The Smith-Kettlewell Technical File

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

William Gerrey, Editor

Issue: sktf-Spring-1990

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

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 Details regarding products, suppliers, and other contact information are original and may be outdated.

Questions about this archive can be sent to









by Robert Gunderson, W2JIO

(The following is a compilation of two articles: "Reforming Electrolytic Capacitors," The Braille Technical Press, Vol. 16, No. 9, September 1965; and "Another Method of Reforming Electrolytic Capacitors," The Braille Technical Press, Vol. 17, No. 3, March 1966; plus a little from your editor.)

We are all familiar with the conventional electrolytic capacitor used in radio and electronic equipment. Its principle of operation is based on the fact that an aluminum electrode, placed in a suitable solution (electrolyte), will form upon its surface a thin film of insulating oxide which will withstand a considerable voltage. This occurs only when the aluminum is the positive electrode of an electrical circuit, and if the polarity is reversed, the film or oxide breaks down and the capacitor simply becomes a conductor. The combination of aluminum and insulating film thus formed, when the aluminum is positively polarized, forms a capacitance which has a high electrostatic capacitance per unit area of film.

The electrodes may be etched foil, or metallic film sprayed on fabric to give maximum surface area. They are separated by paper or gauze, impregnated with an electrolyte made in the form of a fudge-like solid, and the entire assembly is formed into a roll and mounted in a container.

Electrolytic capacitors have considerable leakage current due to their low resistance, and their power factor is relatively high compared to capacitors of the solid-dielectric types. They deteriorate with time, so they have a limited life. This means that they are not as dependable as solid-dielectric types, such as paper, Mylar, mica, or ceramic units.

The formation of the insulating oxide on the aluminum is an electrochemical action formed by applying a positive potential to the aluminum electrode. The thickness of this film, and hence the voltage rating and capacitance of the condenser, is determined by the applied voltage. Capacitor voltage ratings may be from 3 volts to 500 volts. However, as they age, their capacitance value changes, and if they are not in use, they may not withstand their rated voltage due to deterioration of the aluminum oxide.

If you have electrolytic capacitors which have been allowed to age on the shelf, it is well worthwhile to "reform" them before putting them into service. This reforming process consists of connecting a capacitor (using the proper polarity) to a variable-voltage direct-current supply, then gradually raising the voltage until the required working voltage of the capacitor has been reached. In fact, capacitors can be formed to higher voltages than their rated "working voltage"; this means that the oxide becomes thicker, with a consequent loss of capacitance. On the other hand, the capacitor can be formed to a lower voltage, and this means that the capacitance will be increased. The capacitor which has been allowed to operate at a low voltage for a considerable length of time will puncture when applied to its rated working voltage, unless it is reformed to its original rated value.

Reforming a 560-volt capacitor consists of operating the unit at perhaps 100 or 150 volts for 10 or 15 minutes, then gradually running the voltage up to the maximum in small steps until the 450-volt rating is reached. All this takes perhaps 30 to 45 minutes.

Therefore, if you have electrolytic capacitors which appear to be in good condition, you can, more than likely, bring them to operating condition with a little patience, plus a suitable variable-voltage power supply.

Another Method of Reforming Electrolytic Capacitors

by Jim Swail, VE3KF

[Gunderson's Note: Following the article regarding the "forming" of electrolytic capacitors in BTP, Sept. 1965, we received a number of letters describing another method of reforming capacitors of this type. However, Jim's comments came in first, so we'll credit him with the article, but our sincere thanks to others ...]

In this second forming method, the capacitor is connected in series with a resistor of perhaps 100K ohms, and connected (using the proper polarity) to a DC supply voltage whose value is equal to, but not higher than, the rated voltage of the capacitor.

Without this resistor, the capacitor will draw considerable DC current when it is first connected to the supply; as it reforms, the value will decrease. However, with this method, the current is limited by the 100,000-ohm resistor connected in series with the capacitor and the power supply; the current does not become dangerously high, which might cause the capacitor to burn out.

The state of the capacitor may be monitored by measuring the voltage across the resistance and noting that as the capacitor forms, the value of this voltage (across the resistor) decreases, due to the decreasing current drawn by the capacitor.

[Editor's Note: I think the advantage here is to eliminate the creation of high-current-density areas--"hot spots." If unlimited current is available, these "hot spots," once they start drawing current, could further heat up, draw more current, and literally burn through the dielectric. With the current-limiting resistor, thin areas which could be potential hot spots might quietly reform and not burn through. I only think I know what I'm talking about.]

This method provides an automatic means of regulating the voltage across the capacitor. As the leakage current decreases due to the formation of the aluminum oxide within the capacitor, the voltage across the resistor decreases, and we have a more or less automatically adjusting arrangement.

When the voltage ceases to fall across the resistor, the process is complete and the capacitor may be disconnected from the power supply. This process may take a bit longer, although it does have the advantage of not requiring a variable-voltage power supply. [Since this article was written, available values have increased. If you have a unit of 47,000 microfarads, the time constant with 100K would be 79 minutes; do you want to wait around for five time constants? Furthermore, the leakage of those big fellows is high. Can we please make up a rule of thumb about the charging resistor? How about sticking to time constants of about 5 seconds. Thus, the 100K would be fine for 47uF, while the 47,000uF unit would require a 100-ohm resistor.]

[Editor's Note: If the reforming is entirely successful, very little voltage will appear across the resistor; 1V would represent a leakage of 10 microamps (not unusual for capacitors of 100uF or more). Suppose, however, the voltage stabilizes to some value much higher than this. One of two conditions exist: the capacitor is reformed to a voltage lower than that of the power supply and can go no higher, or the capacitor might be reformable, but is just plain leaky. You can guess which it is by watching the voltage across the capacitor after it is disconnected; if this voltage plummets to zero, it is leaky and should be thrown out.]

Remember, electrolytics are polarized and the forming process requires that they be connected with the proper polarity. It makes no difference in which lead the resistor is connected, as long as it is in series with the capacitor.

A Third Method for Antique Equipment

by Bill Gerrey, WA6NPC

Because electrolytic capacitors lose their tolerance for high voltage as they sit around unused, it is risky to plug in antique equipment and hope the electrolytics survive. Since they sometimes can be reformed without damage, it is useful to speak of methods of "warming up" Grandma's radio in a gradual manner.

A variable autotransformer (Variac) can be used to bring old equipment to life slowly. This doesn't protect the capacitors like Jim Swail's method does, but it can sometimes save a premature failure. I have heard of people going to the trouble of removing the tubes from equipment first, and then advancing the Variac--maybe 50V, then 75V, then 100V for 20 minutes, and then ... oh go for broke!

Taking the tubes out will only work if the equipment has a selenium rectifier, or if diodes have been used to replace the rectifier tube. (Don't bother trying this on machines with mercury-vapor rectifiers in them, or some such odd thing.) I wonder how effective this technique is in equipment with tube-type rectifiers; I suppose the proper thing to do is to pull the rectifier, then apply a variable B-plus voltage from the outside.

I don't have much servicing experience; I only offer this as unconfirmed "reforming" information.


by T. V. Cranmer, K4MMB

(The following is reprinted (and updated) from The Braille Technical Press, Vol. 16, No. 10, October 1965. The dental material described is still available, and Tim Cranmer tells me he wouldn't change a thing.)


It is sometimes desirable to enclose electrical components or circuits to protect them from the effects of moisture or to provide mechanical stability. This article will describe procedures for using one material suitable for use as an encapsulating agent.

The material is acrylic plastic, chosen because of its physical properties, ease of handling, and ready availability in all parts of the country. Articles properly embedded in acrylic plastic are given excellent protection. This acrylic may be boiled in water for hours or submerged in many common solvents. The hardened acrylic material is nontoxic. It is not readily flammable. It does not deteriorate with age, and it has excellent electrical properties. Its application may only be limited by the imagination and experience of the user. This writer has used it for fashioning mountings for air-core inductors, for tying bundles of wire together, for plugging holes, and for many other applications.

Acrylic plastics may be purchased from dental supply houses under such trade names as Lang's "Traymix," or Coe Labs "Tray Plastic." Kits are available for approximately $35 which include one pound of the powdered acrylic, a bottle of liquid polymer, measuring spoon and vial, paper cups for mixing, and wooden stirring paddles. Directions included with the kit will be of some help in understanding the uses of the materials. However, you must remember that these directions are intended for dentists and will be limited in scope.

[Lang can supply you with five colors, one of which is clear. Coe Labs seems to only come in blue and white, although the editor does not have their catalog at hand just yet to aid in listing all the choices.]

General Characteristics

When approximately two parts of acrylic powder and one part of polymer liquid (by volume) are mixed, a thick gritty solution results. After a few minutes, the mixture assumes the consistency of heavy molasses. When it is in this state, the mix pours slowly or not at all. If it is touched with the fingers, it will feel sticky and will form tacky strings when the fingers are withdrawn. A few minutes later, the mixture acquires the consistency of dry dough. In this condition, it may be removed from the mixing cup and shaped into any desired form. This dough-like state lasts for just a minute or two (depending upon temperature). After this, the material generates internal heat and becomes a very hard solid mass. Once hardened, the plastic may be filed, ground, or polished as desired.

By varying the proportions of powder and polymer, the initial liquid state can be made quite thick or very fluid. The physical strength of the finished product is somewhat less when the thin mix is used.

Encapsulating Circuits

A transistor oscillator, amplifier, or other circuit may be completely embedded in solid plastic. The following steps give good results:

Select or fabricate a cardboard tray large enough to hold the circuit. Line this tray with thin tinfoil, also available from your dental supply house. (Note, we mean "tin foil"; the term is not used loosely to mean aluminum or other foil.)

Prepare a quantity of acrylic powder and liquid polymer sufficient to fill the tray to a depth of about 1/8 inch. Pour the mixture into the tray immediately while it is still in the liquid state. Allow the material to reach the dough-like state; then gently press the circuit board or circuit components into place, slightly impressing the circuit into the plastic. Next, prepare a large quantity of the plastic material and pour this around and over the circuit until it is completely covered.

Within fifteen minutes, the material will become hardened and may be removed from the tray--with the circuit inside. Tinfoil is necessary to prevent the material from adhering to the paper tray. No lubricants are necessary to facilitate removal of the finished product when tinfoil is employed.

[These companies do not supply tinfoil any more. Rather, the new thing in dental work is to paint the inside of their plaster mold with a "mold release." Tom Fowle says you can use "green soap" as a mold release in plaster molds (the antiseptic soap they used to torture us with as kids). The paint-on stuff is either Lang's "Liquid Foil" or Coe's "Tray Sep." The editor can't say what this would imply for lining a cardboard tray. Andrea at Lang Dental suggested Vaseline for cardboard; this doesn't address the issue of imparting a smooth surface to the plastic, however.

Although tinfoil seems out of fashion nowadays, I did find a manufacturer of it--Buffalo Dental. It costs about $30 for a one-pound roll.]

Packing Plugs and Connectors

Microphone connectors, phone plugs, and similar devices often give trouble due to the flexing of the cables to which they are attached. Broken connectors will be prevented or greatly reduced if these connectors are packed with plastic. A neat job will result if the following procedure is used:

Remove the cap covering the electrical connections to the plug. Prepare a small quantity of plastic mixture and allow it to stand until it reaches the dough-like state. Using the fingers, press the plastic dough into place and shape the material so that when the cover is replaced, there will be enough material to fill it completely. Working quickly, force the cover into place and remove any plastic material that may squeeze out. After 15 minutes, the material will harden and provide permanent protection for the connections.

To prevent the materials from adhering to the fingers, it is a good idea to rub into the fingertips a liberal quantity of hand cream before beginning the job.

Preparing Probes

You may wish to prepare a probe by attaching a sensor--a thermister, a thermocouple or other device--to the end of a flexible cord. A rugged and attractive probe may be made by enclosing the junction of the element and its cord with plastic.

For this application, use the plastic material in the dough-like state. Press the material around the connections and shape it into a cone or cylinder with the axis parallel to the flexible cord. While the material is still soft, it may be wrapped in a strip of tinfoil which will aid in obtaining the desired shape and will impart a better finish to the hardened plastic.

A short length of plastic or rubber tubing may be used as a mold when preparing probes. Close off one end with the tip of your finger, pour it full of liquid plastic, and press the sensor element into position in the liquid material. After the plastic has hardened, the rubber tubing used as a mold may be split with a razor blade and peeled off. This results in a very attractively finished product.


The appearance of the work will be improved if surplus material is filed away. The surface may have powdered texture. This is removed with very fine sandpaper, No. 400. A piece of crocus cloth may then be used to impart an extra-smooth finish, or the material may be buffed with a cloth polishing wheel run at moderate speed.

* * *

If you wish to encapsulate components or circuits, we hope you find this information of use. As to other applications, as we said earlier, your imagination is the limit.

Address List

[Note: Most companies prefer not to sell to customers directly, but to refer you to a local dental supplier. However, Andrea at Lang Dental said that there would be no problem selling directly, as long as the "Traymix" is not being used for dental purposes. You can buy directly from Lang, even from overseas. On the other hand, there are dental supply houses with Coe "Tray Plastic" in every major city; just walking in and buying it hasn't been a problem for me or for Tim Cranmer.]

Buffalo Dental (has tinfoil): 575 Underhill Blvd., Syosset, NY 11791; (516) 496-7200.

Coe Labs: 3737 W. 127th St., Chicago, IL 60658; (312) 568-2100.;

Lang Dental Manufacturing: 175 Messner Drive, Wheeling, IL 60090; (708) 215-6622.


by Bob Prahin, W8DIL


I think this circuit works just as well as the Smith-Kettlewell Auditory Gimmick (SKTF, Spring 1981) but contains fewer parts. An LM358 dual op-amp, the ever-popular NE555, two PC-mount pots, and a handful of resistors and capacitors are all you need to build the thing.

For some 25 years, I've been using my old Transistorized Auditory Gimmick (first published by Bob Gunderson in 1954) to tune my amateur transmitter. Except for a few modifications (like forward-biasing the input circuit to make it more responsive to low voltages), it still contains the original components--including the old germanium transistors. I recently decided to update things around W8DIL; this didn't include replacing the Drake C-Line transmitter and receiver, however.

After rereading the editor's Auditory Gimmick article of Spring 1981, I sought to simplify the circuit a bit. (My, my! Two chips, three transistors, and I had nightmares about "current mirrors" for days.) After playing around for a while, I think I came up with a circuit that works just as well.

Circuit Operation

I've wondered why pin 5 of the 555 has been so unpopular in these pages. (It was used very cleverly to obtain a reference voltage in the Audible VU Meter.) [I just never liked pin 5, and leaving it alone made me very happy. -- Editor.] If pin 5 is taken through a 2.2K resistor to ground, the pitch will rise to its maximum. (If this resistor is decreased much below this value, the tone will disappear. This assumes a 9-volt power supply.) This increase in pitch, however, is not terribly dramatic. A much greater range can be obtained by taking the far end of this resistor through the full range--from 0 to 9 volts.

The resultant pitch range is slightly under 2-1/2 octaves. This seems more than adequate, even for those with "tin ears."

In operation, the DC voltage from the meter is fed to the non-inverting input of the first half of the LM358. There is a sensitivity control (a 25K pot) in the inverting input of this op-amp. The output of the first op-amp is fed to the inverting input of the second op-amp in the 358, with the output of the latter going through the 2.2K resistor to pin 5 of the 555. There is a 10K threshold adjustment in the non-inverting input of the second op-amp. (These adjustments are interdependent.) The rest of the 555 circuitry is unremarkable.

Because this unit has a very high input impedance, it is extremely sensitive to RF fields. It should be built in a metal box which is grounded, via the input jack, to the equipment to which it is connected. Both sides of the circuit are isolated from the input by RF chokes and capacitors--the "common" side of the circuit is not connected to the case. For this reason, the term "circuit common" is used for the "cold" side of the gimmick circuit.

This circuit is intended to be used with meters that are at, or near, ground potential. If you must monitor a meter in the positive high-voltage lead, it will be necessary to build the "gimmick" in a Bakelite box which is then put into a grounded metal box. In this case, insulated couplings to switches and a high-voltage input connector will have to be used.

Gimmick Circuit

The negative of a 9-volt battery goes to pin 4 of the LM358, with this pin 4 also going to the cathode of a 1N4001 diode whose anode goes to circuit common. (This diode assures that the op-amps work at 0.6 volts below ground.) The positive of the battery goes through an SPST on-off switch to VCC. Pin 8 of the 358 goes to VCC.

Pin 1 of the NE555 goes to circuit common, while pins 4 and 8 are tied together and go through a 10-ohm resistor to VCC. Pins 4 and 8 also go through a 100uF capacitor to circuit common (negative at circuit common).

The input connections come through a 2-conductor shielded cable and a tip-ring-sleeve plug. The matching jack of the gimmick has its sleeve grounded to the case. The shield of the cable goes to the sleeve of the plug. (This shield also goes to the case of the equipment with which the gimmick is used.)

The positive lead (tip) goes through a 2.5 millihenry RF choke, thence through a 220K resistor to the non-inverting input of the first half of the LM358, pin 3. Pin 3 also goes through 0.1uF to circuit common. The negative meter lead (ring) goes through a similar 2.5 millihenry RF choke to circuit common. There is a 0.01uF capacitor between the tip and ring of the input jack, and there is another 0.01uF capacitor between the RF chokes on the gimmick side.

The inverting input of the first op-amp (pin 2 of the 358) goes to the arm of a 25K pot (sensitivity control). One end of this pot goes through 1K to circuit common, while the other end goes through 27K to the output (pin 1). Pin 1 also goes through 10K to the inverting input of the second op-amp (pin 6), with this pin 6 also going through a 220K "feed-back" resistor to the output (pin 7).

The non-inverting input of this second op-amp (pin 5) goes to the arm of a 10K pot (threshold adjustment). One end of this pot goes to circuit common, while the other end goes through 100K to VCC.

Pin 7 of the 358 also goes through a 2.2K "coupling" resistor to pin 5 of the 555. Pins 2 and 6 of the 555 are tied together and go through 0.01uF to circuit common. Pins 2 and 6 also go through 10K to pin 7, with pin 7 also going through 220K to the junction of pins 4 and 8. The output (pin 3) goes through a 47-ohm resistor, thence through the speaker to pins 4 and 8.


This device can be made just about as sensitive as you like--it will saturate (achieve maximum pitch) with only a few millivolts of input. Since the sensitivity and threshold controls are interactive, I suggest that you start by setting the sensitivity pot near maximum sensitivity (nearest the 1K resistor), and with the threshold pot set at maximum (nearest the 100K resistor).

Now, with the input shorted, adjust the threshold pot (bringing its arm toward the ground end) until the pitch just begins to rise. This is the correct setting for the threshold.

Connect the gimmick to the meter to be monitored, and adjust the transmitter for maximum power output as indicated by maximum pitch. If maximum pitch occurs before full power has been reached, it will be necessary to back off the sensitivity control a bit. There should be enough "head-room" in the "gimmick" to allow full power output to be achieved without saturating the unit.

Remember--for every new setting of the sensitivity control, the threshold pot will need readjustment. This is done by disconnecting the "gimmick," shorting its input terminals (shorting the tip to the ring), then setting the threshold control to the point where the pitch just begins to rise.

If you know the optimum voltage that is established across the meter terminals during "tune-up," you can adjust the "gimmick" with a battery (such as a 1.5-volt cell), a potentiometer and a voltmeter.

This device is intended to respond to very low input voltages. If your meter develops more than 200 or 400 millivolts across its terminals, you will need to reduce the value of the feedback resistor between pins 6 and 7 of the 358. Some experimentation will be necessary here--try 100K ohms for starters and work down. As this feedback resistor is reduced, it may also be necessary to increase the value of the 10K pot in order to obtain a higher reference voltage on pin 5 of the 358.

The "Fowle Gimmique" (SKTF, Summer 1982) could easily be adapted to this circuit. You would need an NE556 dual timer chip and an LM324 quad op-amp. (One of these op-amps would be disabled by tying the inverting input to its output and connecting its non-inverting input to circuit common.)

The output of one-half of the 556 (the low-frequency oscillator) would go to the Enable pin of the "gimmick" oscillator. The speaker would have to be capacitively coupled; the 47-ohm resistor would go to the positive of a 10uF capacitor, with the negative end of this going through the speaker to ground.

The output of the "comparator" op-amp would then go through the timing resistor to the discharge pin of the low-frequency oscillator. In this way, when the voltage on the arm of the Braille-calibrated potentiometer equals or exceeds that across the meter movement, the "gimmick's" tone would be interrupted, and the meter could be read by noting the position of the Braille dial.

It is suggested that you read Tom's article for further details.

Other modifications--"code-practice oscillator," "continuity checker," "keying monitor," and "field-strength meter," to name just a few--will be left to the ingenuity and imagination of the builder. (Many of these applications are described as attachments in the article "Auditory Gimmick Circuits--Old and New," SKTF, Spring 1981.) Who knows, you could end up with a box full of jacks and switches like Bernie's Little Blue Box.

Parts List

Resistors (1/4-watt 5%, unless otherwise specified):

  • 1--10 ohms 1/2-watt
  • 1--47 ohms 1/2-watt
  • 1--1K
  • 1--2.2K
  • 2--10K
  • 1--27K
  • 1--100K
  • 3--220K

Potentiometers (PC-mount trimmers):

  • 1--10K
  • 1--25K
  • Capacitors:
  • 1--0.01uF Mylar
  • 2--0.01uF ceramic
  • 1--0.1uF ceramic
  • 1--100uF 10-volt electrolytic


  • 1--1N4001
  • 1--LM358
  • 1--NE555


  • 2--2.5mH RF chokes, low-current
  • Speaker, switch, tip-ring-sleeve jack, box, etc.


"Now he's really gone wrong; he publishes a device that he hopes you'll never want." That's right, I hope you will never need this, but here it is for your empowerment.


Some years ago, a family member began to receive troublesome phone calls. Social outcasts usually make those calls because they think you can't do anything about them, and this is a "power" they have over you. Well, I did something:

In the old days (and in some states even today), a law said that if you record a telephone call, there must be regularly occurring beeps--1400Hz, 1/4 second in duration every 15 seconds. I don't think it's the law here, but whenever you call the police to report city life, that beep stands out as an experience.

My hope was that our prankster might have heard of this law, or would at least take notice. So, I built a CMOS beeper that could be left on the line at all times; it would be constantly running, no matter what extension phone you picked up. Because no one wants to talk to a prankster for long, I increased the repetition rate to beep every 7 seconds. (I didn't bother with the tape recorder; what would you do with tapes if you had them?)

One day, a person with improper etiquette did not properly identify himself, but in the course of the brief conversation, he asked, "What's that beeping?" Came the off-handed reply, "Oh, all our phone calls are recorded." The caller hurriedly ducked out of the conversation. No one is quite sure that this caller was the one intended for the trap, but the prankster was never heard from again, and it can be said that the adrenalin of somebody was raised for at least a moment.

This circuit having fallen into disuse, I decided to equip a telephone with it for actual recording purposes. There is an advantage in doing this; if you feed the beeper into the "telephone transmitter" (microphone) circuit, the "side-tone cancellation" (attenuating your own speech to your receiver) works to reduce the level of beeping in the recording. (This is at least true for recordings made by induction off the earpiece.) So that it would be more conventional in operation, I increased the time between beeps to the prescribed 15 seconds.

Description and Circuit

The Telephone Beeper can be built into a small box, perhaps 2-1/2 by 4-1/2 by 1-1/4 inches (Radio Shack 270-221). The circuit fits easily on a piece of perforated board measuring 2.2 by 2.8 inches. With the board mounted toward one end of the box, there will be room for the battery to stand on edge at the other end.

If you happen to have one, a telephone modular jack (Radio Shack 279-353) can be mounted in the side of the cabinet. On the other hand, a salvaged modular telephone cord can emerge from a hole in the cabinet. The on-off switch should be mounted so as to leave room for the battery.

The device is attached to the telephone line by plugging both it, and any extension telephone, into a modular Y-connector (Radio Shack 279-357).

The circuit uses a CMOS dual timer chip, such as the Intersil ICM7556. One timer is a "tone-burst" controller--a very slow free-running oscillator--which operates the 1400Hz oscillator of the second half. The signal is RC coupled to the line; the choice of R and C is such that the ringer voltage is attenuated by perhaps 32dB. An arrangement of diode clamps keeps the ringer voltage from damaging the 7556 chip.

Two values of "timing resistor" are recommended for the tone-burst controller: A 1-megohm unit off pin 1 makes it beep every 7 seconds. A 1.8-megohm resistor makes it beep at the standard interval of once every 15 seconds.

Note the arrangement of the 1400Hz oscillator. With this novel approach, short downward pulses of the tone-burst controller provide charging current for the oscillator. In circuits where current drain doesn't matter, I usually use a control transistor to "inhibit" the operation of a standard 555 oscillator, but this approach eliminates that extravagance.

[My original unit needed an extra CMOS 7555 one-shot to control the beeper; using conventional hookups, you run into duty-cycle arguments and wrong polarities of signals. The invention of the "upside-down 555 oscillator" is that of Tom Fowle, and I want no publication of mine to suggest that I had this technique earlier in my career.]

Finally, if the power-supply bypass capacitor were on the circuit side of the switch, turning off the circuit would not kill the Beeper immediately. Therefore, the battery itself is bypassed.

Circuit for the Telephone Beeper

The negative terminal of the 9V battery is grounded. Its positive terminal is bypassed to ground by 10uF (positive of the cap at this junction). Beyond the positive junction of the bypass capacitor, an SPST on-off switch is in series with the VCC line.

Pin 7 of the ICM7556 is grounded; pin 14 goes to VCC. Pins 4 and 10 are tied together and go to VCC.

Pins 2 and 6 are tied together and go through 10uF to ground (negative of this cap at ground). Pins 2 and 6 also go through 33K to pin 1. Pin 1 goes through a timing resistor to VCC (1 megohm or 1.8 megohm, see text).

The output of this tone-burst section, pin 5, goes through 47K, then through 22K to pins 8 and 12--which are tied together. Pins 8 and 12 also go through 0.01uF to VCC. The junction of the 47K and 22K resistors goes to the cathode of a 1N914 diode; the anode goes to the output of this section, pin 9.

Pin 9 goes through 2.2K, then through 10K, then through a 0.022uF 500-volt capacitor to one side of the telephone line; the other side of the telephone line is grounded. The junction of the 2.2K and 10K resistors goes to the cathode of a 1N4005 rectifier diode whose anode is grounded. The junction also goes to the anode of another 1N4005 whose cathode goes to VCC.

Modifying the Circuit for Direct Inclusion Into the Telephone

I cannot speak for some modern telephones which have capacitor-type microphones in them, and would hence take a different amount of coupling from the beeper. However, with the old Western Electric phone I have--with a carbon "transmitter," the coupling scheme shown works fine. Side-tone cancellation" keeps some of the beeper signal out of the local earpiece.

Caution! Some telephone instruments are still capital equipment of local telephone companies. Make sure any phone you modify is truly your own.

Taking the plastic housing off the phone reveals a network box at the right-rear corner of the base plate. Configurations of network boxes varied over the years, but I can at least tell you about my instrument.

It is probably wise to trace the handset wires with a continuity tester--that's what I did. You can easily tell the difference between a direct wire and other resistances found in the telephone network with a Science Products "Audicator." To do this, unscrew the cover over the mouthpiece and take out the carbon microphone. Behind it, you will find two spring contacts that go to the red and black wires, which are the ones you want.

On my instrument, the four wires from the handset go to the row of screw terminals nearest you--nearest the right-front corner of the network box. The red and black ones from the handset are from the microphone (transmitter); the two white wires are from the earphone (receiver).

Both a red and a white wire share the screw terminal at the right-front corner. The black microphone wire is the next terminal to the left. (The third terminal from the right has the other white earphone wire). The black microphone wire is positive with respect to the red.

The 10K resistor and clamping diodes may be dispensed with on the beeper circuit. The beeper, with its battery, is mounted in the phone where room permits. Then, the ground side of the beeper is connected to the right-front network terminal (with the red and white handset wires). The output of the beeper, pin 9, goes through 22K, then through 0.022uF to the terminal next door--to the black handset wire. (Actually, you can leave all the output clamps and resistors in place, but add another 10K in series with the whole mess.)

Further Notes

This circuit takes about twice the period between beeps to "warm up" and make its first beep. This initial delay is due to how the timer circuit of the 556 chip works (see "Inside the 555," SKTF, Winter 1987). With the upside-down oscillator circuit (see above), the frequency of the tone is more dependent on battery voltage as conventional 555 hookups are. However, the values given will closely approximate the proper 1400 Hz signal, and as the battery runs down, lowering of the pitch will be a signal for you to change it.

The circuit draws about 100 microamps between beeps, with the current drain during a beep depending on whether the phone is off the hook or not. The worst case is if the phone is off the hook, and the beeper is producing its tone; a rough calculation brings the current up to an average value of 450 microamps. Since, with the rapid burst, this current is drawn for only 1/28th of the time, a new average might turn out to be under 120 microamps. At this rate, a 500mAh battery would last over 4000 hours.

Parts List

Resistors (1/4-watt 5%, unless otherwise stated):

  • 1--2.2K
  • 1--10K 1/2-watt
  • 1--22K
  • 1--33K
  • 1--47K
  • 1--1 megohm or 1.8 megohm, depending on the desired timing (see text)


  • 2--10uF 10-volt electrolytics
  • 1--0.022uF 500-volt disc ceramic
  • 1--0.01uF Mylar


  • 1--1N914 small-signal diode
  • 2--1N4005 rectifiers
  • 1--CMOS 556 dual timer, such as the Intersil ICM7556


  • 1--SPST on/off switch
  • 1--salvaged telephone cord, or "modular jack" (Radio Shack 279-353)
  • 1--modular Y-connector, Radio Shack 279-357
  • 1--small cabinet, such as Radio Shack 270-221

Tinning "Tinsel Wire"

In order to make them flexible, the majority of wires used in telephone cords is the so-called "tinsel wire." Tinsel wire consists of very thin strands woven among thread, and it is very hard to solder.

The secret is: Once you have stripped an end, twist the bared end to make sure that it is bundled together. Wrap the bared end with a thin strand of tinned copper wire (borrowed from a piece of regular stranded wire). Wrap this assembly with a very generous amount of solder, and quickly tin it by dragging it across the tip of your soldering iron.

For more on tinning stranded wire, see "Soldering, Part III," SKTF, Spring 1981.



This circuit is quoted from National Semiconductor Applications Note No. AN29, December 1969, Page 10. It is a simple circuit of general interest. Notes on how it might be used to extend the low ranges of the "Multi-Microfarad Meter" by K. Britz (SKTF, Spring 1986) are given.

The op-amp shown was the LM108, once considered the bee's knees. However, the circuit is straightforward, and it can only benefit from using a modern op-amp.

Capacitance Multiplier Circuit

With the values given, this circuit multiplies a 10uF capacitor by a factor of 10,000. If a single op-amp is used, pin 4 goes to a negative supply voltage, while pin 7 goes to the positive supply. (On op-amps which can operate very near ground, the negative supply pin can probably be grounded. If a dual supply is used, the junction of the supplies is taken as ground.)

Between output and inverting input is a 10meg resistor (R2). (This is put there to balance the offset currents on the two inputs, I think. If I'm right, this does not have to be a 1% resistor like the others.)

The capacitor to be multiplied, C1, goes from the non-inverting input to ground. (The sample circuit shows a 10uF electrolytic with its negative end at ground.) The output goes through R3 (1K 1%), then through R1 (10 megohms 1%) to the non-inverting input. The driving point is the junction of R3 and R1; i.e., the multiplied capacitance is seen between the junction of these resistors and ground.

The multiplier is the ratio of R1 over R3 (10 megohms divided by 1000 ohms is 10,000 in this sample circuit).

Unfortunately, besides having a capacitance of 0.1 farad, the value of R3 is always in series with this. The equivalent circuit is the series combination of R3 (1K in this case) and the multiplied capacitance.

The equivalent leakage depends on the offset voltage and offset current ratings of the op-amp used. It is expressed by the formula:

The equivalent leakage equals the offset voltage, plus the offset current times R1, all divided by R3.

Using the Texas Instruments TL071 op-amp as an example: Its offset voltage is 9mV maximum, and its offset current is 100 picoamps maximum. The current offset times R1 is 1mV; the rated offset voltage is 9mV. Ten millivolts divided by R3 (1000 ohms) is 10 microamps. They don't say how the leakage of their 10uF electrolytic contributes to the overall leakage; I would think this could be significant.

How does it work? Our best guess is that as you try to pull up on the driving point, the op-amp draws most of the current, sharing one 10,000th of this current with C1. As C1 charges, the output of the op-amp rises, just as the voltage on a real capacitor would.

Modifying the K. Britz Meter

On the face of it, the binding posts of the Multi-Microfarad meter would go to this multiplier circuit; the negative binding post would go to ground (ground being common to both circuits), and the positive binding post would go to the junction of R3 and R1. The unknown capacitor would go to two new binding posts, the negative one being grounded and the positive one going to the non-inverting input of the op-amp.

The resistance in series with the equivalent capacitance will become significant as small charging resistors are selected with the range switch. (Small resistor values are used in the higher-range positions.)

We'll have to strike a compromise with the resistance of R3. We could change R3 to 100 ohms and make R1 and R2 each 1 megohm. (Remember that R1 and R3 are 1% resistors.) Or, we could do nothing and just tolerate a 10% error when the 10K charging resistor is selected. We needn't go to any higher range than that of 100uF per beep; this is the one which uses the 10K charging resistor, and with the multiplier in place, this would give us a range of 0.01uF per beep.

[I thought of removing pin 6 of the 555 from the binding post and charging resistors, then tying it to the output of the op-amp--trying to sense the charge of the "capacitor" on the other side of the 1K resistance of R3. This is clever, but it doesn't really fix the fact that R3 is still there in the charging string.]

You know that the right thing to do is to charge the test capacitor through a precision current source, time the charge with a digital timer (such as the XR2240), and read the timer's output with a speech board. Oh, no!

[One amp applied to a capacitor of 1 farad will charge it at a rate of 1 volt per second.]

Pin Assignments for Single Op-amps

Op-amps such as the LM108 and LM308 require a compensation capacitor--often 30pF, sometimes 100pF--connected between pins 1 and 8. Also, some op-amps provide for offset-voltage correction by placing a pot between pins 1 and 5, then tying the arm to ground or one of the supply voltages. Various op-amps require different arrangements of this pot; some take the arm to ground, some to the positive supply, and some use a resistor in series with the arm as well. Applications notes for the op-amp of your choice will tell you what the manufacturer recommends.

  • Pin 1--compensation, offset, or maybe unused
  • Pin 2--inverting input
  • Pin 3--non-inverting input
  • Pin 4--minus supply
  • Pin 5--offset, or maybe unused
  • Pin 6--output
  • Pin 7--plus supply
  • Pin 8--compensation, or maybe unused


There are some new guys on the block marketing slates, check-writing guides, and inexpensive "click rules" (the "click rule" is my favorite item of theirs). They are: Community Advocates, Inc., P.O. Box 83304, Lincoln, NE 68501, telephone 402-435-7423.


The following was submitted by Mark Schafer of Lakewood, CA.

I am writing to inform you of the Handicapped Conference on the Interlink System. Interlink is a computer bulletin board message-base that echoes the messages it receives from all over the United States and Canada. It contains conferences that are of interest such as pets, sports, and politics.

As the moderator of the conference, I would like to invite all who are interested in issues that affect the handicapped to join us. The purpose is to share ideas and be of help to one another as we face issues and problems together. All you need to join this conference is a computer and a modem. I have a list of a group of hub centers throughout the United States and Canada; for a copy, write to me at 4558 Albury Ave., Lakewood, CA 90713. If you call and leave a message for the system operator, you can probably find a number that is local to you.

There may be a small charge for the use of the board, but I think you will find that the enjoyment you will realize from getting involved with Interlink will make it worth your while.


The following note comes from the Summer 1989 issue of Handi-Ham World:

"Attention IBM users! Bill Blohm, Handi-Ham student from Idaho, has generously offered to share with any Handi-Ham members his public domain ham radio programs. If you have an IBM computer and would be interested in software for ham radio, please contact Bill Blohm, 809 Davis Avenue, Nampa, ID 83651.


The Consumer Electronics Industry's "1990 Annual Review -- Entertainment and Education, Yesterday, Today and Tomorrow" provides an introduction to one of America's most dynamic and innovative industries. Compiled by the Communications Department of the Electronic Industries Association's Consumer Electronics Group (EIA/CEG), this profile and history of a rapidly changing $44-plus billion retail industry traces the development of consumer electronics product categories such as video, audio, home information equipment, and personal electronics. In addition, this authoritative source of information contains useful statistics on industry-wide sales trends, product by product. All data was compiled by the Association's Marketing Department, which has developed comprehensive statistical reporting programs for over three decades.

For a copy, contact EIA/CEG Communications Department, 2001 Eye Street, N.W., Washington, D.C. 20006, or call 202-457-4919.


I have received the following note from Jack Yeaman, 409 South 1300 East, Salt Lake City, Utah 84102:

"I have many braille copies of the Braille Technical Press. Would these be of any use to you in your training program? If so, I can send them to you."

Thank you, Jack Yeaman. I don't have room for storage, but perhaps I should put your notice in the SKTF Bulletin Board!


An ingenious idea turned out to send your editor on a wild goose chase, but something finally came of it. A modification was suggested for Mr. Britz's instrument, the Multi-Microfarad Meter. It comes from Pim Brouwer of Hayes in Middlesex County, England, who says:

"By the way, remember that meter (SKTF, Spring 1986)--if instead of the timing capacitor governing the one-shot, you used an op-amp (such as the CA3140) as a capacitance multiplier, you could extend the range of the instrument downwards (probably into the picofarad range)."

Leave it to the British to revive the Miller theorem. In World War II, the British surprised everybody with an improvement to radar that made it a great deal more accurate. Based on the Miller theorem, very linear "ramp functions" were achieved with this new invention.

The initial slope of an exponential curve is pretty straight (almost linear). When you apply a voltage to a series RC circuit, the first 10 percent of the first time constant is fairly linear. Therefore, if you could only make large capacitors, you could use the first portion of the curve to generate linear sweeps, or to time things accurately. British engineers were able to expand exponential curves electronically to provide their radar with such linear ramps.

Students of analog computing will remember the Miller theorem whereby the value of a capacitor can be multiplied by the open-loop gain of an op-amp. The capacitor is a feedback loop between output and inverting input, and the resultant capacitance between the inputs appears very high. Well, for a while, this looked pretty good; why not put the unknown capacitor into the integrator and drive the integrator's input with the Multi-Microfarad Meter.

Op-amp manufacturers don't give you precise figures of open-loop gain; they can only tell you that it will be higher than such and such. So, let us construct an inverting amplifier out of two op-amps and build it with a gain of 10,000.

The catch is that both the time constant and the amplitude of the amplifier's output--as it traces the exponential charging curve--are multiplied by the gain of 10,000. The Multi-Microfarad Meter times how long it takes for the charge on a capacitor to get up to 6 volts. Oh fine, for the simulated capacitance on the input to reach this value, the output has to go to minus 60,000 volts. I don't think that my op-amps, nor my capacitors, would stand it.

However, there are other "capacitance multiplier" circuits using op-amps, and we have reprinted one in this issue. Thank you, Mr. Brouwer, for sending me back to school--Al Alden's office--and for a kinky idea.