Smith-Kettlewell TECHNICAL FILE

A QUARTERLY TECHNICAL JOURNAL FOR BLIND AND VISUALLY IMPAIRED READERS

Sponsored by REHABILITATION ENGINEERING CENTER SMITH-KETTLEWELL INSTITUTE OF VISUAL SCIENCES

2318 Fillmore Street, San Francisco, CA 94115

SUMMER 1982, VOL. 3, NO. 3

TABLE OF CONTENTS

Introduction to Phase Lock Loops and the RCA CD4046

The Little Go Beep, I

CORRECTION - "Little Go Beep, II"

The Little Go Beep, II

The Fowle Gimmique

The K3VTA Auditory Gizmo

Lexicography

INTRODUCTION TO PHASE LOCK LOOPS AND THE RCA CD4046

by Al Alden and Bill Gerrey

Abstract

A theoretical discussion of principles will be given along with common uses, such as FM detection, frequency-to-voltage conversion, voltage-to-frequency conversion, tone decoding and frequency multiplication. Although detailed examples of circuits will not be given, a description of a specific phase lock loop chip, the CD4046, should be a sufficient springboard from which the experimenter can begin incorporating phase lock loops (PLL's) into projects. A complete exemplary circuit using the CD4046 can be found in the article, "The Little Go Beep, II," in this issue.

General Description

The three basic elements of a phase lock loop system are: a "VCO" (voltage-controlled oscillator), a "phase detector," and a "loop filter." These items are connected in a "loop" or closed system to be influenced or operated by an external source of frequencies.

The phase detector has two inputs, one for the external source of frequencies and the other for the PLL's VCO. The output of the phase detector is, in turn, used to control the VCO through a low-pass filter (this filter smoothes out the phase detector output so as to present a clean DC signal to the VCO).

In operation, the phase detector produces a voltage in accordance with the difference in phase between the external input signal and the VCO; the detector's output voltage "corrects" the VCO so as to bring it into step with the input signal. From this point on, a drop in frequency of the input signal causes the phase detector to pull the VCO input voltage down, thus lowering the VCO frequency to bring it into sync with the input frequency and phase. By the same token, a rise in the input frequency will cause the phase detector to bring the VCO up to this new frequency with a corresponding increase of the control voltage to the VCO.

With the "phase detector," a "loop filter" to smooth out its output, and a "voltage-controlled oscillator" all connected in a circle, we have a system that runs itself in accordance with an external oscillatory signal. What is the purpose of this exercise in electronic building blocks? To answer this question, we shall take a grand tour of the "loop" to see what information we can pick off the junctions between the building blocks:

  1. The VCO is called upon to replicate the frequency and phase of the input signal. However, by virtue of the loop filter (which we can adjust by our selection of parts), we can control the sensitivity of the system to jitter and noise on the input. The VCO can be made to give us a cleaned-up version of our input signal.
  2. In order for the VCO to be kept in step with the input, a DC control voltage is presented to it. The higher the frequency of our input signal, the higher will be the DC voltage controlling the VCO, and vice versa. We have created a frequency-to-voltage converter. The system will detect "frequency modulation" of its input signal; we extract the information from the varying voltage controlling the VCO. This voltage will dance back and forth between two discrete voltages it the input signal is frequency-shift keyed.
  3. By restricting the frequency range of the VCO, we can make the system very choosy as to the type of input signal it sees fit to lock on; we can arrange for the circuit to ignore everything but a certain input tone. When the desired input is not present, the VCO will free-run and the phase detector will complain about the disorder in the world nowadays. By looking at the phase detector's output, we can get information as to whether or not the VCO is locked onto a signal. For example, each telephone touchtone consists of a pair of frequencies. We could use a pair of PLL's, one for each tone, to detect when a given button has been pressed. Combining their phase detector outputs in an OR gate, we would have a device which would ignore everything but the touchtone it was designed to recognize. Speech, music, the patter of little feet - nothing would satisfy its conditions but the presence of the two necessary tones of that specific pushbutton.
  4. It is easy to divide the frequency of a signal using flip-flops, and you can buy "divide-by-N" counters which will afford division of their clock signal by any integer N. Techniques for multiplying signal frequencies are not so obvious. We can do so, however, by including dividers in the phase lock loop system. By running the VCO through dividers and presenting the slowed-down signal to the >phase detector, we can cause the VCO to run like Alice in Wonderland in order to make its divided frequency track with the input signal. If the divider between the VCO and the phase detector is a programmable divide-by-N unit, we can cause the VCO to run at a selection of multiples of the input frequency by setting the N of the divider. Thus, with a fixed external frequency being presented to the phase lock loop, we can "synthesize" multiples of this frequency, and the output of the VCO can be used to get "channels" whose distance apart is the PLL's input frequency.

In summary, we should note the diversity of unrelated uses to which a phase lock loop system can be put. It is conceivable that an FM transceiver could be built using: a PLL to synthesize the frequency channels, another PLL used as the FM detector, and the VCO portion of still another PLL chip being used as the modulator. Yet another PLL in the receiver audio system could be trained to listen for a specific audible tone on the transmitted signal, giving the radio equipment "selective call" capability. In every case, each PLL is being used in a very different way from its companions.

THE CD4046 PLL CHIP

This chip is discussed at length in "The CMOS Cookbook," by Don Lancaster (Chapter 7); much of the material of this article was gotten from there.

The RCA CD4046 has two discretely separate sections, the VCO and the phase detector circuits. - The loop filter is made up of two resistors and a capacitor which are connected as external components.

Phase detector

The 4046 has two phase detector schemes, giving you a choice as to which would be appropriate. They both have their inputs for the external signal on pin 14; their inputs for receiving the locally-generated VCO signal are on pin 3. Pin 14 can accept signals which are capacitively coupled. Pin 2 is the output of a" narrow-range conventional phase detector, while pin 13 is the output of a new CMOS wide-range phase detector. The features of both detector schemes are discussed below:

The narrowband phase detector is of the "exclusive OR" type. In order for it to work, the signals being compared must be similar. Since the VCO produces a square wave (50 percent duty cycle), the external signal must have a 50 percent duty cycle. The external input can be a sine wave with an amplitude of less than logic levels, since this detector amplifies the incoming frequency to make it into a square wave. This detector is harmonic sensitive; the wave going in must have a 50 percent duty cycle. Once the symmetry of the input signal has been established, the noise rejection of this detector is excellent, and is superior to its wideband counterpart. With this detector, the PLL can be used to "square up" or "clean up" an input signal.

The usable frequency range of the narrowband detector can only be stretched to about plus or minus 30 percent of the "center frequency."

The pulses on the output of the narrowband phase detector are averaged in the loop filter (the loop filter is being used as an integrator). Also it is important to note that this detector runs the VCO in the center of its range (the filter's output at 1/2 VCC) unless the input signal directs otherwise. In other words, this system will free-run at the center frequency.

As stated in the Cookbook, the VCO will be 90 degrees different in phase from the input signal at midband. This phase shift will vary for other frequencies throughout the range.

The wideband phase detector can accept input signals of any duty cycle and of either polarity (not to exceed the power supply voltage). However, the input signals must be at logic levels, and if they are to be capacitively coupled, a pull-up or pull-down resistor should be used to tie pin 14 (the input) to a supply line.

The wideband detector can be used over a frequency range of 1000-to-1. However, noise rejection is not as good with this detector, since it responds to such a wide variety of input pulses.

The output of the wideband detector is different from that of its narrowband counterparts. IIts tri-state output is such as to operate the loop filter as a "sample-and-hold" circuit (it pulls the capacitor to a voltage level and disengages, allowing this voltage to be "sampled" or seen by the VCO without influence until updated by the phase detector). If the input frequency is higher than the VCO, a steady high output from the detector results. If the input frequency is lower than the VCO, a steady low will result. When the frequencies are equal, the output produces pulses which are proportional to the difference in phase between the external input and the VCO (positive for lagging of the VCO and negative for leading by the VCO). When no input signal is present, the VCO rests at its lower extreme, which may be 0 Hz if the oscillator components are so chosen.

Using the wideband detector, there will be no difference in phase between the VCO and the input signal when the system is locked. This is true for all frequencies within the range.

Loop Filter

The output of the phase detector goes through R1, then through R2 in series with C1 to ground. The junction of R1 and R2 goes to the VCO input (pin 9).

The "settling time" of the system is set by R1 and C1. A large time constant will make the loop react sluggishly to changes in the input frequency. If the time constant is too short, the filter will not average enough cycles to present the VCO with a smoothed-out control signal, and the VCO will jump around violently.

Damping of the system is determined by the ratio of R1 to R2. A small R2 makes the loop bounce, overshoot, or even oscillate. An R2 which is large will make the loop take too long to readjust the oscillator to changes on the input.

The math associated with the design of the loop filter is reminiscent of that used in the design of other servo control systems - this is sadly so for your editor, who passed this course largely by accident. The Cookbook suggests that a design can be undertaken by experimenting with component values, starting with a couple of rules of thumb. Some suggestions are given as follows:

The settling time, set primarily by R1 and C1, should be much longer for the narrowband detector than for the wideband one, since the narrowband detector calls upon the filter to average out several pulses. On the other hand, the time constant of the loop filter should be short enough so as not to act as a low-pass filter to the expected changes in input frequency (unless you desire a low-pass filter as would be used to take jitter out of a noisy input signal). You have less to worry about with the wideband detector, since the loop filter is used as a sample and hold circuit and a comparatively short time constant can be tolerated. Obviously, when a divide-by-N counter is inserted between the VCO and the phase detector, the settling time should be chosen with respect to the phase detector's input and not to the VCO's output frequency.

The editor would start by calculating the largest settling time practical using his common sense (very common) in picking an RC time constant short enough so as not to interfere with the tasks the circuit is to perform. From this point I would reduce the value of R1 by 10 to 30 percent to make room for the damping resistor R2.

As far as damping is concerned, the Cookbook recommends picking an R2 of from one-tenth to one-third of R1. The effect of the damping resistor can then be assessed experimentally. An oscilloscope is an ideal instrument for measuring the response of a PLL system to various changes in excitation. However, until we all have one, we will have to be content with using alternative techniques - unfortunately these tricks are not universal to every design problem. For example, long time constants and over-damping would give us a poor frequency response in an FM detector. Overshoot n a tone decoder could be detected by listening to the VCO. Use your creative instincts.

Oscillator

The 4046 has an easy-to-use linear VCO (voltage-controlled oscillator) which the Cookbook suggests using where such a circuit is required. A capacitor and a resistor determine the center frequency and the range, while another resistor can be included to offset or shift this range. The voltage control pin (pin 9, the input) is very high impedance so as not to load the loop filter. An added bonus of the chip is that a voltage follower with its input also on pin 9 gives us a low-impedance version of the control voltage at pin 10, so that we can extract FM information without loading the loop filter. A resistor to ground is required from pin 10 (10K to 100K).

The oscillator's output is a clean square wave at full logic levels. Its output (pin 4) can be directly connected to the local oscillator inputs of the phase comparators (both inputs on pin 3), or the oscillator output can be fed into frequency dividers with their output going to pin 3. The oscillator has an inhibit pin (pin 5).

Grounding pin 5 enables the oscillator, while lifting pin 5 to a logic "1" stops it from operating.

With no offset resistor connected to pin 12, the oscillator frequency will go from 0 to twice its center frequency as the control pin is taken from 0 volts to VCC. As you would expect, its center frequency will be generated when the control pin is halfway between 0 volts and VCC, VCC/2 volts. A resistor on the offset pin will not change the oscillator's frequency range in Hz, but will shift this range up by a given percentage. For example, with an oscillator capacitor (C1) of 0.1uF and a resistor (R1) of 10K ohms, the oscillator's center frequency will be 1kHz. With no offset resistor (R2), the oscillator will cover the range 0 to 2kHz. Including an offset resistor of R2 equals R1 shifts the oscillator frequency so that it covers from 2kHz to 4kHz (as the input goes from 0 to VCC). A table of values is presented at the end of this section.

C1 is connected between pins 6 and 7. R1 goes from pin 11 to ground, and R2 goes from pin 12 to ground. Remember to ground the VCO's inhibit pin, pin 5. The VCO's output will appear at pin 4, with pin 9 being the control voltage input.

TABLE OF FREQUENCY-DETERMINING COMPONENTS

(The frequencies given represent the center frequency with the control pin at 1/2 VCC and with no offset resistor.)

TABLE OF OFFSET RESISTORS

(Inclusion of this resistor will shift the entire VCO range upward. The ratios given represent the newly defined range of the oscillator, where f is the original center frequency.)

Lock-Detection Pin (pin 1)

This pin can be used to generate a signal which will indicate when the PLL is locked onto an input frequency. Not enough is said about this pin in the literature to satisfy your editor; however, a sample circuit is given. When locked, pin 1 goes high except for very short pulses; the narrowband phase detector output, pin 2, apparently generates a signal which is complementary.

The sample circuit shows pin 1 and pin 2, going into a 4001 NOR gate. The output of the gate goes through 100K, then through .00SuF to ground, with a diode connected across the resistor (anode at the 4001). The junction of the resistor and the capacitor goes to both inputs of a second 4001. The output of the second 4001 goes high when the PLL is locked.

The exemplary use of the above circuit is a monopolar FSK detector; the PLL is sharply tuned to the "mark" frequency and the "space" frequency is left to fall outside the range. (Perhaps two such circuits working into the inputs of an RS flip-flop would be a dynamite FSK detector for both mark and space.)

EXAMPLE CIRCUIT

The following circuit is given in the Cookbook as a means of "counting" very low frequencies. (Using a frequency counter under 10Hz takes a little time, and can better be accomplished by hooking a counter onto a PLL multiplier, hence this sample circuit.)

Circuit - Pin 8 of the 4046 is grounded, while pin 16 goes to VCC. Pin 5, the inhibit pin, is grounded.

The external signal, intended to go from 1 to 100Hz, goes into pin 14. The VCO, which is intended to go from 100 to 10,000Hz (thus multiplying the input by 100) has its output going through a divide-by-100 counter to pin 3, the other input of the phase detector. (The oscillator's output is pin 4.) A timing capacitor of .01uF goes between pins 6 and 7, a 6.8K resistor goes from pin 11 to ground, and another 6.8K resistor offsets the oscillator by going from pin 12 to ground.

Pin 13, the output of the phase detector, goes through the loop filter to ground; pin 13 goes through 1meg, then through 100K, then through a high-quality tantalum 2.2uF capacitor to ground (negative of cap at ground). The junction of the two resistors goes to pin 9, the VCO control pin.

In this application, the output is taken from the output of the oscillator, pin 4.

LITTLE GO BEEP, I.

Abstract--This device produces a recurrent beeping sound when turned on; it can then be placed in a location of which the user wishes to keep track. It can serve as an orientation aid for target games, swimming, or for marking a danger zone to be avoided.

Several versions of this device have been built, having various power output requirements. With this article, you can pick a design that suits your purpose.

The size of the resulting device must of course be up to you, although all are battery-operable. High-powered units require a substantial battery supply; we have used as many as six 9V batteries in series. Large loudspeakers are more efficient than smaller ones, giving you another parameter to consider.

The basic beep generating scheme uses a dual timer, such as the NE556. (In our more sophisticated beacons, we have used the Intersil ICM7556 CMOS timer chip to cut down on battery drain between beeps.)

The first half of the 556 is a slow-running clock which determines the rate and duration of the beep. This first half turns off a transistor which otherwise inhibits the oscillator in the second half, the beep oscillator.

Using a bi-polar (nonCMOS) 556, the above arrangement can drive a speaker directly. However, if more volume is desired, a 2-transistor amplifier can be added, driving the speaker directly as the collector load of the final power transistor. If still more volume is desired, up to 6 9V alkaline batteries can be put in series as the collector supply for this power transistor; you can keep your speaker's voice coil nice and warm with this much "English" on it.

Basic Circuit (the Littlest Go Beep)

Pin 7 of the 556 is grounded, while Pin 14 goes to VCC (9V supplied by an alkaline battery). Both "enable" Pins, 4 and 10, go to plus 9V.

Pins 2 and 6 of the first half are tied together and go through 5uF to ground (negative at ground). Pins 2 and 6 also go through 4.7K, then through an optional 20K "duration control" (connected as a rheostat) to Pin 1. Pin 1 also goes through 100K, then through a 1meg rheostat (beep rate control) to plus 9V. The output, pin 5, goes through 10K to the base of a 2N2222 transistor. The emitter of the 2222 is grounded, while its collector goes to pin 13 (the discharge terminal) of the second half of the 556.

Pins 8 and 12 of the 556 are tied together and go through .005uF to ground. Pins 8 and 12 also go through 47K to pin 13. Pin 13 goes through 82K, then through a 1 or 2 megOhm rheostat "pitch control" to plus 9V. The output, pin 9, goes through 47 ohms, then through the speaker to plus 9V.

Rather than run the speaker directly off the 556 (which can only source about 200mA), the speaker can become the collector load of a power transistor which is capable of applying nearly the whole 9V supply to the voice coil. With this arrangement, the 556 is not called upon to source very much current, so we can replace it with its CMOS equivalent (ICM7556); this reduces the battery drain from 7 to about 1mA (between beeps).

You may run into the need for profound decoupling of the 556 supply from that of the power amplifier. For this reason, I recommend that a separate battery be used for each.

Circuit for the Add-On Amplifier (Beefy Go Beep)

Although not absolutely necessary, the NE556 is replaced with its CMOS counterpart, the Intersil ICM7556. Being "pin-compatible", no circuit changes (except for the output resistor) are necessary.

The output, Pin 9, goes through 2.2K (instead of 47 ohms) to the base of a 2N4036 PNP transistor, with this base also going through 10K to Pin 14 of the 556 and to its plus 9V supply (either alkaline or zinc-carbon battery can be used). The emitter of the 4036 goes to this plus 9V line, while its collector goes through 100 ohms to the base of a 2N3055 power transistor.

The base of the 3055 also goes through 1K to ground. Its emitter goes through 1 ohm to ground, while its collector goes through the speaker to plus 9, 18, 27, 36, 45 or 54 volts.

The negative side of both the 556's battery and the amplifier's battery is grounded. The positive side of the 556's battery goes through one section of a DPST switch to Pin 14 and its plus 9V line. The positive side of the amplifier's battery (from 1 to 6 9V alkaline units in series) goes through the other section of the DPST switch to the speaker (the other side of the speaker going to the collector of the 3055 as mentioned).

It may be desirable to have a volume control on your unit. There are several ways to accomplish this, two of which are discussed below:

In amplified units using more than 9V, you can use a single-pole rotary switch to connect the speaker to "taps" on the battery system--between the various 9 volt units in the string. For example, positions on the switch could be used to select one, two, four or six batteries.

The speaker could be connected through a volume control, such as a "rear speaker" unit used with auto radios. We had some 150 ohm linear powerstats that worked okay. In either case, the bottom of the pot goes to the positive battery supply and to one side of the speaker. The top of the pot (the clockwise end) goes to the collector of the output transistor, or to the Pin 9 output of the NE556 (depending upon which circuit was constructed). The arm of the pot goes to the other side of the speaker.

The power output of this device is worth discussing. As we know, the power goes up as the square of the voltage, and a string of batteries will get you up to some theoretically high values. In addition, the power absorbed by the load depends heavily on the frequency (the pitch) of the beep, since the "on" portion of the oscillator's duty cycle is fixed (determined by the 47K resistor going from Pin 13 to Pins 8 and 12).

With the oscillator set at its highest frequency (Pins 8 and 12 going through 82K to plus 9V), and with a supply of 36V, my little 2-inch test speaker ran warm, but survived. However, when I let the oscillator run long enough to measure the current drain (averaging over half an amp), the lab got real quiet.

Parts List - Basic Circuit

Capacitors:

Resistors (1/4W, 5%):

Potentiometers (either panel-mount, fitted with knobs, or PC mount, trimmers):

Semiconductors:

Parts List - Amplifier

Resistors (5%):

Semiconductors:

Switch:

CORRECTION - "Little Go Beep, II" (SKTF, Summer 1982)

In the first mike preamp using a TLO72 (top of page 33), we find, "pin 3 also goes through 4.7K to ground." No, No! In building this preamp, we discovered that its bias must be VREF1, just like pin 2. Therefore, "pin 3 should go through 4.7K to VREF1."

Sorry, a whole battery supply of engineers missed that one.

THE LITTLE GO BEEP, II by Albert Alden

Abstract - This article describes a sound beacon which is triggered by a one-second blast of a 13kHz dog whistle. The beacon itself is of pocket-size, draws less than 15mA from a 9V battery in the standby condition, and can be actuated from a distance of nearly 100 feet if its microphone is facing the whistler. The heart of the system is a phase lock loop (PLL), the theoretical basis of which is discussed in "Introduction to Phase Lock Loops" in this issue.

(Editor's Comments: Vividly, I remember the situation surrounding my discovery in the BTP of Jim Swail's dog whistle beacon (BTP 1965). Instead of attending in class to history and literature professors, I would read BTP under the desk on my lap. The dog whistle article caused enough non sequitur elation to get me caught.

Because Swail's version predated integrated circuits, it had a highly sophisticated audio filter using several transistors; the unit was necessarily large and complicated. I have been keeping my ear to the ground in search of the ideal IC active filter in order to duplicate Swail's work, hoping to generate a circuit that is smaller and easier to build. My able colleague, the author, in his typical flurry of creativity, took a new approach to the design by using a phase lock loop as a "tone decoder" to detect the whistle.

What's it good for? Put it on your seat in a bus, plane, or other public seating area, and experience the freedom of being able to leave and come back again without counting the rows and aisles. Leave it with your towel on the beach so that you can stray a few yards without fear of losing your belongings. Mount an AC-powered version on a hard-to-find door (such as in row housing) so that the desired portal will answer your call. I plan to build one into my suitcase; when I get to the baggage carousel, I'll whistle it up. (At this stage, I cannot guarantee that the device will not false trigger from high-frequency noise in the baggage compartment, thus draining the battery before arrival.

Science for the Blind (SFB Products, P.O. Box 385, Wayne, PA 19004) makes a very fine RF-triggered beacon whose range is 300 feet (this range can be extended to 1000 feet by improving the receiving antenna and by adding a large loudspeaker). However, the sophistication of this system is such that the device is less portable than our "Little Go Beep.")

Recognizing that there is a need for a portable sound beacon which could orient the user as to its location with a buzzing sound when addressed, the development of the "Little Go Beep" was undertaken. Because of its small size and ruggedness, the so-called "silent dog whistle" was chosen as the instrument by which the user activates the beacon's buzzing sound. (It is the editor's opinion that the "silent dog whistle" was so named because its inventor went deaf in its development. However, I was the one who insisted that this instrument be used for the reasons above.)

The following criteria were satisfied by the eventual design:

Circuit Operation

The signal from an electret condenser microphone is amplified in two stages using a low-noise dual op-amp, the Texas Instrument TLO72. After being amplified, the audio signal is squared off with a comparator whose reference is a forward-biased germanium diode (this comparator being a CA3130 op-amp). When the amplified signal exceeds the threshold of the diode, the comparator output jumps from 9 to 0 volts, thus presenting the PLL with a squared-up signal at full logic levels. This signal is then capacitively coupled to the PLL's phase detector input. (Note that in the article, "Into PLL's," the CMOS Cookbook recommends that a pull-up resistor be used with this input if the signal is capacitively coupled. It is the editor's guess that since this circuit is "listening" for a prolonged steady-state 13kHz signal, the floating of the input in standby is harmless; the resistor can go to blazes.)

The RCA CD4046 chip was chosen for two reasons. First, the VCO input voltage and free-running frequency for no input signal are at their lowest (as opposed to resting at the center of their range for most units). Thus the dog whistle, whose pitch is very imprecise, will pull the PLL up to a window of acceptable frequencies. This change is easier to detect than the change we could expect by pulling the VCO to one side of its center frequency which is already in the window. Second, once the input signal has been brought to logic levels, the wide-range detector of the 4046 is less harmonic sensitive than other more conventional PLL designs.

A window of acceptable frequencies is selected by comparing the VCO control voltage with two fixed voltage levels. These fixed levels are gotten from a voltage divider off a 5.2V zener in the PLL chip, and the comparisons with these voltages are made by another TLO72; one half is the "lower" comparator, and the other half is the "upper" comparator. Both comparators will be down when a signal appears in the window between these two fixed voltage levels. One "OR" the other of these comparators will go to logic "1" if the signal is outside the window.

The outputs of the comparators are OR'ed with diodes. Until a signal appears in the window, one OR other of the comparators will forward bias its diode, and a logic high will be maintained on the cathodes of the diodes. A parallel RC combination on the output of the OR gate (the two cathodes) will retain enough charge for the instrument to ignore false short-term signals which land in the window. Both comparator outputs must be low for almost a second for the parallel RC circuit to bleed off and trigger the beacon.

A CMOS 555 timer (Intersil ICM7555) is set up as a one-shot to be triggered by the descending charge on the parallel RC circuit. The time of this one-shot determines the length of the buzzing sound. The 555 triggers a Peizo-electric buzzer.

Circuit

The microphone, a Radio Shack 270-092A, has its shield lead grounded. Its power lead (red) is decoupled from the power supply; it goes through 330 ohms to plus 9V, and it also goes through 1uF to ground (negative end of this bypass capacitor at ground).

The hot signal lead from the mike (the white center conductor of the shielded cable) goes through .0047uF to pin 3, a non-inverting input of a TLO72 dual op-amp. Pin 3 also goes through 4.7K to VREF1.

This first op-amp has a feedback resistor of 510K from output to inverting input (pin 1 to pin 2), and this resistor is shunted by 10pF to prevent oscillation and to suppress noise. The inverting input, pin 2, also goes through 10K to a bias point (VREF1) to be described later.

The output of the first stage, pin 1, goes through .0047uF, then through 4.7K to the inverting input of the second half of the chip (pin 6). The non-inverting input of this second stage, pin 5, goes to VREF1. The feedback on this stage, from pin 7 to pin 6, is also 510K in parallel with 10pF.

Pin 4 of the TLO72 is grounded, while pin 8 goes to plus 9V. VREF1 is gotten from a voltage divider across the power supply; the plus 9V line goes through two 51K resistors in series to ground, with the bottom resistor being shunted by 4.7uF (negative at ground). The junction of the two resistors and the capacitor is VREF1. As stated, this point goes through 10K to pin 2, as well as going directly to pin 5 of the 072.

A CA3130 is used open-loop to square-up the amplified mike signal. Pin 4 of the 3130 is grounded, while pin 7 goes to plus 9V. Pin 7 of the TLO72 goes through .0047uF, then through 8.2K to ground. The junction of this capacitor and resistor goes through 91K (to protect the 3130 from negative input swings) to the inverting input, pin 2, of the 3130. As a voltage reference (VREF2), a forward-biased germanium diode is used. The cathode of the diode is grounded, while its anode goes through 51K to plus 9V. Pin 3 of the 3130, the non-inverting input, goes to the diode's anode.

The output of the comparator, pin 6 of the 3130, goes through .001uF to pin 14 of the CD4046 PLL. Pins 1 and 2 of the PLL are not used. Pin 8 is grounded, while pin 16 goes to plus 9V. Pins 3 and 4 are tied together (this connects the VCO to its phase detector input). Pin 5, an inhibit pin of the VCO, is grounded.

A VCO timing capacitor of .003uF goes between pins 6 and 7. Also for the oscillator, pin 11 goes through 100K to ground. To offset the VCO, pin 12 goes through 51K, then through a 100K rheostat to ground.

Pin 13, the output of the phase detector, goes through 100K, then through 4.7K, then through .1uF to ground (this is the loop filter). The junction of the two resistors goes to the VCO input, pin 9. The buffered VCO input voltage pin 10 has a 51K 1oad resistor to ground.

As yet another reference (VREF3) is a 5.2V zener in the chip with its anode on pin 8 (which is ground) and its cathode on pin 15. Pin 15 goes through 3.9K to p1us 9V.

The two ha1ves of another TLO72 comprise the window comparators. Pin 4 of this chip is grounded, whi1e pin 8 goes to p1us 9V. Pin 10 of the 4046 goes to the inverting input, pin 2, of the "1ower" comparator; pin 10 a1so goes to the non-inverting input, pin 5, of the "upper" comparator. The 5.2V output of the zener, pin 15 of the 4046, goes through three resistors in series to ground - first 180K, then 430K, then 470K to ground. The junction of the 430K and the 470K goes to the non-inverting input of the "lower" comparator, pin 3 of the O72. The junction of the 180K and 430K goes to the inverting input, pin 6, of the "upper" comparator. The outputs of the comparators are OR?ed with diodes; each output (pin 1 and pin 7) goes to the anode of a 1N4148 diode, with the cathodes being tied together and going through the parallel combination of .22uF and 1megohm to ground.

The top of the parallel RC combination goes to pin 2 of an ICM7555 (CMOS 555). Pins 4 and 8 go to plus 9V, while pin 1 is grounded. There is a 1uF timing capacitor going from pin 7 of the 555 to ground (negative at ground). Pins 6 and 7 are tied together and go through 1.5 meg to plus 9V.

The 555's output, pin 3, goes to the control pin, pin 8, of a Star Micronix CMB-12 buzzer. Pin 1 of the buzzer is grounded, while pin 14 goes to plus 9V. (Though these buzzers only have four pins, they are in the corner positions of a 14-pin DIP IC. The speaker holes are closest to pins 7 and 8.)

The 9V line is bypassed by 4.7uF near the PLL (negative at ground). The negative side of the battery is grounded, while the positive side of the battery goes through an on-off switch to the plus 9V line. This completes the circuit.

The 100K rheostat associated with pin 12 should be adjusted so that a 13kHz input tone puts the voltage of pin 10 of the PLL right between the voltages on the two comparators, that is, mid-way between the voltages found at the ends of the 430K resistor in the divider string. For this tone, the dog whistle will be about 1 1/2 turns out from its shortest extreme.

Parts List

Resistors fixed 1/4W 5%

Resistors adjustable

Capacitors

Diodes

Integrated Circuits

Miscellaneous

*Star Micronics, Inc.
200 Park Avenue, Suite 2308
New York, NY 10017
(212) 986-6770

THE FOWLE GIMMIQUE

by Tom Fowle, WA6IVG

Abstract--a "dynamic meter reader" which can indicate relative change as well as quantized information about its input is described. Being a fancified version of the Smith-Kettlewell Gimmick, the reader may wish to review the theory of operation of the original unit (SKTF, Spring, 1981, Page 12).

Over years of work with various auditory meter readers (from those with mechanical choppers to those designed at Smith-Kettlewell, including talking instruments) it has come to my notice that none of the units I have used give a good indication of both dynamic and static meter activity. The most obvious situation in which both indications are needed is in tuning ham transmitters for a "dip" or a "peak" in meter current while noting whether a desired reading has been reached or exceeded. All of this should happen without having to touch the meter reader.

A device of this sort is not limited in usefulness to transmitter tuning. In trouble-shooting circuits, for instance, it is delightful to have a VOM which tells you immediately if it is seeing about 10%, 50% or 100% of the full-scale range. The circuit to be described does a better job of this than other schemes do.

The Fowle Gimmique is based on the Smith-Kettlewell auditory Gimmick circuit described in the Spring, 1981 issue. Like the original, the Gimmique has a VCO whose pitch varies relative to the voltage drop across the meter being read. However, the design of the Gimmique includes a calibrated voltage standard; if this Braille dial has been set to read half scale, a reading which exceeds that scale will cause the VCO to beep (turn on and off) at a rate of perhaps 4 to 6 times a second. The duty cycle of the oscillator creating these pulsations is such that the VCO is on most of the time, thus there is no difficulty in keeping track of pitch changes of the VCO. You still know where the meter is going, and you also know that the reading has passed your pre-set value.

Of course, at any time you can grab the knob of the calibrated standard and adjust it so as to take an exact reading of meter position. You will soon learn the relative pitch of the VCO so that you can guess where the knob has to be adjusted to get a reading quickly. There is no need, as with the original Gimmick, to keep switching between the meter and the calibrated standard. Enough bragging, let's see how it works.

Theory of Operation

There are no conceptual changes in the original circuit, except that a comparator and a slow-running "beep" oscillator have been added. A dual op-amp (LM358) in place of the original CA3130 gives us a spare op-amp to serve as a comparator. A dual timer (NE556) in place of the original 555 gives us a spare section with which to pulse the VCO.

The second half of the dual op-amp, the comparator, has its non-inverting input connected to the meter, while its inverting input goes to the arm of the calibrated pot. The output of this comparator drives the charge resistor of the "beep" oscillator (using the second half of the 556 dual timer) This slow-running "beep" oscillator (operating at about 4 to 6 Hz) drives the emitters of the two 2N 2907 current mirror transistors that control the original VCO (the VCO now being the first half of the 556).

When the voltage across the meter higher than that of the Braille-calibrated dial, the comparator output slams high and the "beep" oscillator starts running, chopping the current source to the VCO section of the 556. A meter reading which is below the setting of the dial will cause the comparator output to go low, thus disabling the slow-running "beep" oscillator whose output will then pull the emitters of the current mirror to VCC. An interesting embellishment is to include a DPDT switch on the comparator's inputs to change polarity or swap its input signals; the instrument will then beep for an input below the setting of the dial and hold steady for a reading which exceeds the voltage of the calibrated pot.

The LM358 dual op-amp does not have offset capability as did the CA3130 in the original circuit. I therefore included a bias circuit for the VCO half of the 556; the trigger and threshold pins are biased through a 1meg resistor supplied by a trim pot going between plus V and ground. This trim pot can then be adjusted to bring the VCO to zero when the input terminals are shorted.

The two models of the Gimmique built so far have exhibited an accuracy which is limited only by the quality of the calibrated standard. Given a pot with good linearity, the Gimmique can be used to generate its own Braille dial. The input signal can either be a voltage source being monitored by a meter of known accuracy, or it can be the meter with which the Gimmique is to be used. In either case, the 1.5 meg calibration rheostat in the Gimmick should be adjusted so that a full-scale input condition causes the "beep" oscillator to take off with the precision linear pot set at maximum. Then adjust this pot and mark the Braille dial for every significant input condition.

Circuit for the Fowle Gimmique

The input terminals, which are shunted by 0.05uF, each go through separate RF chokes (2.5mH), with the far ends of these chokes also being shunted by 0.05uF. The far end of the cold input choke goes to circuit ground. The far end of the hot input choke goes through 220K to Pin 3 (the non-inverting input) of the LM358j the, far end of this choke also goes through 100K to Pin 5 (the non-inverting comparator input) of the LM358. Pin 8 of the 358 goes to plus 9V. Pin 4 goes to the negative of the 9V battery (this is not ground), with Pin 4 also going through a diode to ground (anode at ground). (This diode, a 1N4001, assures that the op-amp will operate at ground by putting the negative supply pin 0.6V below ground.)

Pin 1, the output of the first half of the 358, goes to the base of a 2N2222; the emitter goes to Pin 2, the inverting input. Pin 2 and the emitter go through 1K, then through a 25K rheostat to ground. The collector of the 2222 goes to both the bass and the collector of a 2N2907, the first half of the "current mirror". The collector and base of this 2907 goes to the base of another 2N2907, the second half of the "current mirror". The 2907 emitters are tied together and go to Pin 9 of an NE556--this is the output of the "beep" oscillator.

The remaining collector of the second 2907 goes through 22K to Pins 2 and 6 of the 556, which are tied together. Pins 2 and 6 also go through 0.0047uF to ground, and also through 10K to Pin 1. For offset, Pins 2 and 6 also go through 1meg to the arm of a 1meg pot; the top of this pot goes to plus 9V and the bottom is grounded. Pins 4, 10 and 14 are tied together and go through 10 ohms to plus 9V; Pin 7 is grounded. Pin 14 is by-passed to ground through 100uF (negative at ground). Pin 5, the output, goes through 47 ohms to one side of the speaker, with the other side of the speaker going to Pins 4, 10 and 14 of the 556.

Pins 8 and 12 of the 556 are tied together and go through 1 or 2 uF to ground (negative at ground). (This capacitor sets the beep rate.) Pins 8 and 12 also go through 10K to Pin 13, as well as going through 22K to Pin 7 of the LM358 (the comparator output).

Pin 6 of the LM358, the inverting input of the comparator, goes through 100K to the arm of the Braille-calibrated pot (10K linear precision). The bottom of this pot is grounded, while its top goes through a 1.5 meg rheostat, then through 100K to the cathode of a 6.8V Zener. The Zener's anode is grounded, while its cathode also goes through 470 ohms to plus 9V.

To review, the 1.5 meg rheostat in series with the 10K precision pot is for calibration, the 25K rheostat in the emitter of the 2N2222 is the sensitivity adjustment, and the 1 meg pot associated with the 556 is for offset.

In high RF environments or where hum needs be filtered out of the input signal, the non-inverting inputs of the 358 (Pins 3 and 5) can be by-passed to ground with 0.1uF discs.

Remember, like all meter readers, this circuit should float above chassis ground, and depending on the application (where high voltages are present which are not referenced to ground), the degree of insulation may have to be substantial. Items to be considered include input terminals, speaker! switches and controls, although the insulation inherent in the construction of these components will be sufficient for use around voltages near ground.

PARTS LIST

Capacitors:

RF Chokes:

Resistors (1/4 watt 5%):

Potentiometers:

SemiConductors:

CORRECTION - "The Fowle Gimmique" (SKTF, Summer 1982)

Jim Swail points out that an offset voltage in some LM358's can introduce an error at the triggering point of the "beep oscillator" -- his unit was triggering low by about 4mV. (Data sheets on the 358 list offsets of 2mV nominal and 7mV max.) There are four possible solutions:

  1. The LM358 can be directly replaced with a high-grade op-amp, National LF442ACN, whose offset is 0.5mV nominal and 2mV max. (Available for $6.10 from Hamilton Avnet, 1175 Bordaux, Sunnyvale, CA 94086.)
  2. For the same money, you can buy several 358's so as to hand-pick one with a low offset.
  3. An offset adjustment can be built into the instrument by connecting the bottom end of the calibrated pot through a 500-ohm rheostat to the negative end of the battery, not to ground. The calibration and offset adjustments will interact, however.
  4. One could rebuild the device with two single op-amps having provisions for offset pots (See the 100K pot in the original SK Gimmick, Spring 1981.) With this arrangement, the 1 meg offset pot in the present Fowle Gimmique's oscillator can be eliminated.

THE K3VTA AUDITORY GIZMO

(Taken from BTP, April 1972)

Abstract - The material for this article was sent along to me (Bob Gunderson, Editor) by Gale Conard, K3VTA/4, who put it together from ideas acquired from the article, "The Simplest Keying Monitor," in the October 1971 issue of BTP. Gale has come up with a design which fits neatly into a shirt pocket. It functions as a sensitive field strength indicator, a reliable and pleasant-sounding keying monitor, a handy continuity checker, and a code practice oscillator.

This circuit uses a unijunction transistor (UJT), 2N1671 in a relaxation oscillator configuration. The completed unit contains two closed-circuit phone jacks, J1 and J2. Inserting a plug into J1 opens the negative battery lead feeding the relaxation oscillator so that a telegraph key may be added, or a plug with a pair of test leads may be plugged into this jack to turn the device into a continuity checker. Inserting a phone plug into J2 disconnects the built-in 8-ohm speaker and connects the earphones to the gizmo (low impedance phones work best). The antenna lead for RF pickup may be a piece of flexible hookup wire coming from the instrument, or a small insulating post may be provided for connecting this length of wire.

RF picked up by the antenna is rectified by a 1N34A diode, and the resultant DC signal is presented to the emitter of the UJT.

The circuit can easily be assembled on a piece of vector board (perforated board) and fastened to the inside of an aluminum utility box by any conventional means. This box, which measures 2-1/4 by 1-1/4 by 4-1/4 inches, also houses the speaker and the two jacks, J1 and J2.

Circuit

The pickup antenna goes through an RF choke to ground (RFC1, 2.5mH). The antenna also goes to the anode of a 1N34A diode. The cathode of this diode is bypassed to ground by C3 (.01uF). This cathode also goes to the arm of section 1 of the double-pole double-throw switch (toggle), and also to position No. 1 of section 2 on this switch. The positive of the 9-volt battery goes to position No. 2 of switch section 2, and the negative of this 9-volt battery goes to the tip terminal of key jack J1; the switch and sleeve terminals of J1 are connected together and grounded.

Position No. 1 of section 1 on the toggle switch is unused. Position No. 2 of section 1 goes to the emitter of the unijunction transistor (2N1671). The arm of switch section 2 goes to the base No. 2 of the UJT. The emitter also goes through 0.3uF (made up of three 0.1uF capacitors in parallel), with the other side of this capacitor, C1, grounded. The emitter also goes through 0.3uF (made up of three 0.1uF capacitors in parallel), with the other side of this capacitor, C1, grounded. The emitter also goes through a resistor R1 (5.1 kOhrns) to base No.2 of the UJT, and this No. 2 base also goes through capacitor C2 (0.05uF) to ground.

Base No.1 goes to the tip of the closed-circuit jack J2. The J2 switch contact goes to the hot side of the speaker voice coil, with the other side of the speaker connected to the J2 sleeve and to circuit ground. The impedance of the speaker is 8 ohms, although almost any low-impedance p.m. speaker should perform satisfactorily.

Parts List

Speaker: 8 ohms.

LEXICOGRAPHY

The following books appear in the Catalogue of Recording for the Blind, 215 East 58th Street, New York, NY 10022.

Adams, J., Mac and Robert Noon:

Alerich, Walter N.:

Blumenthal, Philip:

Bohn, Ralph C. and Angus J. MacDonald:

Buban, Peter and Marshall L. Schmidt:

Darr, Jack:

Davenport, Donald A.:

DeFrance, Joseph J.:

Delpit, George H. and B. Steven Johnson:

Doyle, John:

Eberlin, Fred E.:

Geiger, Darrell L.:

Geiger, Darrell L. & Bernard D. Ross:

Gerrish, Howard H.:

Gerrish, Walter C.:

Gillie, Angelo C.:

Gustincic, Jacob J.:

Jacobsen, Frederick O.:

Lease, Alfred A.:

Lemons, Wayne:

Lemons, Wayne & Bill Price:

Miller, Rex & Fred W. Pepper. Jr.:

Mix, Floyd M.:

Molodivitch, Joseph:

Montgomery, Glen:

Morgan, Alfred:

Oleksy, Jerome E.:

Pirrone, Philip A.:

Prenke, Edward M.:

Ross, Frank, Jr.:

Ruchlis, Hy:

Sams, Howard W. & Co.:

Shields, John Potter:

Slurzberg, Maurice et al:

Smith, Carl E.:

Stephenson, George :

Training and Retraining, Inc.:

Ulmer, Robert M.:

Uzunoglu, Basil:

Wilson, Jack:

Wood, Merle W.:

Worthington, Robert M., et al:

Xerox Education Publications:

The following are listed under the heading "Engineering".

Abbott, Waldo & Richard L. Ryder:

Adams, James E.:

Allied Radio Corporation:

American Radio Relay League, Inc.:

Anderson, Charles J. et al:

Belke, R. E. et al:

Belove, Charles et al:

Blake LeMont V.:

Blitzer, Richard:

Boylestad, Robert L.:

Boylestad, Robert & Louis Nashelsky:

Brite, Robert J.:

Brophy, James J.:

Brown, Robert M. & Tom Kneitel:

Buck Engineering Company, Inc.:

Buckwalter, Len:

Bureau of Naval Personnel:

Burke, W. E., et al, Editors:

Carr, Joseph S.:

Castellucis, Richard L.:

Churchman, Lee W.:

Clement, Preston R. & Walter C. Johnson:

Cleveland Institute of Electronics:

Cooper, William David:

Deem, Bill R. et al:

De France, J. J.:

Del Toro, Vincent:

Dorf, Richard C.:

Electrical Wiring Job Sheet:

Energy Concepts, Inc.:

Feldman, Leonard:

FitzGerald, A. E., et al:

Flynn, George:

Gerrish, Howard H.:

Ghirardi, Alfred A. & Jeff E. Dines:

Gillie, Angelo C.:

Graham, Kennard C.:

Grammer, George:

Grob, Bernard:

Grob, Bernard & Milton S. Kiver:

Ham Radio Publishing Group:

Harger, Robert 0.:

Hayt, William H., Jr.:

Hayt, William H., Jr. & Jack E. Kemmerly:

Healey, Timothy J.:

Herrick, Clyde N.:

Hibberd, Robert G.:

Hicks, David E.:

Hoenig, Stuart & E. L. Payne:

Institute of High Fidelity:

Jackson, Herbert:

Jeppesen & Co.:

Johnk, Carl T. A.:

Kaufman, Milton:

Kennedy, George:

Kip, Arthur F.:

Kiver, Lilton, Editor:

Kneitel, Thomas S.:

Lane, Leonard C.:

Larson, Boyd:

Lawton, Sherman P.:

Lenk, John B.:

Lister, Eugene C.:

Loper, Orla E. & Arthur S. Arh:

Lurch, E. Norman:

McKeeman, William M. et al:

McLaughlin, Terence:

McPartland, Joseph F. and the Editors of "Electrical Construction and Maintenance":

Malmstadt, H. V. et al:

Malvino, Albert Paul:

Malvino, Albert Paul & Gregory F. Johnson:

Mann, George B.:

Marcus, Abraham:

Marcus, Abraham & Samuel E. Gendler:

Marcus, Abraham & William Marcus:

Marcus, Abraham & Charles M. Thompson:

Marcus, William & Alex Levy:

Markus, John:

Master Publications:

Matsh, Leander, W.:

Miller, Harry M.:

Milman, Jacob & Christos C. Halkias:

Mims, Forest M.:

Mullin, Ray C.:

Murdock, J. B . :

National Technical School:

Neely, Orville, et al:

Nisbet, Alex:

Noll, Edward M.:

Oringel, Robert S.:

Philco-Ford Communication Education Products:

Philco-Ford Education Operations:

Philco-Ford Technical Education Program:

Philco TechRep. Division, Technical Dept.:

Pierce, J.F. & T.J. Paulus:

Ritchie, George L.:

Ruiz, J.:

Runstein, Robert E.:

Rutkowski, George B.:

Sams, Howard W. & Co.:

Sands, Leo G.:

Schilling, Donald & Charles Belove:

Shoemaker, Alice Klink:

Schwarz, Leland P.:

Schwarz, Martin:

Scott, Ronald E.:

Sessions, Ken, Jr.:

73 Magazine Staff:

Shiers, George:

Shrader, Robert:

Sloot, William:

Slurzberg, Maurice & William Osterheld:

Smith, Donald A.:

Smith, Ralph J.:

Stoffels, Robert E.:

Stone, Harold S. (Editor):

Sunset Books & Sunset Magazine (Ed.):

Tall, Gill & Clifford Martin:

Talley, David:

Tepper, Marvin:

Texas Instruments Learning Center:

Tilton, Edward T. & Douglas A. Blakeslee:

Tinnell, Richard W.:

Tocci, Ronald J.:

Training and Retraining, Inc.:

Tuthill, C. A.:

US Federal Communications Commission:

Van Valkenburg, M. E.:

Van Valkenburg, Nooger & Neville, Inc.:

Veatch, Henry C.:

Villanucci, Robert S. et al:

Williams, Gerald E.:

Zbar, Paul B.: