-up resistor (10K is

A typical circuit would show more than one output going to an
"output bus"; this bus would then go through a common 10K pull-up
resistor to VCC. This arrangement gives a wired "AND"
configuration. For instance, if you connect the first, fifth and
sixth outputs to the output bus, you get a count of 49 (1 plus 16
plus 32). As soon as this count is reached, the common bus will
jump to VCC, where it can be used to reset the timer.

Because of its internal circuitry, applying power sets all
outputs at logic high. When the trigger goes positive, all
outputs go to logic low and remain there until all the utilized
outputs go high again at the end of the prescribed time. Then
the output bus goes high again and remains there until the chip
has been reset by a positive-going pulse and the trigger has
again been activated. Once the device has been triggered,
further triggering will have no effect until the time cycle is

The on-board 5-volt supply is gotten from an emitter follower
whose input goes to a 7-volt Zener diode. Its precision of
regulation is quite poor, but it serves the purpose of supplying
the TTL counter section, and it provides a convenient point to
which reset and trigger signals can be clamped to 5 volts, a
level they should not be permitted to exceed.

Pin 15 provides access to the output of this emitter follower;
the need for it arises when enabling the internal clock's output.
Specifically, the output of the clock (called a "time-base
oscillator") is open collector, like all the rest. In order for
its pulses to be at the appropriate 5-volt level, it is usually
put through an external 20K resistor to this 5-volt pin.
(Enabling the clock is done, therefore, by putting its output,
pin 14, through 20K to pin 15, the 5-volt "regulated" output.)

The internal "time-base" oscillator is very similar to the
circuit of the old familiar 555. The charge of an external
timing capacitor is made to oscillate between 27% and 73% of VCC
(the full supply, not restricted to 5 volts). These voltage
points are gotten off a string of three resistors, connected as a
voltage divider inside the chip. When the external RC circuit on
pin 13 charges to 73%, the capacitor is made to discharge rapidly
to 27%; during the discharge phase, a short pulse can be seen at
pin 14, the oscillator's output (provided that pin 14 has its
required pull-up resistor to 5 volts).

Like the 555, the top junction of the internal voltage divider
(the 73% point) is made available to the user--it goes to pin 12
on the 16-pin packages. A signal can be capacitively coupled to
this pin so as to frequency-modulate the time-base oscillator.
It is also possible to synchronize the time-base oscillator to a
slightly faster string of pulses; these pulses, which should have
a peak-to-peak value of perhaps 60% of VCC, can be capacitively
coupled to pin 12.

Specifications for the TTL Versions
XR2240, XR2242, and Fairchild uA2240
(Radio Shack 276-1735)

For the benefit of newcomers, I would like to state that
service notes and manufacturers' literature are not written to
help the would-be user of their products; rather they are
intended to give employment to the printing industry's less able
apprentices. At best, they are ambiguous; at worst, they are
downright misleading. The literature for this chip is no
exception. I shall not, therefore, reproduce it here. Instead,
I will just give some useful information.

This chip is typically operated with VCC at 12 volts, although
it can be supplied from 4V to 18V. (For supply voltages less
than 4.5V, it is recommended that the 5-volt pin be tied to the
VCC pin.)

Trigger and reset must not exceed 5 volts. The threshold for
these two inputs is between 1.4 and 2 volts.

Trigger and reset currents are both given as 10 microamperes
with an impedance of 25K.

If the trigger and reset are activated simultaneously, the
trigger will take precedence.

The regulated voltage at pin 15 will be between 3.9 and 5.8
volts, depending on your VCC voltage. Regulator current is 5

As for the built-in clock, the minimum external timing resistor
is 100 ohms, with a maximum of 10 megohms. The minimum value for
the external timing capacitor is 0.001uF, with a maximum of
1000uF. The period of the clock pulse is given by C in
microfarads times R in megohms; the frequency is the reciprocal
of this (1 over R times C).
C goes from pin 13 to ground, while R goes from pin 13 to VCC.

Each of the outputs can sink a maximum of 3.2 milliamps, with
1.4 milliamps being the typical value. This means that with all
eight outputs connected, a current of about 12 milliamps is
available. Maximum chip dissipation is 650 milliwatts. Maximum
supply current is 18 milliamps. Response time is given as 0.8

The output of the dividers is a good squarewave with a maximum
toggle rate of 1.5 MHz.

Of course, more than two chips can be cascaded if you want a
time period of several generations.

With several outputs used, some interesting pulse patterns can
be obtained. The frequency of each pulse in the group will be
that of the output nearest to the clock, while the groups will
recur with a period equal to that of the output which is farthest
from the clock.

Pin Connections
XR 2240, uA2240
Radio Shack 276-1735

  • Pin 1 : output binary 1
  • Pin 2 : output binary 2
  • Pin 3 : output binary 4
  • Pin 4 : output binary 8
  • Pin 5 : output binary 16
  • Pin 6 : output binary 32
  • Pin 7 : output binary 64
  • Pin 8 : output binary 128
  • Pin 9 : ground (minus V)
  • Pin 10 : reset (high for reset)
  • Pin 11 : trigger (high to trigger)
  • Pin 12 : "mod" (control pin for
    FM'ing, etc.)
  • Pin 13 : time base (for external R and C)
  • Pin 14 : output for internal clock and
    input for external clock
  • Pin 15 : regulator output
  • Pin 16 : VCC (plus supply voltage)

Pin Connections

(This little brother works the same, but not all the outputs
are available. A consequence of this is that a count of 256
cannot be gotten by AND'ing the outputs; the best you can get is
128. Note also that the regulated 5 volts is not brought out.
Not only are the trigger and reset clamped to 5V, but the time-
base output is clamped there internally. A 22K pull-up resistor
is needed from pin 8 to VCC to enable this "output.")

  • Pin 1 : VCC
  • Pin 2 : output binary 1
  • Pin 3 : output binary 128
  • Pin 4 : ground (minus V)
  • Pin 5 : reset (high for reset)
  • Pin 6 : trigger (high to trigger)
  • Pin 7 : time base (for external R and C)
  • Pin 8 : output for internal clock and input for external clock

Sample Circuits for the TTL Versions

Circuit for Mono-Stable Operation

Pin 9 is grounded; pin 16
goes to plus 12 volts (VCC). Minus 12 volts is grounded. Pin 10
goes through 47K to the output bus. Pin 11 goes through 47K,
then through a single-pole, single-throw switch to VCC (so that
you can trigger the device manually). (These 47K resistors are
necessary because the trigger and reset pins are clamped
internally to 5 volts.)

Pin 12 goes through 0.01uF to ground. (This bypass suppresses
noise on the internal resistor string which determines the
charge/discharge threshold voltages.) Pin 13 goes through the timing
capacitor to ground, and pin 13 also goes through the timing
resistor to VCC. Pin 14 goes through 20K to pin 15. This will
enable the clock, once the circuit has been triggered.

The output bus goes through 10K to VCC. If desired, each
output (pins 1 through 8) can go through its own single-pole,
single-throw switch to the output bus, so that you can select any
or all of the outputs.

Once the circuit has been triggered, any further closure of the
switch will have no effect until the timing cycle has completed
and the device is reset.

As the clock frequency is reduced, the accuracy of the final
selected time will slip a bit. This may or may not matter,
depending on your application. At best, the accuracy will be
0.5%, at worst 5%. As with the 555 oscillator, the dependence of
the charge/
discharge points is arranged to change with VCC; the result is
that the oscillator frequency is not particularly affected by
changes in the supply.

For any given time setting, accuracy will be best when the
clock frequency is high and the programmed count large. The
accuracy can be further improved by cascading two chips; the
"count" can then be made extremely large.

The makers of nicads give a wide tolerance for charging time
(for instance, 14 to 16 hours for standard units), so I used only
one chip in my charger. This meant applying a clock pulse with a
period of about 211 seconds (0.00474 Hz). One chip can multiply
the period of the clock pulse by 256; two chips can multiply it
by 65,536.

The following circuit shows how to cascade two chips. You will
notice that pin 16 of IC2 is not used; instead the second chip
gets its VCC through the common connection of both pins 15. This
makes for economy of power and greater stability. Also note the
connection of pin 13 on the second chip, which is shunted to
ground by a 1K resistor. Drive to the second chip is applied to
pin 14.

Circuit for Cascading Two Chips

Pins 9 of both chips are
grounded, along with minus 12 volts. Pin 16 of the first chip
goes to VCC. Pin 16 on chip number 2 is not used. Pins 10 of
both units are tied together and go through 47K to the output bus
of the second chip. Pins 11 on both units are tied together and
go through 47K, then through a single-pole, single-throw switch
to VCC for manual triggering.

Each pin 12 goes through its own 0.01uF capacitor to ground.
Pin 13 on IC1 goes through the timing capacitor to ground as well
as going through the timing resistor to VCC. Pin 13 on IC2 goes
through 1K to ground. Pin 14 on IC1 goes through 20K to pin 15
on this chip, which is tied to pin 15 on IC2. Pin 14 on IC2 goes
through 47K to VCC, and this pin 14 of IC2 also goes directly to
the output bus of IC1. All outputs on IC1 (pins 1 through 8) go
to their own output bus, which then goes through 10K to VCC. As
already mentioned, this output bus of IC1 goes to pin 14 on IC2.
All outputs on IC2 go to their own output bus, which then goes
through 10K to VCC.

Testing the Chip

If you want to demonstrate how the chip
works, connect it up for mono-stable operation, but instead of
taking pin 10 through the resistor to the output bus, take it
through 47K, then through another normally open pushbutton to
VCC. Take pin 13 through 0.05uF to ground. Also take pin 13
through 120 ohms, then through a 5K rheostat to VCC. Connect pin
14 through 20K to pin 15.

Connect one end of a 10K resistor to VCC, with the other end
firmly attached to the tip of a braille stylus (to serve as a
test probe). Connect the cold input of your test amplifier to
ground, and connect its hot input lead through 1 megohm to the

Now connect the 12-volt supply and push the button to reset the
circuit. Push the switch that triggers the counter, and with
your probe on pin 1, turn the rheostat until you hear top C (4187
Hz). The clock will now be running at 8372 Hz. Move the probe
to pin 2, and the signal will be exactly one octave lower. Move
to pin 3 to bring the signal down another octave. Continue thus
until you are on pin 8; the signal will now be at 32.7 Hz.

You may think all this is fooling around, but I always feel
more confident in designing a project if I am sure the device can
do what they say it will; listening to it is the most direct way
of testing it. It is the best-known hair remover--finding that
after all your hard building, the dashed thing won't behave

The Intersil CMOS Versions
ICM7240, ICM7250, ICM7260

The ICM7240 is said to replace the 2240 "in most applications."
Being CMOS has implications which should be listed here:

Firstly, no 5-volt requirements or restrictions are necessary.
Pin 15 of the 7240 has no connection. (Remember this when
cascading them; pin 16 must be used to power the second chip.)
Since trigger and reset inputs are not clamped internally, no 47K
resistors need be provided for them; reset can go directly to the
output bus, and the trigger switch can take pin 11 straight to
VCC. The supply voltage can range from 2 to 16 volts.

The outputs are "open-drain"; they still need pull-up resistors
to work (including pin 14 for enabling the clock). The clock
output can only sink 1mA; specifications on the other outputs are
more complicated.

As is often done with CMOS output specs, the manufacturers plot
voltage vs. current; the more current you ask of these drains,
the higher will be their voltage at logic low. With a 5V supply,
each drain pulls down as if it were a 40-ohm resistor; with a 15V
supply, this "resistor" is about 16 ohms. The "resistance"
sharply increases at some point. In addition, there is a maximum
of 50mA which you are not supposed to exceed, as well as a total
dissipation of 200 milliwatts that you are not to exceed.

The ICM7250 AND ICM7260 have BCD counters in them. Two digits
are available: pins 1 through 4 are "1," "2," "4," and "8,"
respectively, while pins 5 through 8 are "10," "20," "40," and
"80." The "carry out" is pin 15 on these chips. The time-base
hookup and all other pins are the same as the 2240's.

The ICM7250 can count from 0 through 99. The ICM7260 is
internally programmed to count from 0 through 59 (ideal for
counting minutes and seconds). The carry bit goes high at 50.

To cascade these, the carry output of the chip whose time-base
is working (this chip being responsible for the least significant
digits) goes through a pull-up resistor to VCC (perhaps 100K); it
is then tied directly to pin 14 (the counter input and disabled
clock output) of the following chip.

It should be mentioned that the 2240's can also be used to
count through 99 or 59, and this may be the cheapest way to go.
To make these reset at 60, for example, pins 3, 4, 5, and 6 (4
plus 8 plus 16 plus 32 equals 60) can be taken to the output bus,
leaving pins 1, 2, 7, and 8 unconnected. A second chip could be
coupled to this output bus by way of pin 14, as described

Have fun with these excellent chips.

[It seems that on the U.S. side of the pond, Radio Shack does
not stock these items. However, the editor was surprised to find
all but the ICM7260 available from Jameco. The 7240 is about
three times the price of the 2240, but it's still only five


by Tom Fowle, WA6IVG


A simple electronic thermometer is described using a
common transistor as a temperature sensor. This circuit is
particularly suited for use with various audible meter readers.

Over many years several of us have seen references to circuits
which use the temperature coefficient of the base emitter
junction in a transistor as a sensor. However, trying this out
is just one of those many projects we never got around to. The
circuits varied widely in complexity, from just the transistor, a
resistor, a battery and meter, to one which used two op-amps and

a plus and minus 9-volt supply.

In planning the details of how to build such a unit, including
a dedicated audible meter reader, I found that we could end up
with three 9-volt batteries; this struck me as just a bit much.
So I enlisted the help of our ever faithful Al Alden and Jay
Williams for prototyping this circuit.

This one uses one battery for both the temperature sensor and
its op-amp. As for the meter reader, you can build and calibrate
this thermometer portion, and then use whichever meter you happen
to have available (even a talking meter, if you haven't gotten
tired of its constant nagging). Use with a built-in meter
reader, probably the better way, will be included here.

Calibration does require that you establish two known, fairly
solid temperature references. This will probably mean borrowing
a sighted friend with a thermometer, but there are a few tricks
you can try if you have a freezer and a stove. (Don't worry,
this will be described later.)

The sensor transistor is a 2N2222, or just about any old NPN
silicon unit. The temperature coefficient of the transistor's
base-emitter junction is negative; thus the base emitter drop
goes down about 1.6 millivolts per degree Fahrenheit of
temperature increase. What's more, this function seems to be
pretty linear throughout any thinkable temperature range. It is
true, however, that the actual voltage gotten from any two
transistors at a particular temperature will vary. Therefore, if
you wish to have several sensors--say, one outdoors and one near
the unit--you will need to match them. This is done using the
completed unit itself, and will be discussed after the circuit.

Circuit Operation

The sensor transistor has its emitter
grounded. Its collector and base are connected together and go through a resistor to the plus V line; thus, the base-emitter
junction is forward-biased by a current source. (The voltage at
the base of the sensor transistor is 700 or 800 millivolts, and
the resistor supplying current for the sensor is fairly large.
Therefore, the resistor becomes a current source and the dying
battery makes little difference in the current through the
transistor and, therefore, the voltage across its junction.)

This grand junction goes to the inverting input of an op-amp
with a gain of about 4, and has a voltage reference on its non-
inverting input. Specifically, the non-inverting input of the
op-amp goes to the arm of a pot whose bottom is grounded; the top
of this pot must be connected to a temperature-stable voltage
reference so as to keep the dying battery from changing the
circuit's calibration. This pot lets you set the output of the
op-amp to 0 volts--for whatever the coldest temperature you want
to measure may be.

The output of the op-amp goes to the top of another control
whose bottom is grounded; the arm of this control is the
thermometer's output. This "scaling" control sets the overall
gain of the system. This output can be connected to a high-
impedance voltmeter--or, via a voltage follower, to any

As the base emitter drop of the transistor increases with
decreasing temperature (about 1.6 millivolts per degree
Fahrenheit), the output of the amplifier rises four times as
much. The pot on the non-inverting input, "zero adjust," is set
to match your lowest desired reading--or your cold standard. The
pot which scales the output for the volt meter, "temperature-span
adjust," is set for your maximum reading--or hot standard. These
controls interact, so some fiddling will be necessary.

The transistor can be attached to a cable, or mounted directly
into a plug. Selecting matched transistors, you can stash a few
at strategic points around the place for quick checking--using a
switch to interrogate them.
Use a twisted pair for connecting them at any distance.

If you are building-in a dedicated meter reader for this unit,
the temperature-span adjust pot can be eliminated, and the output
of the op-amp can go directly to the plus meter reader input. In
this case, the meter reader's calibration control takes the job
of setting the upper limit (and span) of your temperature range.

There is another advantage in having a dedicated meter reader
for this device; the temperature-stable reference voltage can be
used in the meter reader for its calibration reference. This is
the recommended procedure, since the common old zener we usually
use turns out to drift all over the map with temperature. You
can't measure temperature unless you have something temperature-
stable to measure with. For this purpose, we use the LM336
temperature-stable 2.5 volt reference diode.


A 9-volt battery has its negative at ground, and its
positive going through an SPST on-off switch to the plus 9V line.
An LM358 dual op-amp is used. Pin 4 of the LM358 is grounded,
while pin 8 goes to plus 9V. If a high-impedance voltmeter is
available, the second op-amp has pins 5 and 6 grounded, so as to
put it out of commission.

The sensor transistor, a 2N2222, 2N3904, or equivalent, has its
emitter grounded, while its base and collector are tied together
and go through 15K to plus 9V. The junction of the transistor
and resistor goes through 5.1K to pin 2, the inverting input of
the first op-amp in the package. Between pins 1 and 2 is a 20K
feedback resistor.

Pin 3, the non-inverting input, goes to the arm of a 10K 10-
turn PC-mount pot (zero adjust). The bottom of this pot is
grounded, while its top goes to the reference voltage. This
reference is the cathode, or "plus" terminal, of an LM336 diode.
This diode has its anode, "minus" terminal, grounded. Its
"adjust" pin is not used (see below). The cathode also goes
through 2.2K to plus 9 volts.

Pin 1 of the op-amp, its output, goes to the top of another 10K
10-turn PC-mount pot ("temperature-span adjust"). The bottom of
this pot is grounded, while the arm can be considered as the
thermometer output. If something other than a high-impedance
VTVM is used, the arm of this adjustment goes to pin 5 of the
LM358. Pins 6 and 7 are connected together--pin 7 being the
output of a voltage follower which can drive any voltmeter.

For an internal read-out system, I chose to use the meter
reader circuit which was published under the moniker "The Fowle
Gimmique" (SKTF, Summer 1982). With this Gimmique system (a play
on words relating to the much simpler "Auditory Gimmick"), the
pitch of an oscillator indicates the relative temperature
reading. An actual quantified reading can be taken by turning a
braille-calibrated control until the tone of the oscillator
"pulsates." The point where pulsation begins is determined by a
comparator to be the point at which the voltage of the calibrated
standard matches that of the thermometer's output. The circuit
for the "Fowle Gimmique is given in brief here for the full

Those familiar with this meter reader will note a couple of
omissions--such as RF filters not necessary here--and the offset
pot on the VCO which is not necessary because you never need the
oscillator to go to zero (by then, you would be so cold you
wouldn't care what the beeper was doing anyway).

The input to the meter reader is taken from pin 1 of the
thermometer's LM358 (as mentioned, the "temperature-span
adjustment" is omitted). That pin 1 goes through 220K to pin 3
of another LM358 op-amp, with pin 3 going through 0.1uF to ground. The op-amp has pin 4 grounded and pin 8 taken to plus V.
The thermometer's pin 1 also goes through 100K to pin 5 of the
latter LM358, and pin 5 goes through 0.1uF to ground.

Pin 6 of the latter LM358 goes to the arm of a 10K precision
linear pot (fitted with a pointer and braille dial). The bottom
of this pot is grounded. The top of this pot goes through a 100K
10-turn rheostat to the cathode of the LM336 reference diode.
(This 100K rheostat, calibration for the Gimmique, is the new
"temperature-span" adjustment.) The cathode of the LM336 also
goes through 2.2K to the plus V line.

Pin 1 of this chip goes to the base of a 2N2222, whose emitter
goes to its pin 2. The emitter also goes through 1K in series
with a 25K rheostat to ground. The collector of the 2N2222 goes
to both the base and collector of a 2N2907, and also to the base
of a second 2N2907. The emitters of these two transistors are
tied together and go to pin 9 of an NE556 dual timer chip.

Pin 7 of the op-amp goes through 22K to pins 8 and 12 of the
NE556; these pins also go through 10K to the 556 pin 13. Pins 8
and 12 also go to the positive of a 2uF electrolytic capacitor
whose negative is grounded.

The free collector of the second 2N2907 goes through 22K to
pins 2 and 6 of the NE556, and thence through 10K to pin 1. Pins
2 and 6 also go through 0.0047uF to ground. The NE556 has pins
4, 10 and 14 tied together and going through 10 ohms to plus V,
which is bypassed to ground by 100uF (negative at ground). Pin 7
is grounded.

Pin 5 goes through a 47-ohm 1/2-watt resistor, thence through
the speaker coil to plus V.

The LM336 temperature-stable reference diode has three leads
and comes in a TO92 package (this is the little partial cylinder). With the leads up and the flat side toward you, the
connections are, from left to right: ground (anode), plus
(cathode), and adjust. The best temperature stability of this
unit is when the voltage across the diode is 2.49 volts. If you
get one which comes close to this, you are lucky. If the voltage
across the unit lands at more like 2.41V, you are in the worst
case area of the curves. In this case you might wish to add
the simple adjust circuit and bring the thing up to 2.49V to
improve its stability.

Adjust Circuit for the LM336

A pot having a nominal value of
5K or so is connected in series between two diodes across the
336. In other words, the plus goes to the anode of a 1N914 diode
whose cathode goes to the top of a 5K pot. The bottom of this
pot goes to the anode of a similar 1N914 whose cathode is
grounded. The arm of the pot goes to the adjust pin of the

Matching Sensors

If you are going to use several sensors--and
why not plan on doing so; you will want to some day anyhow--you
will need to match up at least two transistors. So when getting
parts, buy a basket full of 2N2222's; the ones in metal cans cost
a little more, but should settle out to the temperature of their
surroundings more quickly. The instrument itself can be used to
select matched units.

Now build the machine, but don't attempt to calibrate it yet.
Have the braille dial set up so you can take readings. Put a
couple of clips on the input terminals, and turn the beast on.
Grab a transistor and clip its emitter to the ground clip, and
clip both its base and collector to the "hot" clip.
Now turn the zero adjust pot until you get some moderately
annoying pitch to listen to. Twist the calibrated pot and try to
find a point where the pulsating stops and the tone becomes
steady. If the tone never pulsates, the temperature-span
adjustment (now the 100K rheostat in series with the braille-
calibrated pot) needs to be turned to less resistance. If it
never starts pulsating, you need higher resistance in this
calibration rheostat.

The 25K rheostat in the emitter of the meter reader's 2N2222
affects only the range of pitch of the oscillator. You can
adjust this to your liking, once you put the instrument into
service. For the matching of transistors, you may want to at
least set it so that saturation of the pitch at a high frequency
never occurs during the tests.

Take hold of the sensor transistor, or put it in a cup of cold
or hot water. If the pitch of the tone changes, the beastie
basically works. If the pitch doesn't move, but the "pulsation
point" on the dial changes, then mess with the 25K rheostat in
the emitter of the Gimmique's 2N2222. Pick a stable medium--a
cup of water--and note the reading after things settle out. Do
with a bunch of transistors and see if you can find a few


You can calibrate this unit for just about any
range you want; the main consideration is accuracy. For
instance, if you want a range of minus 30 degrees F. to 120
degrees F., your scale would have to cover 150 degrees. This
would mean that you need a dial and calibrated pot capable of
greater than 1 percent accuracy to read down to 1 degree. Figure
out what range you want, then what precision is practical with
the pot and dial-making systems you have (or the resolution of
the meter you intend to use).

Next, you can begin worrying about how to develop standards
within that range; best results will be gotten with temperature
standards near, or at, each end. If you don't have access to a
good thermometer and reader, the ice and boiling baths are a
couple of pretty good standards to fiddle with.

Ice Bath

I am told by our scientific types that a bath of ice
in water, which is allowed to stand for a few minutes, will
produce a pretty reliable 32 degrees Fahrenheit. This should be
at, or near, enough the bottom of
your range, unless you live somewhere that has real winters.

Hot Bath

If you boil water, preferably distilled, in a clean
glass pot, its temperature will be a pretty good 212 degrees F.
(This shouldn't hurt the transistors, but I can tell you it's
hard on fingers.)

Once standards have been established near the ends of your
desired range, go back and forth between them, adjusting the two
pots until things fit your dial. In other words, put the
transistor in the ice bath and set the zero adjust to get a
reading of 32 degrees. Now put the transistor in the hot bath
and, after screaming, adjust the temperature-span control to get
this point correct on your dial. (Remember that the "temperature-
span adjustment" is now the 100K rheostat in series with the
brailled pot.) Now go back and do them both again several times
until they are correct (they will interact, so three or four
successive attempts will be necessary).

Note that when you put the sensor in water, the change in
reading occurs pretty quickly, and the thing settles down within
ten seconds or so. But when you just leave it out in the air, it

can take five minutes or more to get settled so as to agree with
a nearby visual thermometer. This is due to poor coupling
between the air and the sensor. Don't let this worry you; visual
thermometers do the same thing--especially mercury ones, which
have a comparatively high heat capacity.

For my calibration, I had access to an AFB talking thermometer
with which I measured the temperature of a cup of water that was
just good and hot, as well as that of an ice bath. I chose to
set up a scale covering minus 30 to plus 130 degrees F. I have a
fancy dial maker, and I used a high-precision pot, rated for a
linearity of plus/minus 1/2 percent; my overly optimistic
markings every 2 degrees are proving to correlate rather closely.

If you really want to resolve every degree, you'd better plan
on a range of 50 degrees or less, because you probably can't get
a dial and pot combination having a resolution and accuracy of
less than two percent.

Parts List
Basic Thermometer

  • Resistors (1/4-watt 5%, unless otherwise stated):
    • 1--2.2K
    • 1--5.1K
    • 1--15K
    • 1--20K
    • 2--10K 10-turn PC-mount pots
  • Semiconductors:
    • 1--LM336 precision voltage standard
    • 1--LM358 dual op-amp
    • 1 or several matched units--2N2222, 2N3904, or any
      general-purpose NPN's.
    • Metal cans preferred (see text).
  • Miscellaneous:
    • Appropriate plugs and jacks
    • SPST on-off switch, plus appropriate switch for multiple
    • 1--9V battery connector

Parts List
Incorporated Read-Out
(The "Fowle Gimmique")

  • Capacitors:
    • 1--0.0047uF mylar or mica
    • 2--0.1uF disc ceramic or mylar
    • 1--2uF or 2.2uF 10V electrolytic
    • 1--100uF 10V electrolytic
  • Resistors (1/4-watt 5%, unless otherwise stated):
    • 1--10-ohm
    • 1--47-ohm, 1/2-watt
    • 1--1K
    • 2--10K
    • 2--22K
    • 1--100K
    • 1--220K
    • 1--10K precision linear pot, fitted with a Braille scale
    • 1--25K 10-turn PC-mount pot, used as a rheostat
    • 1--100K 10-turn PC-mount pot, used as a rheostat
  • Semiconductors:
    • 1--2N2222
    • 2--2N2907's
    • 1--LM358 dual op-amp
    • 1--NE556 dual timer
    • (Note that the LM336 also appears in this circuit, although
      this unit is common to both circuits.)

  • Miscellaneous:
    • (The switches, plugs and jacks, and the battery connector
      are common to both circuits.)
    • 1--small loudspeaker



This set of circuit boards from Radio Shack (No. 277-
1012) is designed for remote control of toy cars. Since it is
now available as a "component," its low cost (about $16) makes it
appealing for other applications: responder beacons, athletic
equipment, alarms, and simple environmental control systems are


The set comes in the form of two circuit boards, a long slender
transmitter board (1 by 3-3/4 inches), and an almost-rectangular
board of 1-1/2 by 3 inches (one long side having the corners
approached by slight-angle cuts). Radio Shack calls this a
"four-function radio controller"--forward, back, right and left.
I would say it has about
2-1/2 functions for the following reasons:

One of the "detector" circuits in the receiver is looking for
audio noise of any form; this noise directs the main drive motor
forward. The receiver itself is a super-regenerative type (what
we used to call a "rush box" in the early days of VHF radio); its
self-oscillations produce a loud rushing noise until a signal
comes along to "quench" regeneration, thus bringing on a blessed
quietness. As soon as the receiver is turned on--with the
transmitter turned off--its own noise directs the toy car
forward. Presenting it with a dead carrier causes the drive
motor to be reversed; a lack of audio noise causes this reversal,
not true detection of the carrier.

The transmitter is able to generate three kinds of signals: a
dead carrier (reverse), a carrier modulated at 500Hz (left), and
a carrier modulated at 3000Hz (right). The presence of either
tone, while being analyzed by a low- and high-pass filter to
reverse the drive to the steering motor, is seen as "noise" to
the controller of the drive motor; thus, the car moves forward as
well as turns.

Therefore, the four functions are not independent, since the
"forward" one always responds to "steering." What's more, this
receiver is listening on the main Citizens' Band, and being a
simple "rush box," it's broad as a hippopotamus and can be
by any signal--any dead carrier will send you in reverse.

It may sound like I'm giving this set of boards a bad review.
Actually, I am not; its give-away price and ready availability
make it alluring in the worst of cases. It should be used in non-life-or-death applications. (For example, we designed a set
of two remote-controlled bases for "Beep Baseball" using this
product.) On the other hand, don't use this simple set to
control your wheelchair, or you'll do the Hula to "Whistlin'
Pete" on Channel 16.

These boards are of by-gone technology--not a chip to be seen.
Since they contain literally dozens of transistors, their full
circuit diagrams will not be presented here, although enough
specific information will be given to intelligently get on and
off the board with control switches and items to be driven.
(Full circuit information is given in the pamphlet that comes
with the set.)


  • Transmitter:
    • Operating Voltage--6 to 9.5 V
    • Current Drain at 9V--25mA max., 18 nom.
    • Operating Frequency--27.145MHz
    • Left-Turn Signal--500Hz AM
    • Right-Turn Signal--3000Hz AM
  • Receiver:
    • Receiver Operating Voltage--6.5 to 9.5 V
    • Receiver Operating Current--18mA max.,
      12mA nom.
    • Motor Driver Voltage--3V (two flashlight cells)
    • Driver Motor Current--300mA for "main drive"; 200mA for "steering"
  • Range:
  • The range of this system is an issue that cannot be treated
    simply. First of all, the choice of band means that interference
    will be significant, and will depend heavily on atmospheric
    conditions. Then, being a long wavelength, structures of
    civilization are very absorptive. The receiver being simple,
    sensitivity will vary widely between sets. With my first set, I
    could only cover a room; with my second, I was able to get semi-
    reliable response from the receiver along the length of my 60-
    foot house.

The Transmitter

This transmitter is crystal-controlled on the frequency of
27.145MHz. Besides this, its board contains a multivibrator whose
frequency can be selected for 500Hz and 3000Hz; this is done by
pulling pairs of biasing resistors high. In addition to
supplying power to the board, a control switch must either ground
the bottom end of an emitter resistor in the final stage, or
connect this emitter resistor to one of the multivibrator's
outputs, thus setting the transmitter up to be amplitude-
modulated at a selected "steering" frequency.

Orient the board, with its component side up, so that the
crystal (the tallest item on the board) is at the bottom-left
corner. This puts one of its long edges toward you; a mounting
hole is provided at either end (the one on the right is not
centered). Although hard to feel through the forest of
components on the top, a large hole is present in the center of
the board.

Above the large center hole are several small holes for
connection to switches. One of these, a hole very near the edge
of the board and 1-3/4 inches from the left end, is "Point 1," a
ground connection; this is not provided for power, but for
grounding the emitter resistor of the final stage. Directly
below "Point 1" is a pair of holes, "Point 2," only one of which
is used; this goes to the bottom end of a 68-ohm resistor off the
transmitter's final emitter (it is to be grounded or tied to the
multivibrator). Just to the right of "Point 2," at the upper
edge of the large center hole and looking like a third in the
two-hole set of "2," is "Point 3"; this is one of the
multivibrator's collectors, which is used to modulate "Point 2."

A large choke stands up at the "one o'clock" position of the
center hole, just to the right of the previous three points.
Standing on one of its leads, the choke's other lead is bent
double and sent down through a hole near the edge of the board.
Even closer to the board's edge, and slightly to the right of the
choke's bent-over lead, is a hole marked "Ant"; it would like to
see 20 inches of wire.

Immediately to the choke's right are two disc capacitors.
Beyond these are two diodes which are lying flat and oriented
vertically. Directly above these diodes, near the edge of the
board, is "Point 4." This point, which adjoins both anodes, goes
through the diodes to a pair of 15K resistors; when tied high (to
the plus 9 volts), this makes the multivibrator run at 3000Hz
("right"). Five-eighths of an inch beyond "4" (and 5/8 of an
inch from the right end) are two 82K resistors standing on end,
their bodies being very near the top edge. Right between them is
"Point 5." This point, which adjoins these 82K resistors, makes
the multivibrator work at 500Hz when tied high ("left"). (There
is no need to take points "4" or "5" low. In fact, you probably
shouldn't in the case of "5"; just leave them open when not in

A hole in the top right corner of the board is ground--the same
as "Point 1"--and goes to the negative of the 9V battery. Ground
is also found at the left-end mounting hole. (Because of a lazy
draftsman, plus 9V appears at the right-hand mounting hole;
thanks, Pal.) The proper connection for plus 9V is 1/4-inch
below the corner ground lead, and slightly farther in from the

Their Hookup

The "forward/reverse" switch is DPDT with no
center off. The arm of section A goes to plus 9V. Position 1 of A (forward) is not used; position 2 of A (reverse) goes to the
"Plus 9V" point on the board. The arm of the B section goes to
"Point 2," the transmitter's emitter resistor. Position 1 of B
(forward) goes to "Point 3," the multivibrator output. Position
2 of B (reverse) goes to "Point 1," which is ground.
The "steering" switch is DPDT; it has a center off and is
momentary in either direction. Positions 1 and 2 of the A
section are jumpered together and go to the "plus 9V" point on
the board (available on A2 of the other switch). The arm of A
goes to the positive of the battery (available at the arm of A on
the other switch). The arm of section B of the steering switch
goes to the "Plus 9V" point, available at either position of its
own A section. On the B section, its "left" position goes to
"Point 5;" its "right" position goes to "Point 4."

Alternative Arrangements

The most obvious thing to do first
is, with three pushbuttons, provide switching to establish the
three transmitter conditions. A DPDT pushbutton can be directly
substituted for the forward/reverse switch; this will be called the "dead-carrier" switch.
Normally-open DPST pushbuttons can do the two tone
functions--"high" and "low." These would be wired as follows:

The swingers of both sections of each "tone" switch go to the
"Plus 9V" point on the board. The "on" position of pole A of
each switch goes to the positive of the battery. The "on"
position of pole B of the "low" switch goes to "Point 5"; the
"on" position of B on the "high" switch goes to "Point 4."

There are applications where you would want a carrier all the
time; buttons would modulate the carrier with "low" and "high"
beeps without interruption. In populated areas where the
Citizens' Band is very active, this reduces false triggering of
the tone detectors in the receiver by coarse voices. Also with
this arrangement, a crude carrier detection can be accomplished
with gates. Thus, if the range of the system were exceeded (if
the blind skier got away from his guide, for example), the bearer
of the receiver would know that he was out of reach. The
following hookup, though unwieldy, will accomplish this:

A simple SPST toggle switch would be used to connect the
positive of the battery to the "Plus 9V" point. (Naturally, the
negative of the battery is grounded.)

Triple-pole, double-throw pushbuttons are needed. This is true
since each "tone" switch has to: (a) tie the appropriate
multivibrator resistors high; (b) take the transmitter's emitter
resistor to the multivibrator; (c) relinquish this emitter resistor from ground, so that it can be modulated.
"Point 2," the emitter resistor, goes to the arms of poles A
and B of the "low" switch, as well as going to the arm of pole B
on the "high" switch. The normally closed position of pole A on
the "low" switch goes to the arm of pole A on the "high" switch;
the normally closed position of pole A on the "high" switch is grounded by way of "Point 1." On both pushbuttons, the normally
open contact of pole B goes to "Point 3," the output of the

On both switches, the arm of pole C goes to the "Plus 9V"
point. On the "low" switch, the normally open position of pole C
goes to "Point 5." On the "high" switch, the normally open
contact of C goes to "Point 4."

Wouldn't it be nice if you could leave the emitter resistor
connected to the multivibrator all the time? You could if, when
the biasing resistors are floating, a multivibrator collector sat
low. No such luck--without modification to the circuit board.
I will leave it up to the ambitious student to try the inverting
circuit that follows:

The transmitter's emitter resistor, "Point 2," goes to the
collector of a transistor whose emitter is grounded to "Point 1."
The base of this inverting transistor goes through a resistor,
perhaps 4.7K, to "Point 3," the output of the multivibrator. I
hope this doesn't change the frequencies enough to affect
operation; the pull-up resistor off "Point 3" now has a little
extra current flowing through it during the "off" cycle.

The Receiver

The receiver is not crystal-controlled. It is super-
regenerative, and is tuned by a slug-tuned coil (the only coil on
its board).

You will notice that its board is not strictly rectangular.
With the component side up, put the long straight edge toward
you. The tuning coil will now be in the upper-left corner. The
upper edge will taper toward you slightly, near the corners.
Also, in the center of the upper edge is a notch--about 3/8 of an
inch deep and 1/8 of an inch wide.

Further examination will reveal that there are several tiny
niches up there as well; these can be used as alternative
connection points, just in case you don't like soldering wires
into holes (with two exceptions, both a hole and a niche gain
access to each trace). I suspect that these niches were intended
for push-in terminals of some sort. (You may need to find the
tiny niches in among many "scallops"; in fabrication, the boards
were perforated and then broken off, thus leaving serrated edges
on one side or another.)

A hole for the antenna is provided in the upper-left corner.
Actually, the pad under this corner reveals that there are
several places the antenna could go--a niche to the right of the
corner hole, and a hole farther down the left edge, to the coil's
left. A 12-inch length of wire is recommended.

There is no conventional hole for a ground lead; they tell you
to take the emitter of TR17 as ground. As luck would have it,
there is a niche adjacent to TR17; this niche, which is
surrounded by a ground pad, is on the right end of the board,
about 1/4 inch down from the upper-right corner. Ground is also
available at the mounting holes, which are in the corners nearest
you. (Look out! I was elated to find a solder hole near the
left-hand mounting hole, only to find that this went to some
obscure place in the circuit and is never used.)

Along the upper edge of the board, just before the large notch,
is the "Plus 9V"
point. Either the tiny niche or the hole behind it can be used.

Directly to the right of the large notch is one of the main
drive motor connections, M2B. This is just a hole, without a
niche. A quarter inch beyond M2B--5/8 of an inch to the right of
the large notch--is M2A, the other main drive connection. M2A
has a niche and a hole behind it.

Connections to the steering motor occur along the slanted cut,
just before the upper-right corner. This upper-right slanted
section has two niches, widely spaced; their companion holes are
not directly behind them, but between them--they are closer
together and positioned near the edge of the board. The left-
most connection is M1B, with the one nearest the upper-right
corner being M1A.

This whole upper-right section of the board is dominated by a
block of 12 transistors--three deep and four across. The "Plus
3V" point is near the upper edge, exactly centered with respect
to the transistor group; i.e., a niche and a hole slightly to its
right reside between M2A and M1B, just before the gentle bend.

Their Hookup

The "Plus 9V" point goes directly to the positive
of the 9-volt battery. The "Plus 3V" point goes to the positive
of its two flashlight cells. A DPST on-off switch is used. The
swinger of one section goes to the negative of the 9-volt
battery; the swinger of the other section goes to the negative of
the 3-volt battery. The two "on" contacts are tied together and
go to ground.

The main drive motor (a 3V 300mA unit) is shunted by 0.047uF or
greater; it is then connected directly between M2A and M2B. The
steering motor (a 3V 200mA unit) is shunted by 0.047uF or
greater; it is then put in series with an optional 10-ohm, 2-watt
resistor, this combination going between points M1A and M1B.

With any noise in the receiver (transmitter off, left turn, or
right turn), M2A will be positive with respect to M2B. With the
500Hz left-turn signal, M1A will be positive with respect to M1B.
With either straight receiver noise or a strong dead carrier, the
M1 steering outputs are floating; they are not committed to anything when a steering signal isn't being received. With light
loads on these outputs, I measured 2.4 volts when they were

The motors are directly driven by the upper eight of those
twelve transistors. On each motor leg is a complementary pair of
transistors; pulling one leg up and the other down--or vice
versa--allows the motor to be powered in both polarities.
Unfortunately, these transistors have their emitters going to the
3-volt "rails," and the collectors of each pair are connected
together to drive a leg of the motor. Thus connected, they have
a nasty habit of all turning on at once.

I first made the mistake of connecting the 3-volt battery
without its accompanying 9-volt unit (the latter being innocently
dead). When I smelled something, I grasped the board, only to
find that the drivers were hotter than the hinges of Hades--the
pairs tried to turn on simultaneously. This happened even though
there were no motors connected. (The next time Little Junior
tearfully runs up to you with a toy car that is billowing clouds
of smoke, you can console him with the statement that, "If I had
caught it in time, I could have saved this favorite of yours with
a new battery a while ago.") Keep that receiver battery going.

The next thing I noticed was that applying more voltage (I
would prefer 5V instead of 3V) was not taken kindly either. The
current goes way up when this is done. For example, without a
motor, it can draw 85mA all by itself using a 3-volt battery;
going up to 4.5-volts, it reached 350mA.

With all of this, it still remains a practical device. All you
need is to find some low-voltage relays (reed relays, probably).
By placing diodes in series with two of them
--one facing one way and the second facing the other--a pair of
relays could be controlled from each set of output terminals.
Control signals can be referenced to ground as well.


I don't want to run motors, necessarily, with
this set. I want to control logic and steer electrons with this
thing. Furthermore, why let the receiver make me use two battery
, and warm up my projects with its own dissipation problems. As it happens, there are modifications that cure all ills.
The "luck of the Irish" was with us when they did the production
engineering on this receiver; these modifications turn out to be
relatively easy.

The block of twelve transistors involved in driving the motors
can simply be dispensed with. As mentioned, eight of these are
directly coupled; these are mounted near the output connections.
The four behind these are involved in a zany driving scheme
(there are some low-value resistors that will disappear also).
The circuits you are left with are as follows:

An NPN common-emitter transistor is being driven by a voltage
doubler that rectifies the receiver noise. In its collector is a
5.6K resistor which is being supplied from, with the luck of the Irish, a 5.1V zener--just the sort of thing to drive logic.
High- and low-pass filters for detection of the tones feed
voltage doublers that rectify and drive a pair of common-emitter
NPN transistors as well. Their outputs each go through 33K
resistors to the "Plus 3V" point--which can now be run from 5V.
Since there is no reason to separate the supplies any more, a 5-
volt regulator and the receiver can all be supplied by a single
9-volt battery.

As a final touch of good fortune, there are exposed resistor
leads on top of the board for easy access to all these points.
In fact, there is even a convenient point from which the audio
can be retrieved; if you have a strong stomach, you can listen to
an evening's whoop-dee-doo of our fine upstanding Citizenry.
(This audio output is recommended by the manufacturer for tuning
up the transmitter and receiver using a VOM; this simply will not
work, but I can tell you how.)

[Tuning of the receiver is easy--by ear. Throw the transmitter
to "reverse" (a dead carrier) and tune the only coil on the board
to the center of the signal, using an amplifier on the point
described above to listen for minimum noise. The transmitter has
two coils you can adjust. I suppose you could find the quench
voltage in the receiver so as to get a sort of "S-meter." I
didn't use this to tune my transmitter; I have an audible field
strength meter (as was described in "Auditory Gimmicks--Old and
New," SKTF, Spring 1981). (There is an error in that circuit;
there is supposed to be an RF choke from antenna to ground.
Otherwise, you can't beat it.)]

Ravaging the Board

I used "flush-cutting" diagonal cutters to
nip off the offending components. I will describe them as they
go; pay attention to what you're doing, since components you want
are right nearby. For the sake of consistency, I will preserve
the orientation of the board that we are used to--the longest
edge nearest you.

These transistors are of the type whose packages are plastic,
having a "half-round" shape. (With the leads down and the flat
side toward you, the three leads are, from left to right:
emitter, base, collector.) Although their leads are "in-line,"
they have been bent so as to afford mounting in the board with
the familiar triangular configuration (the base being brought
toward the flat side).

The twelve we want to get rid of are in a rather uniform matrix
in the upper-right section of the board. Immediately to the left
of this matrix is TR6, the forward/
reverse output unit we want to save, so be careful. The top two
rows of four have no coupling resistors in among them. Nip these
off with dispatch.

In the next line of defense, there are some resistors. These
all go as well. One is off to the right--a 22-ohm unit. Moving
to the left, leaning up against the transistors that remain, is
another 22-ohm unit, then two 10-ohm ones. Nipping off the first
three can be done with glee. The left-most one is dangerously
close to TR6; I cut its farthest leg, and then bent its body back
and forth until it broke free, rather than cutting it free.

Now we have four more transistors to remove. TR6 is slightly
out of line with these, which will help you avoid it.

TR5 is a transistor in the lower-right corner; you want this,
since it is the high-beep detector. However, immediately ahead
of it (toward the oddly shaped edge of the board) is a 10-ohm
resistor. This can go too, but just behind it and slightly to
its left is TR5's 33K pull-up resistor, which you want.
Strangely enough, the 33K pull-up unit is not in direct line with
TR5--you might guess it belonged to TR4. (I began to make heavy
use of my continuity tester to tell the difference between high-
value and low-value resistors. If you cut the exposed lead free
first, testing between this and the bottom end will assuredly
tell you which you got, by the pitch.)

TR5's 33K resistor is mounted with its exposed lead just to the
right of its body. Immediately to its left are two resistors
which are turned 90 degrees from this, their bodies are most
exposed, and their upper leads seem to disappear in the forest.
The one next to the first pull-up you saved is 10-ohms; the 33K
unit for TR4 is similarly mounted and next door. In order to
avoid cutting the wrong thing, I grabbed the top of the 10-ohm
unit with locking forceps and took my time bending and breaking
it free.

The next items, a diode in front of TR4, and a 3.3K resistor to
its left, are unwanted. However, they don't cause any harm or
crowd your future efforts. In fact, the exposed lead of the 3.3K
unit goes to the collector of TR6 (via a round-about trace), and
this will be your tie point. (This 3.3K unit is right in front
of an electrolytic capacitor; its exposed lead is to the right of
its body--oddly enough, away from TR6.)

Making Connections

Although slightly tricky, wires are
"merely?" soldered to the leads which emerge from the tops of the
resistors. Using stranded wire with nicely tinned ends, I make a
hook in the wire, loop this under the resistor's lead, do what I
can to wrap a turn around the lead, and solder it there. This is
done in four places:

[When circuit boards began to get crowded enough to necessitate
standing components up on end, the Japanese had the ingenious
idea of coating axial leads of components to reduce accidental
short circuiting atop the board. This coating seems to be
attacked by the flux (as it should be), making soldering
possible; however, you may have to try a couple of times before
wetting occurs (it's not like soldering bare copper). Furthermore, in using such component leads as test points (before you
attack them with solder), you will only get good connection by
attaching your clips near the board--where heat during
installation has flawed this coating.]

The easiest to reach is TR5's 33K resistor, which appears
slightly to the left of TR5. To the left of this is the top of
TR4's resistor. Then, skipping the diode, connect a lead to the
top of the 3.3K resistor in front of the electrolytic capacitor,
at the bottom-left corner of the clearing. The "Plus 3V" point
is where it always was, just before the bend in the edge of the
board; this can be tied to the "Plus 9V" point if you wish. The
old M2 and M1 points have no connection.

The receiver's audio output can be found at the top of another
resistor. As you run your thumb along the bottom edge of the
board, you will notice that, right in the center, a place exists
where components are not as close to the edge as everywhere else.
On the left of this "cove?" is an electrolytic capacitor; on the
right is a diode standing on end. Right between these, at the
back of the cove, is a resistor whose exposed lead is to the left
of its body. A high-impedance amplifier can be connected between
here and ground.

Output Circuits of What's Left

The emitters of TR5 (nearest
the right-hand end of the board) and TR4 (to the left of TR5) are
grounded. Their collectors each go through 33K to VCC (5 to 9
volts, as applied to the old "Plus 3V" point). The emitter of
TR6 is grounded as well. Its collector goes through 5.6K to the
cathode of a 5.1V Zener diode, the anode of which is grounded.
To suppress the tremendous amount of receiver noise that is
unavoidable, this collector is bypassed to ground by 33uF. This
collector also goes through 3.3K to nowhere; and to the cathode
of a diode, the anode of which goes nowhere.

Logic Considerations

CMOS buffers are probably advisable. Not only am I not sure
that the "outputs" can sink enough current to drive TTL, but CMOS
logic has the inherent ability to ignore noise. These circuits
are driven by noise; they're bound to be noisy.

Five-Volt Logic Setup

(Do your own logic as per the "Truth
Table" below; this sets up the proper power arrangements.) The
negative of the 9-volt battery is grounded. Its positive goes
through an on-off switch to the "Plus 9V" point--just to the left
of the big notch. This point also goes to the input of a 7805
regulator, the common terminal of which is grounded. This input
is bypassed by the parallel combination of 100uF (10V electrolytic, with negative at ground) and 0.1uF (disc ceramic). The
5-volt output of the 7805 is bypassed by a similar parallel combination of 100uF and 0.1uF as well; this output goes to the
former "Plus 3V" point, as well as to the VCC line of any logic.

Truth Table

All outputs pull down when their noise is
presented to them. The collector of TR5 (the right-most
transistor) pulls down for the 3000Hz tone, and is high for
everything else. The collector of TR4 pulls down for its 500Hz
tone, and is otherwise high. The collector of TR6 (on the left
side of the "clearing") only goes high for a dead carrier; it
goes low for any other noise (specifically, transmitter off and
steering tones).

Two-Function System with Pseudo Carrier Detection

This setup is to be used if the transmitter is always "on".
This arrangement has two advantages: The dead carrier keeps
other zealots of radio from directly controlling your system with
voices and whistles. The logic described can trigger a signal
that indicates that the receiver is out of range of the

The two tone outputs are combined in a "NAND" gate; the
presence of either tone causes this gate's output to go high.
This output is combined in an "OR" gate with the noise-detector's
output (TR6); either silence or a tone will bring an input of
this gate high. If: noise causes TR6 to pull low, AND, the NAND
output remains low, saying "This noise isn't a tone," the OR gate
goes low to trigger an alarm. (By de Morgan's Theorem, an OR
gate does AND all low inputs--this one AND that one have to be
low before its state changes.)

Now that we have our logic straight, we have to use de Morgan's
Theorem to combine these functions in one gate chip. (I refer to
Statement 1 of de Morgan, "Gabbing About Gates," SKTF, Summer
1981.) Ol' Morg said that a NAND gate is the same as an OR gate
being given negative inputs. A quad NAND gate gives us just
enough logic elements;
two of the gates will have their inputs tied together, and will
be used as inverters.

Carrier Drop-Out Alarm Circuit

A CD4011 quad NAND chip is
used. Pin 7 goes to ground; pin 14 goes to plus 5V. Pins 8 and
9 are tied together to make an inverter; pin 10 is its output.
Likewise, pins 12 and 13 are tied together to make an inverter,
pin 11 is its output.

Pin 6 goes to the collector of TR5, the high-tone detector.
Pin 5 goes to the collector of TR4, the low-tone detector.
The resulting output, pin 4, is inverted by running it to pins 8
and 9. The collector of TR6 is inverted by running it to pins 12
and 13. The inverter outputs, pins 11 and 10, go to pins 1 and
2, respectively. The output of the system, pin 3, triggers an
alarm when it goes low.


By Bob Gunderson


This simple alarm grew out of a need to protect our motor
home--a 21-foot LeisureCraft. This is simply a house on a 1-ton
truck, and it is ideal for camping away from home. It is indeed
a delight to be able to visit your friends and to bring your bed
and board along with you. Then, too, there is nothing quite like
saying goodnight to your host, and spending a quiet hour with
your best gal.

However, even more so than "back at the ranch," a motor home is
vulnerable to burglary and vandalism, and a reliable alarm system
is necessary to protect your "home away from home."

Many alarms are subject to RF interference; they trigger
indiscriminately, frightening you to death and tempting you to
disable them after a few false alarms. This one, by virtue of
its simplicity and freedom from sensitive electronics, is quite
immune to RF. (No alarm is foolproof, but it does keep the
untrained thief from breaking into your home. As the old saying
goes: "Locks are made only for honest people." The professional
crook can find his way into any abode.)

The heart of the system is based on two relays, one of which
keys an alarm sound; the other acts as a time-delay which
prevents the first from closing unless a protective circuit is
broken for more than a few seconds. The only problem is finding
high-resistance relays which will close on 12.6 volts; the ones I
used were World War II surplus items. Their coil resistance is
10,000 ohms, and this means that they will close at a bit more
than 1 milliampere. It is quite likely that 5,000-ohm relays
will work just as well (or better, since the closing current is
twice that of the 10,000-ohm units). It would be well for you
to check the relays on 12 volts before you attempt this project.
Before going further, the best idea is simply to describe the
circuit; its operation will become obvious to you.

Normally open switches (sensors) are made to operate from the
doors and windows. (More on the actual types used will be

presented later.) These switches, which are held closed by the
doors and windows, are all in series; opening one of them will
"break the loop" and trigger the alarm.

The alarm set switch consists of two parallel-connected
switches--a key switch hidden somewhere on the outside of the
motor home, and a toggle switch on the alarm panel inside. This
allows you to set the alarm from inside or outside. One side of
this parallel combination goes through a 5-ampere fuse to the
positive of the 12-volt battery which operates the vehicle. The
other end of this parallel connection goes through the door and
window switches, with the other side of this series loop going to
the top of the coil of relay RY1 (high-resistance relay). The
other side of RY1's coil goes to the battery negative and to
common ground. The junction of the on-off (parallel) switches
and the series-connected door and window switches will hereafter
be referred to as Point X. (With the alarm set and with all the
sensor switches closed, RY1 is pulled in.)

The arm of RY1 goes to Point X, and the normally closed contact
of this relay goes through a 2.2K ohm resistor (1 watt) to one
end of the coil of relay RY2 (high resistance); the other side of
the RY2 coil goes to the battery negative. The RY2 coil is
shunted by a large electrolytic capacitor (2,000 to 4,000uF) with
the negative side of this capacitor connected to the negative
side of the coil. The normally closed contact of RY2 is not

The alarm is provided with two signals--one being a buzzer
mounted on its panel, and the other a transistorized horn or
warbler (siren) which is used when you are away from the vehicle.
These signals are selected by switch S2 (DPDT toggle), also
mounted on the instrument panel. With this switch in the
"buzzer" position, the loop which protects the windows is short-
circuited, so that the windows may be left open during the night.
However, if the door is opened, the buzzer sounds.

The arm of RY2 goes to the swinger of switch S2A, with position
1 going to the hot side of the buzzer, and position 2 going to
the horn. The other ends of buzzer and horn go to battery
negative. The normally open contact of RY2 goes to Point X. The
other half of switch S2 (S2B) short-circuits the window loop when
this switch is in the "buzzer" position.

Magnetic switches or simple microswitches may be used for the
two loops. To use the magnetic switches, a small bar magnet is
mounted to the door or window, adjacent to the switch which is
fastened to the fixed door or window frame. These switches may
be purchased as SPST (normally closed or normally open), or
single-pole, double-throw types are also available. The presence
of the magnet operates a small reed inside the switch. I
originally used magnetic units, but found that somehow they went
bad and actually burned out. I have since replaced all switches
with the dependable microswitches. (The magnetic switches you

get might be huskier than the ones I had.) The microswitches I
used have a lever which closes the contacts with door and windows
closed; they open when the lever is released.


When the alarm is set, relay No. 1 closes. This
opens the circuit through the coil of RY2 before it has a chance
to close. This retarded action of RY2 is provided by the
capacitor and resistor associated with the RY2 coil, since the
uncharged capacitor is virtually a short circuit, and takes
finite time to charge. Thus, RY2 remains de-energized, and RY1
remains closed. If the "set" circuit is opened, RY1 simply falls
back, and nothing happens. However, if any portion of the door
or window loops open, RY1 drops out, and the normally closed RY1
contacts complete the circuit through the resistor and the RY2
coil, and RY2 closes. This will energize the signalling device,
and it will stay energized as long as the circuit through the
loop is open. However, if, upon noting the signal, the thief
slams the door, the alarm will continue to sound until the
voltage across the capacitor has fallen to a value sufficiently
low to open relay RY2, whereupon the whole system is reset.
Should it happen that the thief thinks the system has stopped,
opening the door or window recycles the system. I designed it
with the idea that I didn't wish to catch the thief, and perhaps
become a dead hero; rather, I would prefer to scare him off.

I have provided an additional feature--a means for energizing a
small radio transmitter using the noise produced by the horn.
This portion of the system is a simple two-stage audio amplifier,
with a microphone at its input, and a silicon-controlled
rectifier connected as a switch to turn on the "paging system,"
and to transmit the sound of the alarm to a receiver I carry when
we're away from the camper. The rectifier is in its "off"
position until the gate signal reaches the firing point, at which
time the SCR closes and turns on the transmitter. What's more,
the transmitter will remain on even after the alarm has reset.
The only way it can be turned off is to open a switch in series
with the SCR.


by Jean LeBorgne

Although the Japanese Solder Guide (see "The JA3TBW Solder
Guide," SKTF, Spring 1983) is a great idea, I think this
modification constitutes an improvement. Before, I always had
trouble feeding the solder; it would buckle as it entered the
tube, and I would even lose control of the guide while attempting
to feed solder.

I designed a feeder sleeve which holds the solder above the top
end of the tube, and which permits a controlled amount of solder
to be fed. It is made of a piece of aluminum bar stock with a
hole drilled through its length to accommodate the stainless-
steel solder tube. At its upper end, a hole is drilled and
tapped for a setscrew which can be tightened down against the
solder. With this setscrew, the feeder sleeve can be set to feed
a predetermined length of solder, since the tube and the feeder
sleeve will only telescope as far as the shaft collar permits
(the shaft collar being part of the original Japanese Guide).
(This system is similar to a "metered" syringe, which has stops
on the plunger to regulate dosage.)

In operation, the solder is passed first through the feeder
sleeve, then the solder guide. The solder is positioned so that

it is even with the bottom end of the tube. Next, the feeder
tube is slid down onto the guide to a point just above the shaft
collar (perhaps leaving a space of 1/4 or 3/8 inch). The
setscrew at the upper end of the sleeve is then tightened against
the solder so that it can be fed by pressing down on the sleeve.
Put the bottom end of the solder tube against the connection and
heat the materials with the iron, using the tube to guide your
iron down onto the connection. When the solder melts, the sleeve
will telescope down to the shaft collar.

To prevent the flux from "gluing" the solder in the tube,
either withdraw the tube from the feeder, or loosen the setscrew
in the feeder and feed solder forward through the
system--immediately after removing the tool from the connection.

My feeder sleeve is made from 1/2-inch aluminum bar stock, and
is cut to about 1-1/2 inches in length. Then, a hole is bored
along its length which is 0.080 in diameter (using a No. 46
drill). Near the top of the guide, a hole is drilled and tapped
to fit a No. 3-48 or 4-40 machine screw--this is the setscrew
which holds the solder. I did not have a proper thumb screw,
which would be preferred, so I simply threaded a nut along a 1/2-
inch screw and tightened this against the head, thus giving me a
handle for my fingers.

[Editor's Comments: Very nice! Thank you, Mr. LeBorgne.
Playing around with different materials for the feeder might be
fun. For example, I'd like to try 3/8-inch Teflon bar
stock--something nice and "soapy-feeling" (self-lubricating).
Delrin plastic would work, but you might contaminate your iron on
it once in a while. I wish I could pass these on to you, dear
readers, but the machining costs are more than we could bear.
Nevertheless, we can pass along more of the original solder
guides for those who need them.]