TABLE OF CONTENTS

The Intersil ICL 8038 Function-Generator Chip

A Highly Portable Function-Generator using the ICL 8038

The Smith-Kettlewell Universal Tone Indexer

The Smith-Kettlewell Speaker Phase Tester

The Silkiest MCW Signal on the Air

Save Your Batteries with a Timer Switch

Summary of SKTF Survey

Editor's Crystal Ball

THE INTERSIL ICL 8038 FUNCTION-GENERATOR CHIP

Abstract

The INTERSIL ICL 8038 is capable
of producing sinusoidal, square, triangular,
saw tooth and pulse waveforms over an extremely wide range of frequencies. In addition, the frequency can be voltage-controlled

so as to be made easily tunable over a range
of 10-to-1 or more. Though its outputs are
not low impedance, a full-fledged function
generator can be built with the inclusion of
a buffer amplifier. Other applications
include use as a highly linear VCO and as
special test oscillators; we have used this
IC in our tape indexer and various Morse Code
devices. Sample circuits of a general sort
will be given here, while specific hookups
appear in articles to follow.

Brief (?) Description

The ICL 8038 is a 14-pin package, two pins
of which are not used (13 and 14). It has
three output pins, one for sine, one for
triangle, and one for squarewave (the square-
wave needs an external pull-up resistor). It
has a high-impedance voltage-control pin, in
addition to which it provides a VREF on still
another pin so that you can anchor the VCO
pin to VREF if external electrical control of
the frequency is not desired.

The oscillator works by charging and
discharging an external timing capacitor.
The charge and discharge times are each
controlled separately by resistors; these can
be made equal or vastly different to generate
a wide variety of wave shapes. A near
approximation of symmetry between the two
half-cycles can be had with these resistors
combined into one, resulting in a circuit of
extreme simplicity.

Fine "shaping" of the sinewave can be
accomplished by attaching trim pots to two
pins provided for this purpose. Compromise-
trimming with a single fixed resistor is also
commonly done.

Besides taking care of the VCO input (for
example, jumpering it to its VREF pin), a
usable test oscillator can be made with a
single resistor and capacitor. Beyond this,
the versatility of this chip affords connection in many configurations; this, not its
complexity, is what makes this article so
frighteningly long. Please read on in good
spirits.

Theory of Operation

[Ah, something simple for a change.]

The heart of the 8038 is a triangle-wave
oscillator based on the charge and discharge
of a capacitor. The workings of this oscillator are best illustrated by looking at an
equivalent circuit:

Equivalent Circuit of the Triangle
Generator

Two "active current sources" act
on a timing capacitor, the bottom end of
which goes to the -V line. The top end of
the capacitor goes to the output of a current
source "I" coming off the plus supply. The
top of the capacitor also goes through a
single-pole single-throw switch to a current
sink of -2I, which comes off the -V line.
With the switch open, the capacitor is
charged by current "I," while closing the
switch will discharge the capacitor with the
net current of -I (the sum of the source and
the sink).

One input each of two comparators senses
the voltage at the top of the capacitor; one
comparator is set to trigger at 2/3V, while
the other is set to trigger at 1/3V. The
outputs of the two comparators operate a
flip-flop, one output of which is used to
open and close the SPST switch (turning the
current sink on and off).

The capacitor is charged by the current
source until a voltage of 2/3V is reached, at
which point the upper comparator directs the
flip-flop to close the switch and establish
the discharge current. The capacitor is then
bled off by a net current of -I until the
lower comparator directs the flip-flop to
disengage the current sink, whereupon the
charge cycle resumes.

Squarewave Generation

The free output of
the above flip-flop drives the base of an NPN
transistor in the common-emitter connection.
The collector of this transistor is open (a
pull-up resistor must be connected externally
if a squarewave output is desired).

Sinewave Generation

The sinusoidal output
is gotten by performing a "wave-shaping
operation" on the triangle. Progressive
loading is imposed on the triangle wave as it
approaches its peaks--"rounding it off" to
get a simulated sinewave. (Under the best
conditions where trimming of the circuit has
been done externally, a total harmonic distortion of better than 1/2% can be expected.)

A string of eight switching transistors
connects load resistors of various values
across a high-impedance version of the
triangle--four progressive loads being
imposed on each half of the cycle. The
points at which these various loads are
applied are determined by the voltages found
along a string of nine resistors in a voltage
divider contained in the chip. The top and
bottom junctions on this divider string are
made available on pins 1 and 12, respectively, so that further "trimming" can be
done to minimize distortion.

Wave Symmetry and Frequency Adjustments

The actual value of the charge and discharge
currents can be set individually with external resistors. Therefore, a fast charge and
slow discharge, or vice-versa, can be purposely set up; in this way, the triangle can
be made into a sawtooth and the squarewave
can be made into pulses. The variations
afforded can be extreme--the duty cycle can
be varied from less than 1% to greater than
99%.

Frequency determination depends on three
elements listed below:

1. Different capacitors can be connected; a wide-range function generator might
   have several such units to be selected by a
   "range switch."
2. Since the charge and discharge
   currents are "programmable" using the aforementioned resistors, changing these will
   directly affect the frequency.
3. The charge and discharge currents
   are also voltage-controlled via a VCO input
   pin. Furthermore, once the duty cycle is
   chosen, the charge and discharge times
   "track" very well over a range of greater
   than 10-to-1 as the voltage on this pin is
   varied. Finally, a voltage divider within
   the chip provides a reference (VREF on pin 7)
   to which the VCO control pin can be tied when
   external voltage control of the oscillator is
   not intended.

Specifications

[Editor's note--Much confusion is apparent
in looking through the literature at specs
which refer to the quality of waveforms,
distortion and the like. This is probably
due to the fact that several grades of this
chip are available, and that no consistent
mention as to which chip is being demonstrated in the spec is apparent. The most
commonly available unit (which one is likely
to find at Radio Shack and JAMECO, for example) is the unit of commercial grade having
the suffix CC; i.e., ICL 8038CC.]

The ICL 8038 can be used over a range of
frequencies from 0.01Hz to 1/2MHz. (The
distortion of the sine output begins to
increase above 10kHz; some circuit configurations only guarantee a distortion
figure of 10% as 1/2MHz is approached.) The
frequency will be relatively independent of
the supply voltage if the VCO input is tied
to its VREF pin (both the reference voltage
and the triggering points will track so as to
hold the frequency constant). If the VCO
input voltage is controlled independently
from an external circuit, the supply voltage
will directly affect the frequency, thus
making a regulated supply necessary.

Although advertisements boast that the VCO
will operate over a range of 1000-to-1, this
range is hard to achieve in practical terms;
the control pin must be pulled slightly above
VCC to cover this range, and mistracking of
the current source and sink lead to asymmetry
in the waveforms. It seems to the Editor
that a different chip, such as the VCO
section in the CD4046 PLL IC (See SKTF,
Summer 1982) would be a better choice where a
wide range of frequencies as a function of
voltage is required. By my experiments, I
have found that a more realistic statement of
VCO frequency range is about 16-to-1 or less.

Supply Voltage

The range is stated to
be +5 to +18 volts. Actually, the chip runs
off a single supply of from 10 to 36 volts
(drawing a minimum of 12mA, typically 20mA);
however, since the output waveforms are
"centered" within this region, a split supply
can be used, the common connection of which
can then be considered a "signal ground." In
other words, the waves should be symmetrical
in amplitude about 1/2VCC. On the other
hand, if capacitor coupling is used on the
outputs, the split supply can often be
avoided.

The range of the VCO will to some extent
depend on the supply voltage used. Mistracking of the current source and sink
occurs as the VCO pin approaches VCC; it is
caused by the junction potentials in the
active devices becoming significant. At
higher supply voltages, the region in which
these problems can be expected becomes less
significant as compared to the acceptable
range of input voltages on the VCO. It is
probable for this reason that they list the
lowest supply voltage as 10V. [Out of
ignorance, I have run these devices off 9
volts in countless projects, and I have never
had a chip misbehave. In these projects,
however, I have usually affixed the VCO input
to the VREF pin, thereby avoiding the
troublesome region.]

Output Amplitudes

From the section "Theory
of Operation," it can be seen that the output
amplitudes vary in direct proportion to the
supply voltage. The peak-to-peak values are
listed below:

• Triangle--30% of VCC min., 33% typ.
• Sine--20% of VCC min., 22% typ.
• Square--the entire range of VCC, with
  the exception of the output transistor's saturation voltage; 0.2V typ.,
  0.5V max. (sinking a current of 2mA).
  When this transistor is open, the
  leakage current is 1uA max.
• Squarewave rise and fall time--100
  nanoseconds rise, 40 nanoseconds fall
  (with a pull-up resistor of 4.7K and
  a supply of 20V).

Frequency Formulas and/or the timing of
each half-cycle:

• f (using separate and equal R_a and
  R_b) = 0.3/RC (R is either R_a or R_b;
  C is the timing capacitor)
• f (using single resistor) = 0.15/RC
• T₁ (positive half-cycle) = (5/3)R_aC
  (R_a goes from pin 4 to VCC; C is
  the timing capacitor from pin 10
  to -V.)
• T₂ (negative half-cycle) = (5/3)
  [R_aR_bC/2R_a-R_b] (R_b goes from pin 5 to
  VCC.)

Recommended Loading

The literature says
very little about "output impedance." In
most examples, however, the sine and triangle
outputs are loaded with 100K. By experimenting I have found that load resistances of
less than 100K degrade performance; the output of the triangle begins to fall off and
clearly audible distortion of the sinewave is
noticeable. Perhaps a load resistance of
500K is optimal. Specifications for the
squarewave's pull-up resistor are not discussed directly in the literature, but the
rise- and fall-time specs are given for a
collector current of 4mA. This would dictate
using a 4.7K pull-up resistor for a supply
voltage of 20V.
VCO Input and the VREF Pin--The VCO input
is very high impedance; though never specified, the internal circuit of the chip shows
this pin as being the input of an emitter
follower (more or less, the internal circuit
being very complicated). My experience is
that this input impedance is very high, well
over a megOhm.

The lowest output frequency is gotten with
this pin being taken to, or slightly above,
VCC (clamped with a diode to limit this
voltage at 0.6V above VCC). The frequency
increases and varies directly as the control
voltage drops (measured with respect to VCC
and not with respect to ground). In other
words, the frequency goes up as this control
pin is brought down from VCC. The lowest
voltage which should be allowed to appear on
this pin (resulting in the highest frequency
for any given hookup) is 2/3 of VCC plus
about 2 volts. For example, with the supply
of 15V, the VCO pin will function over a
range of VCC down to 2/3 of 15, +2--12V
above the minus supply (a range of 3V below
VCC).

The VREF pin (to which the VCO can be
directly tied) consists of a voltage divider;
VCC goes through 10K, then through 40K to -V,
with the junction of these resistors being
VREF, available on pin 7. Even with the
control pin tied to pin 7, a modulation
signal for FM'ing the chip can be applied to
the control pin through a coupling capacitor;
the resultant "input impedance" of this

arrangement is the parallel combination of
the 10K and 40K resistors, about 8K. This
impedance can be increased by running the
control pin through an external resistor to
VREF; the resultant impedance at the control
pin will be 8K plus R, where R is the value
of the external resistor.

Using the ICL 8038
Illustrated by Exemplary Circuits

Simple 1kHz Test Oscillator

The supply
pins are pins 11 and 6; pin 11 goes to -V,
while pin 6 goes to +V (VCC). Pin 10 goes
through a 0.015uF timing capacitor to pin 11
and to -V. Pins 4 and 5 (the current source
and sink programming pins) are tied together
and go through 10K to pin 6 and to VCC.
(This being a single resistor common to both
halves of the cycle, the formula for determining frequency is: f = 0.15/RC.) Pin 8,
the VCO input, goes to pin 7 (VREF).

Pin 2 is a pleasant sounding sinewave (its
distortion has not been minimized), while pin
3 is a triangle wave. In order to get a
squarewave from pin 9, its output, a pull-up
resistor from pin 9 to VCC is necessary; let
us run pin 9 through 10K to pin 6 and to VCC.
Now that we have connected up the squarewave,
however, switching spikes from this section
will no doubt be present on the power supply
and affect the performance of the rest of the
chip. It is recommended, therefore, that the
VCO pin be bypassed to VCC through 0.1uF.

Pins 2 and 3 (the sine and triangle outputs) can each go through 100K, then through
0.1uF to their respective hot output terminals. The squarewave is less susceptible to
loading (in our example it will have a 10K
source resistance in the positive direction)
and can directly go through a coupling capacitor to its output jack. Since the outputs
are capacitively coupled, the cold side of
the output can be taken as -V. If a split
supply were used (perhaps made up of two 9V
batteries in series), the common connection
could be taken as ground; the coupling
capacitors could be eliminated, since the
waveforms are nearly symmetrical with respect
to one-half of the supply voltage.

A noticeable improvement in harmonic distortion of the sinewave output can be gotten
by running pin 12 through 82K to -V; this
constitutes a very rough adjustment of the

sinewave shaping network. In later circuits
where care is taken to equalize the charge
and discharge time (thus tuning for temporal
symmetry), more complicated external networks
for fine trimming of the sinewave circuit
will be given. The next step up from the
above would be to change the 82K resistor to
a 100K rheostat.

Variable Frequency Test Oscillator

We can
make the above circuit adjustable in frequency so as to cover a range of perhaps
100Hz to 1kHz. There are two ways to achieve
this:

1. We can make the current source
   resistor off pins 4 and 5 variable. (The
   literature specifies practical limits on
   resistors associated with pins 4 and 5; they
   should not be less than 500 Ohms or greater
   than 1 megOhm.) In our example, pins 4 and
   5, which are tied together, will go through
   10K, then through a 100K rheostat to VCC.
   The disadvantage of this system is that the
   frequency does not vary linearly as the
   rheostat is adjusted; this adjustment becomes
   very critical at the high-frequency end.
2. Another way of doing it is to use
   the VCO control pin. Instead of going to
   VREF, pin 8 (which is still bypassed to VCC
   by 0.1uF) goes to the arm of a pot. The
   bottom of this pot goes to VCC, while its top
   end goes to VREF, pin 7. To get the maximum
   range from this system, the resistance of the
   pot should be high enough so as not to pull
   VREF up to VCC; I usually use a linear pot of
   about 250K. A slight loading effect (and
   hence lowering of the frequency) can be offset by decreasing the value of the resistor
   on pins 4 and 5, perhaps to 9.1K in this
   example. This system will give you a good
   linear frequency range of greater than 10-
   to-1. Failure to go to 0 frequency, as well
   as asymmetry in the output waveforms, are
   unavoidable effects which dictate that the
   very bottom end of this adjustment be
   ignored.

Minimizing Distortion

There are two items on this agenda. First,
we must equalize the charge and discharge
currents in the timing capacitor so that the
upper and lower halves of the cycle are of
equal duration. Second, we can adjust the
"sine converter" for optimum wave shaping.
Both procedures require embellishments on the
basic circuit. The improved circuits and
adjustment procedures are listed below:

Temporal Symmetry

With pins 4 and 5
jumpered together, the current source and
sink share the same sampling resistor; there
is no way of "balancing" them or adjusting
them individually. We could connect each pin
through a separate rheostat; pin 4 (dedicated
to the "charge" half of the cycle) would go
through a rheostat to VCC, while pin 5 (dedicated to the "discharge" half of the cycle)
would go through another identical rheostat
to VCC. A system more commonly used, however, is to run pins 4 and 5 each through a
resistor, with the far ends of these resistors going to the ends of a single pot which
is shared between them. The arm of this pot
goes to VCC. Let us rebuild our 1kHz test
oscillator to incorporate this feature:

Model 2, 1kHz Test Oscillator with
Temporal Symmetry Adjustment

As before, pin
6 goes to VCC, while pin 11 goes to -V. Pins
7 and 8 are tied together and go through a
0.1uF bypass capacitor to VCC. Pin 10 goes
through a 0.015uF timing capacitor to -V.
Pin 4 goes through 18K to the top of a 2K
pot; pin 5 goes through 18K to the bottom of
this 2K pot. The arm of the pot goes to VCC.

[The resistance values off pins 4 and 5 and
the timing capacitor were this time chosen by
using the formula:
f = 0.3/RC
(where R is either R_a or R_b).

In this case, R_a is the resistor from pin 4
through its portion of the pot to its wiper.
R_b is the resistor from pin 5 plus its
portion of the pot to its wiper. Using the
above formula, the frequency would come out
to be 1kHz if R_a and R_b were 20K. Since
standard values of resistors and capacitors
preclude this possibility, the above values
are given; considering the nature of the VCO
pin and its associated VREF, it is easier to
lower the frequency than to raise it.]

We will have need for the squarewave output; it goes through 10K to VCC.

The usual way to assess the symmetry is to
look at the squarewave on an oscilloscope;
the pot is then adjusted so that the top and
bottom portions of the squarewave are of
equal length. However, until we all have
"scopes," I recommend the following
alternatives:

1. Pad the timing capacitor with a
   large value unit, say from 0.5uF on up
   through 5uF. Connect the squarewave output
   to a test amplifier so as to permit adjustment by ear. The 5uF unit will produce a
   series of clicks which you can adjust for
   equal timing, similar to leveling a mechanical metronome. I prefer a fast buzz, such as
   the 60hZ gotten with a 0.5uF unit. This
   latter system will emit a pure buzz when the
   adjustment is correct; any deviation from
   this point will give the illusion of producing two frequencies (sort of like
   listening to the echoes of the old LORAN
   system on 160 meters).
2. The charge and discharge currents to
   which the timing capacitor is subjected can
   be measured with a sensitive milliammeter,
   and the adjustment can be made so as to make
   these equal in magnitude. With a pot connected across the supply (I used a 5K unit
   between pins 6 and 11), connect the arm of
   the pot through a milliammeter to pin 10 (the
   top of the timing capacitor). This can be
   done with the timing capacitor left in place.
   Turn the pot down so that you are sure its
   wiper has gone below 1/3 VCC and then advance
   it to about mid-position; take note of the
   current. Next, advance the pot until you are
   sure it has gone above 2/3 VCC, and then back
   it down to about mid-position; take note of
   this current. (Of course, the polarity of
   the meter will have to be alternated, since
   these currents will be in opposite directions.) Make the necessary adjustment of the
   pot on pins 4 and 5 so as to equalize these
   currents. Caution--do not believe a reading
   while the pot is outside the limits of 1/3 to
   2/3VCC; the circuits in the chip do funny
   things when forced to their extremes.

Once temporal symmetry has been accomplished, the square and triangle waves will
be as pure as the driven snow.

Shaping the Sinewave

Eight junctions on a
voltage divider determine the "breakpoints"
or points of progressive loading which make a
sinewave out of a triangle. The top and
bottom junctions (pin 1 and pin 12, respectively) are made available so that this
divider string may be externally influenced
to achieve a more precise replica of a sine-
wave. Left untouched (without any trimming),
you can expect the harmonic distortion to be
about 5% or so. Two schemes, one simple and
one elaborate, are commonly used to vastly
improve the above distortion figure.

The simple setup merely consists of running
pin 12 through 82K to -V. One might expect
that this would offset the "sine converter,"
and that another resistor should be used at
pin 1; this is not the case, and the distortion can be reduced to perhaps 0.8% with
this simple approach. Replacing this resistor with a 100K rheostat is often done to get
an additional slight improvement.

The fanciest scheme is as follows:

Pin 12 goes to the arm of a 100K pot; the
bottom of this pot goes to -V, while its top
end goes through 10K to VCC. Similarly, pin
1 goes to the arm of another 100K pot; the
bottom of this pot goes to VCC, while its top
end goes through 10K to -V. These two units
are then jockeyed in combination to get minimum distortion, better than 0.5%.

The usual way of assessing the distortion
and trimming it for minimum is to examine the
wave on an oscilloscope, especially while
comparing this with a true sinewave superimposed on it. An alternative would be to
run the sinewave output through a notch
filter which is tuned to the fundamental
frequency; the remaining signal will be
comprised of the harmonics which are to be
minimized. The residual signal can be tuned
for a minimum output, either audibly or with
the help of an AC VTVM.

A notch filter can be made by combining a
second-order low-pass and a second-order
high-pass filter (both with the same cutoff
frequency) in a summing amplifier (an inverting op-amp configuration with two equal
input resistors). [The low- and high-pass
filters are described in "Op-Amps III"; the
summing amplifier is discussed in "Op-Amps I"
(Summer 1983 and Winter 1983, respectively).]
On the other hand, a notch filter using the
Reticon 5620 (Summer 1983) is easy to build
and is tunable to boot.

A 20-to-20,000 Hertz
Audio Oscillator

This comes from an Intersil pamphlet,
"Everything You've Always Wanted to Know
about the 8038." The intention was to show
how the device could be used over its advertised frequency range of 1000-to-1. I would
never build this circuit; it calls for such
items as a logarithmic rheostat (which would
be expensive), and it is loaded with trimming
adjustments. However, it is an outstanding
example of how to hang every imaginable thing
onto this chip; any one of these tricks may
be individually useful to you.

Circuit

Plus and minus 15V supplies are
used; their common connection is grounded.
Power to the chip is taken across the whole
30V. Pin ll goes to -l5V, while pin 6 goes
through a diode to +l5V (cathode toward pin
6). (This diode will be explained later.)
Pin l0 goes through a 0.0047uF timing capacitor to -l5V. Pins 4 and 5 each go through
4.7K, with the far ends of these resistors
going to the ends of a lK symmetry adjustment
pot. The arm of this pot goes to pin 6, the
plus supply pin. Pin 5 also goes through a
l0meg rheostat to -l5V (to be explained
later). For frequency adjustment, pins 7 and
8 are tied together and go through a l00K log
pot, connected as a rheostat, to +l5V. Pins
7 and 8 are also bypassed to +l5V by 0.luF
(made necessary only when the squarewave pin
is connected).

The squarewave output, pin 9, goes through
a l5K pull-up resistor to +l5V.

For trimming distortion in the sinewave,
this circuit shows pin l2 going through a
l00K rheostat to -l5V. For completeness
here, however, we will restate the fanciest
arrangement: Pin l2 goes to the arm of a
l00K pot; the bottom of this pot goes to
-l5V, while the top of the pot goes through
l0K to +l5V. Pin l goes to the arm of a l00K
pot; the bottom of this pot goes to +l5V,
while the top of the pot goes through l0K to
-l5V.

Notes on the Above Circuit

The diode in
the plus supply line enables you to take pins
7 and 8, used here for frequency control, to
about 0.6V above the supply pin, pin 6. This
allows you to get very low frequencies (without changing the timing capacitor, which is
another way of getting them).

The l0meg rheostat on pin 5 is necessary
because, as pins 7 and 8 approach +l5V, the
duty cycle shifts due to transistor mismatches in the current source elements.
These mismatches can be compensated for by
bleeding a small amount of current away from
pin 5, hence the l0meg rheostat.

Intersil recommends making the duty cycle
adjustment (using the lK pot off pins 4 and
5) with the oscillator set to its highest
frequency--with the frequency control wide
open. Then, with the frequency control set
to generate the lowest frequency of interest,
decrease the l0meg rheostat to again balance
the temporal symmetry.

Finally, they also caution that, unless the
supply has excellent dynamic regulation,
transients from the squarewave will find
their way onto the triangle, and hence onto
the sinewave. It may be necessary to bypass
both supply pins to ground near the chip.
They don't recommend a particular value; try
0.l's for starters.

A HIGHLY PORTABLE FUNCTION-GENERATOR
USING THE ICL 8038

I have never had the time to build this
instrument; I designed it using the information contained in the cogent, well-written
article above. It is an instrument which I
would love to have; perhaps an ambitious
reader might build one for me for presentation on one of my birthdays (March l2,
April l, October 3l, etc.)--Thank you very
much.

As designed, this instrument produces sine,
square, and triangle waveforms whose peak-
to-peak amplitude is adjustable over two
ranges--0 to 10V, and 0 to 3V. Its five
tunable frequency ranges are: 0 to 10Hz, 0
to 100Hz, 0 to 1kHz, 0 to 10kHz, and 0 to
20kHz. If a semi-precious linear pot is used
for the frequency control, it is well worth
attaching a 10- or 20-division braille dial
so that the frequency can be read directly.
(The bottom 10% of the scale will be
extremely inaccurate and should be ignored.)

Circuit

A split supply is used, ±9V to
±15V. The negative side of the plus supply
is grounded, while its positive terminal goes
through one pole on a DPST on-off switch to
the +V line. The positive side of the negative supply is grounded, while its negative
terminal goes through the other pole on the
switch to the -V line. The +V line is bypassed to ground by l00uF (negative at
ground). The -V line is also bypassed to
ground by l00uF (positive at ground).

Pin ll of the 8038 goes to -V; pin ll is
also bypassed to ground by 0.luF (located
near the chip). Pin 6 goes to +V; pin 6 is
also bypassed to ground by 0.luF (located
near the chip). Pin 8, the VCO control pin,
is bypassed to the +V line by 0.luF (located
near the chip).

Pin 7, the VREF pin, goes through a l0K
calibration rheostat, then through a l00K
linear pot to the +V line. The arm of this
l00K unit (the frequency control) goes to
pin 8.

Pin 4 goes through 24K to one end of a l0K
symmetry-adjustment pot; pin 5 goes through
24K to the other end of this pot. The arm of
this pot goes to the +V line.

Pin l0, intended for the timing capacitor,
goes to the arm of a 5-position rotary switch
(range switch). Position l goes through luF
to -V (25V tantalum with its positive toward
position l). Position 2 goes through 0.luF
(Mylar) to -V. Position 3 goes through
0.0luF (Mylar or mica) to -V. Position 4
goes through 0.00luF (mica) to -V. Position
5 goes through 500pF (mica) to -V.

The fanciest distortion trimming is connected. Pin l2 goes to the arm of a l00K
pot; the bottom of this pot goes to -V, while
the top goes through l0K to +V. Pin l goes
to the arm of another l00K pot; the bottom of
this pot goes to +V, while the top goes
through l0K to -V.

A 3-pole, 3-position switch is used to
select the functions; one pole disables the
squarewave output when not in use, one pole
feeds the desired oscillator output into the
buffer amplifier, and the third changes the
gain of the buffer to make their amplitudes
equal.

Pin 9, the squarewave output, goes through
l0K to position 3 of pole No. l; the arm of
this pole goes to +V.

On pole No. 2, position l goes to the sine
output, pin 2 of the 8038. Position 2 goes
to the triangle output, pin 3. Position 3
goes to the squarewave output, pin 9. The
arm of this switch goes through a 500K output
level control to ground. (This control can
be log, linear, audio taper, as you wish.)

On pole No. 3, position l goes through Rl
to ground, position 2 goes through R2 to
ground, and position 3 goes through R3 to
ground. These resistors will be tabulated
later on the basis of supply voltage.

A fast op-amp is used as a buffer; I have
always like the Signetics NE535 which does
not need external compensation. Pin 4 of the
op-amp goes to -V; pin 7 goes to +V. Between
pins 6 and 2, from output to inverting input,
is a selectable feedback resistor to provide
two gains. Pin 6 goes through 43K, then
through 20K to pin 2. Across the 43K resistor is an SPST gain switch.

Pin 2 of the op-amp, the inverting input,
goes to the arm of pole No. 3 of the waveform
selector switch. Pin 3 of the op-amp, the
non-inverting input, goes to the arm of the
output level control.

The cold output terminal of the instrument
is grounded. The hot output terminal comes
directly from pin 6 of the op-amp.

Table of Input Resistors
R1, R2, and R3

For a split supply of ±9V:

• R1--7.5K
• R2--l2K
• R3--36K

For a split supply of ±l2V:

• Rl--l0K
• R2--l6K
• R3--47K

For a split supply of ±l5V:

• Rl--l3K
• R2--20K
• R3--62K

Happiness is a Warm Screwdriver

With the frequency control turned up, the
temporal symmetry adjustment (the l0K pot off
pins 4 and 5) should be adjusted first while
looking at the squarewave output. (The range
switch can be in any position as appropriate
given the technique used; see previous
article. My favorite technique is the trick
with the current meter, although the symmetry
can be adjusted audibly with the range switch
in one of the two lower positions.)

Next, adjust the l0K rheostat off pin 7 so
that, with the frequency control turned all
the way up, frequencies close to the maxima
specified are gotten. A compromise here or
there may be necessary due to distributed
capacitance, etc. Because of loading on the
VREF pin by the frequency control circuit, it
may be impossible to reach the desired
maxima, even with the calibration rheostat
closed. If you encounter this problem, put a
slight load on the VREF pin in the opposite
direction; i.e., put a resistor of 470K or
680K from pin 7 to -V.

Trimming of the sinewave is perhaps best
done visually on an oscilloscope. However,
feeding the sinewave through a notch filter
tuned to its fundamental frequency will
present the distortion products audibly so
that they can be tuned for a minimum. If
worse comes to worst, listen to the sinewave
with your naked ear and tune the trim pots
for the purest tone (maximum "oooooh" and
minimum "eeeeeh").

Parts List

Resistors (1/4 watt, 5%):

• 3--l0K
• l--20K
• 2--24K
• l--43K
• Rl, R2, R3--see table above

Potentiometers (screwdriver-adjust trim
pots):

• 2--l0K
• 2--l00K
• Linear precision pot, l00K
• Output level control, 500K

Capacitors:

• 2--l00uF electrolytic 25V
• l--luF tantalum 25V
• 3--0.luF disc
• l--0.luF Mylar
• l--0.0luF Mylar or mica
• l--0.00luF mica
• l--500pF mica

Integrated circuits:

• l--ICL 8038, Radio Shack 276-2334
• l--Signetics NE535 or other fast op-amp

Switches:

• l--SPST toggle
• l--DPST toggle
• l--3-pole, 3-position rotary
• l--single-pole, 5-position rotary

THE SMITH-KETTLEWELL
UNIVERSAL TONE INDEXER

Abstract

A low-frequency tone generator
using the ICL 8038 chip is described. As an
outboard unit, this device can be plugged
into any tape recorder for the purpose of
adding low-frequency tones to mark chapter
headings, pages, etc.

Description

Two versions of our tone indexer will be
described; one has a built-in microphone,
while the other has provision for connecting
a remote microphone which can be used to
control the tones or to start and stop the
tape through the remote jack of the tape
recorder. The more complicated unit,
intended for use with a dynamic microphone
having a remote switch, contains the following bells and whistles:

1. Two cables emerge from the box; one
   plugs into the mike input of the tape
   recorder, and the other plugs into the tape
   recorder's remote control jack.
2. A jack to receive the microphone plug
   is mounted at one end of the tone indexer.
3. Located one centimeter away from the
   mike jack is a "remote control jack" through
   which the tape recorder can be turned on and
   off in the usual way.
4. Located one centimeter away and on the
   other side of the microphone jack is a
   "remote tone jack" which permits the tones to
   be controlled by the switch on the microphone.
5. A "line-input" or "auxiliary" jack is
   included to permit recording off of high-
   level sources (such as another tape recorder). High-level material fed into this jack
   is attenuated so as to be of an appropriate
   level to feed into the recorder's mike input.

One final feature of my design is that no
current is drawn by the tone generator
between button pushes. The unit containing a
microphone draws very little current (500uA);
an on-off switch is nevertheless provided on
this unit. No on-off switch need be included
for the other version, since the pushbutton
serves this function.

It should be mentioned that tone indexers
are commercially available. One can be
gotten from Science for the Blind, P.O. Box
385, Wayne, PA l9087. The other is very
recent; it is a companion to the "Talkman"
tape machine, available from BIT, 24l
Crescent Street, Waltham, MA 02545.

[Chris Mackey (Mrs. H.V. Mackey, 202 West
Grangeville Blvd., Hanford, CA 93230), in
various talks on tone indexing, describes a
simple system which can be used with tape
machines on which separate input level controls allow mixing of a line input with the
mike. Into the line input, she inserts a
patch cord. A perfectly good 60 Hertz hum is
then gotten by touching the tip of the plug
at the free end of the patch cord, taking
care not to ground yourself via the shield of
the plug or by holding onto the mike,
machine, etc.

Over-cultivated technologists should note
that this is "engineering" in the purest
sense of the word; this approach is
"ingenious." This tells the true story of
the trade; the two words, engineering and
ingenious, stem from the common Latin root,
"ingenium."]

Circuit Operation

The tone generator consists of an ICL 8038
function generator chip connected in its
simplest form. Its sinewave output is fed
directly to the mike input via a resistor,
rather than being combined in an op-amp
mixer. The version intended for use with a
dynamic mike actually uses the microphone
element as part of its attenuator (to attenuate the tone and present only a fraction of
it to the input jack). Condenser microphones, such as that which is contained in
the other version, provide no low-impedance
path for this primitive system to work.
Therefore, the values of resistors chosen are
different for the two versions. I could have
played around with these values in order to
strike some sort of compromise, but I decided
that if I made this primitive system do too
much, it would do nothing very well.

The frequency of the oscillator is variable
over a range of from 25 to 150 Hertz. A PC-
mount trim pot (connected as a rheostat) is
used for this purpose; a hole was drilled in
the box so that this adjustment could be
accessed from the outside.

Circuit for the Remote-Controlled Version

The negative side of the battery goes through
a normally open pushbutton switch to ground.
Across the switch is connected a 1/16 inch
remote tone jack; the sleeve of this jack is
grounded, while its tip goes to the negative
side of the battery. Pin 11 of the ICL 8038
is grounded. Pin 6 of the 8038 goes through
100 ohms to the positive side of the battery;
pin 6 is bypassed to ground by 10uF (negative
at ground). (The latter components, comprising what looks like a decoupling network,
are included to reduce transients when the
oscillator is turned on and off.)

Pin 10 of the 8038 goes through 0.1uF to
ground. Pins 4 and 5 are tied together and
go through 10K, then through a 50K rheostat
(pitch control) to pin 6. Pins 7 and 8 are
tied together.

The sinewave output, pin 2, goes through
0.luF, then through 330K to the tip of the
microphone jack--this jack is a closed-
circuit 1/8 inch mini phone jack. The switch
contact of this jack goes through 330 ohms to
the sleeve, with the sleeve of this jack
being grounded. (This 330 ohm resistor on
the switch contact is in the circuit when the
mike is unplugged, thus providing the oscillator and the line input jack with the
necessary voltage divider.)

The tip of the mike jack also goes through
33K to the tip of the high-level input jack.
The sleeve of this latter jack is grounded.
(This can be an open circuit jack, since the
switch contact serves no purpose; it could
even be an RCA jack if you wish.)

The cold output terminal (the sleeve of a
suitable plug for your tape recorder input)
is grounded. The hot output terminal (the
tip of the plug) goes directly to the tip of
the mike jack. (This plug is usually a 1/8
inch mini phone plug.)

So as to permit starting and stopping of
the tape recorder, a remote control jack is
provided and connected as follows. The
sleeve and switch contact of closed-circuit
1/16 inch jack are tied together and go to +-the sleeve of the remote control plug (this
plug is usually 1/16 inch sub-mini phone
plug). The tip of the jack goes to the tip
of the plug. Sometimes the remote jack in
the tape recorder is off ground; its sleeve
is sometimes not tied to the sleeve of the
mike jack. For this reason, it is unwise
(also unnecessary) to ground the sleeve in
the Indexer circuit.

Circuit for Self-Contained Version

A Radio
Shack 270-092A condenser mike element is
used. The shield of its output cable is
grounded; the center conductor (mike output)
goes through 0.22uF, then through 2.2K to
ground. Its power connection (a red wire
separate from the shielded cable) goes
through an SPST on-off switch to the positive
battery terminal.

The negative side of the battery is
grounded. The positive battery terminal goes
to one side of a normally open pushbutton
switch. The other side of the pushbutton
goes through 100 ohms, then through 10uF to
ground (negative at ground). The junction of
this resistor and capacitor goes to pin 6 of
the 8038. Pin ll of the 8038 is grounded.

Pins 7 and 8 of the 8038 are tied together.
Pin l0 goes through a 0.luF timing capacitor
to ground. Pins 4 and 5 are tied together
and go through l0K, then through a 50K rheostat (frequency adjust) to pin 6.

Pin 2 goes through 0.luF, then through 680K
to the junction of the 2.2K resistor and the
coupling capacitor in the mike output circuit. A single mike plug on shielded cable
is used for connection to the tape machine.
The sleeve of this plug is grounded, while
its tip goes to the top of the 2.2K resistor,
to the mike's coupling capacitor, and to the
680K dropping resistor.

Mounting Considerations

The version which contains its own microphone has a problem of picking up vibrations
as the unit is handled and operated. Though
I have not done so in our prototype, it might
be wise to rubber mount the microphone.
Rubber grommets of sufficient size (an inner
diameter of 5/16 inch) are large and thick-fitting them onto such a small box would be
quite a problem. Instead, I recommend gluing
the microphone into a rubber "washer" (perhaps cut from rubber gasket material) and
gluing the assembly into a hole which is
larger than the microphone (treating the
rubber as a "mounting flange").

The double plugs on remote control microphones often have a spacing of 10mm (1 centimeter) between centers of their prongs. It
is handy to note that 10mm is very close to
4/10 inch.

In making holes for these jacks, I use
vectorboard with tenth-inch hole spacing as a
template. For example, in this instrument we
want 3 holes in a row with a space of 4/10
inch between them. I determined the position
at which I wanted the microphone jack (the
middle one); I drilled a hole through the
vectorboard into the box in this position.
From this point, I drilled again in hole No.
4 (to the right), and again in hole "minus 4"
(to the left). Very little filing was necessary to make these jacks comfortably fit
their intended plug.

You can dodge this whole issue by using one
of Radio Shack's "Universal Replacement
Microphones" on which the two plugs extend
from separate short lengths of cable.

Parts List

Remote-Controlled Indexer:

• 1--100 ohm, 1/4 watt
• 1--330 ohm, 1/4 watt
• 1--10K, 1/4 watt
• 1--33K, 1/4 watt
• 1--330K, 1/4 watt
• 1--50K trim pot connected as a rheostat
• 1--10uF, 10 volt
• 2--0.1uF discs
• 1--ICL 8038
• 1--1/16 sub-mini phone jack, open-circuit
• 1--1/16 sub-mini phone jack, closed-circuit
• 1--1/8 inch mini phone jack, open-circuit
  (or other type)
• 1--1/8 inch mini phone jack, closed-circuit
• 1--1/8 inch mini plug on shielded cable
• 1--1/16 inch sub-mini plug on unshielded wire
• 1--normally open pushbutton switch

Indexer with Microphone:

• 1--l00 ohm, l/4 watt
• l--2.2K, l/4 watt
• l--l0K, l/4 watt
• l--680K, l/4 watt
• l--50K trim pot connected as rheostat
• l--l0uF, l0V electrolytic
• l--0.22uF disc or Mylar
• 2--0.luF discs
• l--ICL 8038
• l--Radio Shack 270-092A microphone
• l--SPST on-off switch
• l--normally open pushbutton switch

THE SMITH-KETTLEWELL
SPEAKER PHASE TESTER

Abstract

This battery-operated unit
plugs into "auxiliary" inputs of any hi-fi
amplifier. It simulates two popular phase-
testing signals which are often found on test
records. It permits testing of speaker
phasing without disconnecting one of the
speakers.

I have always had trouble putting my
speakers in phase. First of all, quick
comparisons are not possible because my
speakers are connected with screw terminals.
They have to be tested as connected one way;
then after a significant period of time
during which one is being switched around, a
listening test is performed with the phase
reversed. With this device, however, a
direct "A/B comparison" can be made which
will tell you if your speakers are in or out
of phase.

The simplest way of making a testing jig is
to build a polarity reversing switch into one
of the speaker lines. Making such a jig
detachable, however, means that you must
account for all the different kinds of
speaker plugs, binding posts, etc. It seemed
to me that a couple of common tests such as
those found on test records could be
generated using an ICL 8038 function-
generator chip. Both the polarity-reversing
switch and the eventual speaker phase tester
prototype are described here.

Simple Polarity-Reversing Switch

Let us
get this out of the way before describing the
more complicated device. A double-pole
double-throw switch (without center off) is
used. Position l of pole A goes to position
2 of pole B. Position 1 of pole B goes to
position 2 of pole A. The output terminals
of the amplifier are connected: one to
position l of pole A and the other to
position l of pole B. The arms of the switch
drive the speaker.

With the switch in position l, the speaker
lead on the arm of pole A is connected to
position l of pole A; i.e., the polarity is
not reversed. With the switch flicked to
position 2, the polarity is reversed.

With a monaural signal feeding both
channels of the amplifier, try the two
positions of the switch; the speakers are in
phase when the monaural material can be
clearly localized as being in the center-between the two speakers.

The Fancy Speaker Phase Tester

An 8038 is connected so as to generate two
waveforms: One is a 75Hz sinewave. The
other is a square pulse with a duty cycle of
about 5%. Two phono plugs on a cable emerge
from the instrument. A pushbutton switch on
the front of the tester determines the phase
relationship of the signals available at the
plugs. Normally, these signals are in phase;
when the switch is depressed, the signals are
l80 degrees out of phase.

When the 75Hz tone is used, a difference in
loudness can usually be discerned. With the
speakers in phase, their motion is such as to
be additive; the listener hears the combined
power of both signals. With the speakers out
of phase, their motion is such as to cancel;
a decrease in loudness of the tone should be
evident. (This system falls short of being
ideal if standing waves in the room are such
that the listener finds himself in a node of
the soundwave. For this reason, make your
cables long enough so that you can move
closer to and away from the speakers.)

The pulses are intended to simulate a
metronome which, if the speakers are working
in phase, can be localized as being in the
center between the two speakers. If the
speakers are out of phase, the clicking sound
cannot be localized; it may sound as if it
comes from the sides or from nowhere in
particular. Stepping from side to side or
operating the balance control will give
further evidence of this difference.

Circuit

Pin ll of an ICL 8038 is grounded,
as is the negative side of the 9V battery.
The positive terminal of the 9V battery goes
through an on-off switch to pin 6 of the 8038
and to the VCC line. Pins 7 and 8 of the
8038 are tied together and bypassed to the
VCC line by 0.luF.

Pin l0 of the 8038 goes through 0.luF to
ground. Pin 4 goes through 36K to one end of
a l0K symmetry adjustment pot. Pin 5 goes
through 820K, then through 36K to the other
end of this pot. The arm of this pot goes to
the VCC line. (A switch to be described will
short out this 820K resistor.)

Pin 9, the squarewave output, goes through
47K to VCC. For crude shaping of the sine-
wave, pin l2 goes through 82K to ground.

Two inverters in cascade provide buffered
signals of both phase polarities. An LM358
or other dual op-amp is used. Pin 4 is
grounded, while pin 8 goes to VCC. The non-
inverting inputs, pins 3 and 5, are tied
together and go to a reference voltage. The
reference is gotten off a voltage divider;
two resistors of 22K are connected in series
between VCC and ground. The junction of
these resistors is bypassed to ground by luF
(negative at ground).

Pin 7, the output of the second inverter,
goes through a l00K feedback resistor to pin
6, the inverting input. Pin 6 also goes
through l00K to pin l, the output of the
previous stage. Between pins l and 2 is
connected a feedback resistor of l00K.

Pin 2, the input of the first stage, goes
through 100K to the arm of a double-pole
double-throw selector switch. Position 2 of
this pole goes through 220K to the squarewave
output, pin 9 of the 8038. Position l goes
to the sinewave output, pin 2 of the 8038.
(These connections are made with shielded
cable so as to keep the pulses from pin 9 out
of the sinewave signal.)

Position l of the second pole goes to pin 5
of the 8038, while the arm of this pole goes
to the junction of the 820K and 36K resistors.

Pin l of the op-amp, the output of the
first stage, goes through 0.luF to the center
conductor of one of the output cables; the
shield of this cable is grounded. The center
conductor of the other output cable goes to
the arm of a single-pole double-throw push-
button switch; its shield is left floating so
as to avoid ground loops. The normally
closed contact goes to the junction of the
output capacitor and the center conductor of
the first output cable. The normally open
contact goes through 0.luF to pin 7 of the
op-amp, the output of the second stage.

The sinewave test does not work well if the
harmonic content is high. For this reason,
the symmetry adjustment was provided to at
least enable you to balance both half-cycles.
This can be done by turning the treble
control of the hi-fi amplifier up and its
bass control down--then adjusting the l0K pot
for minimum high-frequency buzz. When using
the sinewave to test the speaker phase, on
the other hand, turning the treble control
down and the bass control up will aid in
presenting an optimal signal.

Parts List

Resistors (l/4 watt, 5%)

• 2--22K
• 2--36K
• 1--47K
• l--82K
• 4--l00K
• l--220K
• l--820K
• l--l0K trim pot

Capacitors

• 4--0.luF disk or Mylar
• l--luF electrolytic

Semiconductors

• l--ICL 8038
• l--LM358

Switches & Plugs

• l--SPST toggle
• l--DPDT toggle
• l--SPDT pushbutton
• 2--RCA plugs on shielded cables 6 ft. or longer

THE SILKIEST MCW SIGNAL ON THE AIR

Introduction

Here in the Bay Area, there

are 2-meter repeater and symplex frequencies
on which MCW operation is tolerated and even
encouraged. Most of the folks place their
microphone next to the speaker in their
keying monitor; a few others use a rasty-
sounding oscillator made with a 555 timer,
or the like, piping this directly into their
mike input. Room echo and/or the excessive
generation of pairs of sidebands are not my
style, and I set about giving WA6NPC a tone
with some class.

There are other uses for this device:
Tom Fowle, a computer technician at Smith-
Kettlewell, uses it as a Morse Code input/
output device for his computer. It makes a
high-grade non-fatiguing code practice
oscillator for anyone teaching the Code or
making Code practice tapes. Our Talking Book
readers will note that this device is what I
use to put the Morse Code identification tags
into the Table of Contents. We have even
used this circuit to make a tactile keying
monitor for the deaf.

Description

The heart of this unit is an
ICL 8038 connected in the conventional way.
The only variation is the addition of the
keying transistor which shorts out the timing
capacitor when the key is up. As a built-in
monitor, the device contains an LM386 audio
amplifier.

The unit is built in a 2 by 3-l/2 by 6 inch
metal utility box, metal being essential if
it is to be used around RF. The front panel
contains a volume control, a pitch control,
an on/off switch which also doubles as the
transmit/receive switch, a key jack, and an
earphone jack.

The rear panel contains a 4-pin jack by
which a patch cord connects the unit to your
transceiver; a high-level output jack (which
can be patched into a tape recorder) is also
located on the rear. (The volume control
does not affect the level at these output
jacks.) Finally, an internal trim pot sets
the level as desired into your transmitter's
microphone input circuit.

Embellishments would include another simple
open-circuit jack to serve as a low-level
output suitable to feed the mike input of
tape recorders which do not have "auxiliary."
Use of a shorting-type key jack will allow
the unit to be used as a general purpose test
oscillator, say for checking wow and flutter
in tape recorders. A sine/triangle switch
would permit you to choose the tone which
best suits your mood.

As shown, the available pitch range is from
a few hundred cycles to nearly l500 Hz (about
a 4-to-l range). Later on, appropriate
modifications will be given to provide the
tactile user with a range from l50 to about
300 Hz.

Circuit

A double-pole single-throw on-off
switch is used. One pole operates the push-
to-talk circuit of your transceiver. The
other pole of the switch goes between the
positive 9V battery lead and the VCC line.
The negative side of the battery is grounded.
The VCC line is bypassed to ground by l00uF
(negative at ground).

Pin ll of the 8038 is grounded, while pin 6
goes to the VCC line. Pins 4 and 5 are tied
together and go through l0K to VCC.

Pin l0 goes through a 0.0luF timing capacitor to ground. Pin l0 also goes to the
collector of a 2N2222, with the emitter of
this transistor being grounded. The 2222
base goes through 33K to VCC. This base also
goes to the tip contact of the key jack. The
sleeve of this jack (and possibly its switch
contact; see "Embellishments") is grounded.
The transistor base is also bypassed to
ground by the parallel combination of 0.0luF
and 0.22uF (the latter being a key-click
filter).

Pin 7, VREF, goes through a l00K linear pot
(pitch control), then through 33K to VCC.
The arm of this pot goes to pin 8, the VCO
control pin.

To take some of the harmonics out of the
sinewave, pin l2 goes through 82K to ground.

Pin 2, the sine output, or pin 3, the
triangle output--as you prefer--goes through
0.luF, then through 220K to the top of the
low-level output adjustment (a 2K trim pot).
The bottom of this pot is grounded, while its
arm goes to the hot mike input of the transceiver. This arm is also bypassed to ground
by 0.0luF.

The junction of the above 0.luF and 220K
goes to the hot output of the high-level
recording jack. The low-level recording jack
can either go to the arm of the 2K output-
level adjustment, or it can go to the top of
this pot so as not to be affected by the
level setting. The sleeves of these output
jacks are grounded.

The selected output, either pin 2 or pin 3
of the 8038, also goes through 0.luF, then
through 220K to the top of the volume control
(l0K audio taper). The bottom of this control is grounded.

An LM386 is used for the internal power
amplifier; pins 2 and 4 of this chip are
grounded, while pin 3, the input, goes to
the arm of the volume control. Across pins 2
and 3 is connected a 0.0luF disk capacitor
(to keep stray animals out of the amplifier).
If you want the full l25mW out of the 386,
connect pin 8 through lK, then through l0uF
to pin l (positive of the capacitor toward
pin l).

Pin 6 of the 386 goes through l0 ohms to
VCC; pin 6 is also bypassed to ground by
250uF (negative at ground). Pin 7 is bypassed to ground by 25uF (negative at
ground).

Between pins 4 and 5 is 0.22uF. Pin 5 also
goes through l00uF (positive toward pin 5) to
the tip of the earphone jack. The sleeve of
this jack goes to one side of the speaker, as
well as going to ground. The other side of
the speaker goes to the switch contact of
this jack.

I have only used my prototype of this
instrument with very low power transceivers,
l.5 to 2 watts. Even so, any attempt to use
an indoor antenna severely affects the performance of the device. If you have trouble
with RF interference, it may be necessary to
insert RF chokes in the hot leads of all
jacks; then bypass each jack with 0.0luF
(located on the jack terminals). Further,
the ground lead from the circuit board to the
metal chassis should contain an RF choke to
completely isolate the circuit from RF
present on the box. (Unless low-resistance
high-frequency RF chokes are used, the chokes
will introduce a prohibitively high resistance in the speaker/earphone system. If
chokes are included, eliminate the earphone
jack entirely, or wind the output leads in a
toroid on a ferrite coil form.)

Our tactile version is identical with two
modifications. First, the timing capacitor
is increased to 0.047uF; this lowers the
frequency to a maximum of about 300 Hertz.
Next, the range of the pitch control is
restricted; pin 7 goes through the l00K pot,
then through a l00K resistor to VCC. The
power delivered by the LM386 should be enough
to drive a fairly efficient transducer.

Parts List

Resistors (l/4 watt, 5%)

• l--l0 ohm
• l--lK
• l--l0K
• 2--33K
• l--82K
• 2--220K

Potentiometers

• l--2K PC-mount
• l--l0K volume control
• l--l00K linear pitch control

Capacitors (disk ceramic, 30V)

• 4--0.0luF
• 2--0.luF
• 2--0.22uF

Capacitors (electrolytic, l0V)

• l--l0uF
• l--25uF
• 2--l00uF
• l--250uF

Semiconductors

• l--2N2222
• l--ICL8038
• l--LM386

Jacks & Switches

• l--Open- or closed-circuit key jack
• l--4-pin mike connector
• l--RCA phono jack
• l--Closed-circuit earphone jack
• l--DPST switch

Miscellaneous

• l--Loudspeaker
• l--Metal box
• l--9V battery

SAVE YOUR BATTERIES WITH A TIMER SWITCH

Abstract

Circuits for giving your battery
projects a timer switch are given here. As
their switching element, they use recently
available "VMOS" power FET's; a survey of
these is given at the end of this article.

How many times in the hot pursuit of a DX
station on the Clipperton Islands have you
reached for your "Auditory Gimmick" only to
find that it has been left on for weeks and
you cannot now tune your transmitter? No?
Well then, how many times have you, in your
attempt to capture a rare recording of Nellie
Melba singing, "I Didn't Raise My Boy to be
in Software," have you reached over to find
that your audible VU meter has been left on
all night? In any case, the replacement
batteries are a needless expense and the
inconvenience is untenable.

Nowadays, items such as calculators and the
"Speak-and-Spell" toy turn themselves off
automatically so as to save their batteries.
The internal program of these devices
contains a digital counter (probably running
off the main clock) which ticks away the
minutes until a BIT in the string orders a
main switch to open. My first thought was to
take this approach using a count-down timer
chip such as the XR2242 (discussed in our
"Universal Nicad Battery Charger," SKTF,
Winter 1981). Besides the switching element,
this would necessitate adding a whole new
chip to every project--too much work.

On the other end of the spectrum is an
approach of extreme simplicity. A power FET

can be used as the switch, with its gate
being controlled by a simple RC circuit. The
gate circuit is arranged so that, as the
capacitor charges, the gate gradually loses
its forward bias; this will result in the
gradual opening of the FET channel. (This
scheme was written up in "Popular Electronics," June 1982; the article, "Ultra-
Simple VMOS Timer," was co-authored by
Forrest Mims and Istvan Mohos.) Its disadvantage is that the switch does not
abruptly open; the project just fades away.

I sought the middle ground, a simple
circuit with solid on-off characteristics--
one which could be expanded to include automatic resetting if the project stays in use.
In my basic circuit, I use a VMOS power FET

as a switch, this in conjunction with a
"latching transistor." They hold each other
on until bias begins to fall away, then they
act to turn each other off. A third transistor can be added to dump the charge in the
timing capacitor, resetting the timer from
activity in the project.

Circuit Operation

In all cases, the
battery goes to the source of the FET, while
the drain goes to the project. (The negative
or positive battery lead can be switched by
choosing an appropriate gender for the FET--
the negative side can be switched with an N-
channel unit, while a P-channel is required
for switching the positive lead.) Being
"enhancement-mode" FET's, they will be open
(biased at cutoff) when their gate potential
is near that of the source; bringing the gate
toward the other battery lead will cause the
"switch" to close (biasing the FET into
saturation).

From the opposite supply line (the one not
containing the FET switch) is a transistor
which, when turned on, forward-biases the
gate of the FET so as to hold it on. Not
only does this cause power to be applied to
the project, but the FET drain also forward-
biases the base of the transistor through a
resistor and capacitor in series. In other
words, turning on the transistor causes the
FET to turn on; the FET in turn holds the
transistor on.
Eventually, however, a charge is developed
on the capacitor in the transistor's base
lead, and the base bias begins to fall away.
When this happens, the transistor begins to
relinquish its hold on the FET gate, and the
FET comes out of saturation. But the FET is
what has been holding the transistor on.
Immediately as the FET opens, the base of the
transistor is driven in the direction of
reverse bias; its latching action collapses
and the gate of the FET goes back to the
source, thus cutting off the switch.

The "hang time" (the length of time the
circuit stays latched) is about three time-
constants of the RC circuit on the base of
the latching transistor. With the 47uF
capacitor and the 680K resistor shown, your
project will stay energized for about 95
seconds. In the refreshable version, where a
third transistor dumps the charge on the 47uF
capacitor through a 22 ohm resistor, a five
millisecond pulse constitutes five time-
constants of this latter network, thus
resetting the timer.

Circuits

Mirror-image circuits for both P- and
N-channel FET's will be given. My favorite
uses a P-channel unit in the plus supply
line; its refresher circuit requires downward
pulses which are usually available from
projects containing NE555 oscillators. On
the other hand, N-channel units seem to be
more readily available, units of this gender
being stocked at Radio Shack.

P-Channel Timer Switch for the Positive
Supply Line

The source of the FET (P-channel
such as the Siliconix VP0300M) goes to the
positive side of the battery. The negative
side of the battery goes to the -V line of
the project. The drain of the FET goes to
the VCC line of the project. Between gate
and source of the FET is 10K.

The gate of the FET also goes to the
collector of an NPN transistor (2N2222), with
the emitter of this transistor going to the
-V line. The base of this transistor goes
through 680K, then through 47uF to VCC and to
the FET drain (positive of the capacitor at
the drain). Between the base and emitter of
the transistor is a diode (1N914, 1N4148,
etc.), with its cathode on the base.

The junction of the capacitor and the
resistor goes through a normally open push-
button "on" switch to the plus battery lead

and to the source of the FET. A normally
open pushbutton "off" switch goes between the
base and the emitter of the 2N2222.

Refresher Circuit for the Above

A
medium-power PNP transistor is used (2N4036,
something which can handle a peak collector
current of nearly half an amp). Its emitter
goes to the VCC line and to the positive end
of the capacitor. Its collector goes through
22 ohms to the junction of the capacitor and
the resistor. Its base goes through 2.2K
to a source of downward pulses which, when
the project is dormant, stays high. (Such
downward pulses, for example, can be found
at the output of the 555 or 556 in the Smith-
Kettlewell Gimmick and/or the Fowle Gimmique.)

In the case of our VU meter, where the
output of the audio oscillator only goes up
to 5V and does not rest at VCC, the refresher
transistor base can be capacitively coupled
to the source of pulses. This base circuit
can be modified as follows:

The base of the PNP refresher transistor
goes through 10K to its emitter. This base
also goes through 1K, then through 0.47uF
(positive toward the resistor) to the output
of one VU channel.

N-Channel Timer for the Negative Supply
Line

The advantage of this version is that
the builder has so many VMOS/FET's to choose
from. The positive battery terminal goes to
the VCC line of the project. The negative
battery lead goes to the source of an N-
channel FET (VN0300M, VN10KM, etc.). The
drain of this FET goes to the project
"ground" (its -V line). The FET gate goes
through 10K to its source.

This FET gate also goes to the collector of
a PNP transistor (2N2907), with the emitter
of this transistor going to VCC and to the
positive side of the battery. The base of
this "latching transistor" goes through 680K,
then through 47uF to project ground and to
the drain of the FET (negative of the capacitor at ground). The junction of this
resistor and capacitor goes through a
normally open pushbutton "on" switch to the
source of the FET and to the negative battery
lead. Between the emitter and the base of
the above transistor is a diode, anode on the
base. Also between this base and emitter is
a normally open pushbutton "off" switch.

Refresher Circuit for the Above

A medium-
power NPN transistor is used (2N2219, good
for about 1/2 amp peak). Its emitter is
grounded, while its collector goes through
22 ohms to the junction of the capacitor and
the resistor. The base of this refresher
transistor goes through 2.2K to a source of
positive-going pulses in the project; when
dormant, these pulses should rest at logic
zero.

VMOS Power FET's

These are metal-oxide-semiconductor field-
effect transistors (MOS/FET's); the "V" in
their designation refers to geometry of the
channel and its interface with the gate.
(For a more complete discussion of their
entrails, refer to Siliconix Applications
Note AN76-3.) As opposed to junction FET's
(JFET's) which conduct until reverse bias on
their gate "pinches off" conductivity, these
MOS units operate in the "enhancement mode."
FET's of the enhancement-mode type are
normally cut off until a forward bias
(bringing the gate away from the source
and toward the drain) causes the channel to
conduct. Being devices of the "insulated-
gate" variety, the input resistance is
extremely high--it can be considered
infinite.

To a continuity tester, the gate--with
respect to either end of the channel--will
look open for either polarity. For whatever
reason, the channel looks like a diode to the
tester, the FET being used properly with this
"diode" back-biased. (The literature never
says whether this is an actual diode included
for protection of the device, or whether it
is the nature of the semiconductor device.
However, forward-current specs are given for
the diode; they are generally quite high.)
The cases and/or tabs of all these FET's are
common to their drain, giving you a good way
of finding the drain with a tester.

P- and N-channel units bearing the same
number (such as the VP0300 and the VN0300)
seem not to be "complementary" devices; such
specs as their gate-source threshold voltage
and their "on-resistance" differ markedly. I
have, therefore, shied away from using sister
devices to control a dual supply; they would
not turn off at the same time, etc.

Siliconix VP0300 P-Channel

(Note: The M suffix denotes a TO237 package; the B suffix denotes a TO39 package.)

General Parameters:

• Maximum drain-source voltage--30V
• Maximum drain-gate voltage--30V
• Maximum gate-source voltage--+40V
• Gate-source threshold voltage--minimum
  -2V, maximum -4.5V (measured with gate
  tied to drain and with 1mA drain
  current)
• Continuous drain current--0.48 amps for 0300M, 1.5 amps for 0300B
• Power dissipation (case temperature of
  25 degrees C)--1 watt for the 0300M,
  6.25 watts for the 0300B
• Drain-source on-resistance--2.25 ohms
  (with a drain-source voltage of -10V and
  a drain current of 1 amp)
• Forward transconductants--200 milliSiemens
  (with a drain-source voltage of -15V,
  and a drain current of 0.5 amp)

Drain-Source Diode:

• Forward on voltage--1.5V minimum at 0.5 amp

Siliconix VN0300 N-Channel

(Units with the M suffix are in a TO237
case; a D suffix denotes a TO220 package.)

General Parameters:

• Maximum drain-source voltage--30V
• Maximum gate-source voltage--+40V
• Continuous drain current--0.7 amp continuous (3 amps pulsed) for the 0300M;
  2.5 amps continuous (3 amps pulsed) for
  the 0300D
• Gate-source threshold voltage--0.8V
  minimum, 2.5V maximum (measured with
  gate tied to drain and with 1mA drain
  current.)
• Power dissipation (case temperature of
  25 degrees C)--1 watt for the 0300M,
  20 watts for the 0300D
• Drain-source on-resistance--3.3 ohms
  (with a gate-source voltage of 5V and a
  drain current of 0.3 amp), same for both
  case styles
• Forward transconductants--200 milliSiemens
  (with a drain-source voltage of 15V and
  a drain current of 0.5 amp.)

Drain-Source Diode:

• Forward on voltage--minus 0.9V at minus
  1 amp

Siliconix VN10KM N-Channel

(This device, which comes in a TO237
package, is an N-channel unit in which a
Zener diode has been included to protect the
gate.)

General Parameters:

• Maximum drain-source voltage--60V
• Gate-source voltage--+15V to -0.3V
  (contains Zener diode)
• Continuous drain current--0.3 amp (pulsed 1
  amp)
• Gate-source threshold voltage--0.8V
  minimum, 2.5V maximum (measured with
  gate tied to drain and with 1mA drain
  current)
• Power dissipation (case temperature of 25
  degrees C)--1 watt
• Drain-source on-resistance--5 ohms maximum
  (gate-source voltage of 10V and a drain
  current of 500mA)
• Forward transconductants--100 milliSiemens
  minimum (drain-source voltage of 15V and
  drain current of 400mA)

Drain-Source Diode:

• -0.85V at -1 amp

(No maxima are listed for the Zener currents, although they show a forward breakdown
voltage of 15V at 10uA.)

Pin Connections
for the Above Siliconix Devices

VN10KM, VN0300M, VP0300M (TO237):

With the pins facing up and the flat
side of the package toward you, the
three leads are, from left to right:
drain, gate, source.

VP0300B (TO39):

With the leads facing up and the tab to
the left, the three leads are, from left
to right: source, gate, drain.

VN0300D (TO220):

With the mounting surface toward you and
the pins facing up, the three leads are:
gate, drain, source.

(The above devices are available from
Hamilton Avnet, 1175 Bordaux, Sunnyvale, CA
94086.)

International Rectifier
IRFD-123 N-Channel
Radio Shack 276-2073

General Parameters:

• Drain-source voltage--60V
• Drain-source current--400mA
• Power dissipation--1 watt
• Drain-source on-resistance--3.2 ohms

Pin Connections:

This is a 4-pin DIP package, pins l and 2
of which are internally tied together; pins l
and 2 are the drain. Pin 3 is the gate, and
pin 4 is the source.

International Rectifier
IRF511 N-Channel
Radio Shack 276-2072

(Note: This device is no longer listed by

International Rectifier.)

General Parameters:

• Drain-source voltage--60V
• Drain current--3 amps
• Drain-source on-resistance--0.6 ohms

Pin connections:

This device is in a TO220 case. With the
mounting surface toward you and the pins
facing up, the three leads are, from left to
right: gate, drain, source.

SUMMARY OF SKTF SURVEY

By Lesley Brabyn

Surveys were sent out to 352 subscribers of
the SKTF in February of l983. So far, ll4
replies have been received and a preliminary
analysis of the first 85 has provided the
following information:

Twenty-nine subscribers report that they
have built a total of 55 devices described in
the SKTF. The most popular of these has been
the Auditory Gimmick, with l3 made, followed
by the Meter Reader and Sonalert Tester, each
with 8. Twelve of the 85 had commissioned
either a volunteer or paid contractor to make
a total of l9 additional SKTF devices, again
the most popular being the Auditory Gimmick,
with ll fabricated. Many other respondents
indicated that they are intending to build
projects in the future, but are presently
lacking the time to do so.

Fifty-four subscribers had built projects
not described in SKTF, but the majority of
these (74%) had found information contained
in the magazine to be helpful or encouraging
in the building of these outside devices.
Subscribers listed a total of 40 different
devices they would consider buying if they
were commercially available at a reasonable
price, the most popular being the V.U. Meter,
Audible Compass, Little-Go-Beep, and general
meter readers, respectively.

A majority of those replying found the
articles on soldering useful (84%) and comprehensible (95%). Sixty-seven percent
replied that they could now perform certain
tasks more easily than they could before
reading this material. Subscribers found
helpful and/or useful articles on "techniques
other than soldering" (point-to-point wiring,
wire wrapping, solderless breadboards - 82%),
tape splicing and cassette repair (62%), tape
editing (49%), component discussions (85%),
and instructional materials (66%). Most felt
that they could not have met their requirements for information of this type through
currently available sources other than SKTF.

Subscribers listed a total of 52 different
SKTF articles as being the most enjoyable to
read, with Soldering, Digital Electronics,
Speechboards, and Gabbing About Gates ranked
highest in popularity, respectively. Most
readers reported using the information
obtained through SKTF in connection with
their hobbies, but many also used it in their
work as well. An overwhelming 96% found SKTF
to be personally encouraging to them, and 92%
found it helped to expand their concept of
blind people in technical fields. Furthermore, 89% felt that, since reading SKTF, they
would be more confident in attempting a task
about which a counselor or teacher might say,
"It can't be done."

Most subscribers seemed to keep their
copies of SKTF to themselves, although some
shared them with a few others. A reader
living in Japan gave the magazine its widest
exposure by translating articles into
Japanese and publishing them in a braille
technical magazine there. Most readers had
"no idea" of what the maximum number of
people might be who could benefit from SKTF,
and guesses ranged from "a very few" to
"250,000." Both Talking Book and Braille
readers seemed generally pleased with the
quality of reproduction and formatting for
these versions. As such a small number of
Large Print readers responded to the survey
(only 2 out of the 85 currently analyzed), no
statement can really be made about this
version as yet.

Analysis of the remaining replies is
continuing, and an in-depth report will be
available soon.

[Editor's Note: Whereas most survey
responses are considered substantial if they
are 5% or so, you have broken all records in
submitting your entries. This Editor thanks
you profoundly; I feel as though you have
done me a great personal favor.]

EDITOR'S CRYSTAL BALL

If I can rouse myself from the reverie over
your kind words of support and the results of
the evaluation survey, it is time once again
to point the way toward new directions. (Not
that I have any illusions about fulfilling
all my past promises, but that never stops
me--making new ones keeps me on my toes.)

I want to make up some tape recordings of
lab activities which beginners could use as
step-by-step building instructions. The
first such tape will be a recording of three
people building identical continuity testers
--giving a blow-by-blow description of
gathering components, cutting the perforated
board, wiring the circuit, mounting the hardware, and testing the completed instrument.
Another tape will be a recording of me as I
build a more complicated IC project; this
will have to include some decision-making
about the size of the board, parts layout,

etc.

* * * * * *

Now that our Training Program is in full
swing and has proven itself, I plan to make
live recordings--strictly for encouragement--
of beginners making decisions about which
soldering techniques best address their
needs, etc. No such tape would be complete
without witnessing an experienced builder
(like me) try three or four times to make a
good solder connection, or without hearing
said builder do some damage. Although
edited, these latter recordings will show
processes being done in real time; the
listener will get an idea as to how long it
takes to make a solder connection and how

long it takes to measure out the location for
mounting a control.

Still another tape will contain examples of
Yours Truly working with an experienced
technical reader, going through a schematic
diagram and doing some basic math.

The first tape to be available is the tape
of building the continuity testers. Please
send me a 90-minute cassette, clearly marked
with your address, if you want this recorded.
The availability of the others will be
announced subsequently.

* * * * * *

I would like to compile information, perhaps in a series of articles, regarding the
training of technical readers for personal
use--getting people to describe schematic
diagrams, read equations, etc. A lot of
papers and seminars covering this subject
have been given at transcribers' conventions
(good reference material), but I have seen
nothing written from the viewpoint of
students, hobbyists and professionals working
on a one-to-one basis with their peers.
Those of you who have thoughts on this
subject, please pass your ideas along to me.

* * * * * *

The study on our prototype PC-board kits
has not been completed, though the preliminary results are encouraging. (Using
Thermoform models of the board to illustrate
various stages of completion seems viable.)
Eventually, if this concept is to become a
reality, there are some basic questions to be
answered:

If a manufacturer were to supply all the
components and accessories for a given
project (the cabinet, jacks, controls,
hardware, etc.), the eventual product would
be tremendously expensive; it would approach
the price of a ready-made instrument. (For
beginners without a "junkbox," and who are
not experienced in ordering components, this
would be the best approach.)

On the other hand, the kit could contain
only the circuit board and hard-to-get
components, thus making it available at a
fraction of the cost of the former. This
would mean, of course, that the eventual
product would be appropriate for more
experienced builders. I would like your
feedback as to which approach should be taken
initially; which would you buy if it were
available.

Our first prototype was of the basic "null-
type" meter reader ("Basic Analog Meter
Reader," SKTF, Spring l98l). This design
having been completed, this may be the first
kit produced. Upon completion of our study,
I would like to design a kit for the dog
whistle beacon ("Little-Go-Beep II," SKTF
Summer l982). It will be necessary to know
how many of each we should get made, and in
which form (the complete collection of parts,
or only the circuit board). Once again, I
would like to hear from you.

* * * * * *

Our subscribership is in need of another
boost. For whatever reason, our readership
is down from the previous year by 20%. We
are taking steps to get notices in various
magazines and into the hands of education and
rehab personnel, but it is my experience that
the most effective advertising is word-of-
mouth, as done by you. Please, help us grow
stronger and more secure by "talking us up"
within your organizations, and on the air.
(If a compatriot resists, twist his telegraph
key.)

To simplify our mailing lists and bookkeeping, we have been converting subscriptions to
coincide with the calendar year; eventually
everyone will renew his subscription in
January. Except for a few, your subscription
for the year of l984 is now due. I don't
want to lose anybody, so please keep this in
mind. (Those who need to renew will get a
reminder notice in January.)

* * * * * *

Your response to the survey, along with
what we have accomplished in the magazine
this year, catapults me into the New Year
with a light heart and good spirits. I wish
the same, combined with prosperity, for all
of you.