SKTF -- Fall 1989

The Smith-Kettlewell Technical File

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

William Gerrey, Editor

Issue: [current-page:title]

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

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 Details regarding products, suppliers, and other contact information are original and may be outdated.

Questions about this archive can be sent to


The Ring Detector: A Telephone Interface

Accessing the Fluke PM2525/623 Multimeter

The Smith-Kettlewell "Talk & Tones" Volt/Ohm/Milliammeter

The Robert W. Gunderson Memorial Fund

Editor's Corner


By Albert Alden


Since we got a new phone system here at Smith-Kettlewell, the standard-issue "Unity Phones," as they are called, sound so much alike that they've kept us all running to see if the phone we hear is our own. In an effort to distinguish my telephone ring from others in the same vicinity, I thought I would try something "spectacular": a 10-inch diameter fire alarm bell. This article describes an interface circuit used to ring the bell. The circuit consists of an opto-isolator, a retriggerable one-shot multivibrator, and a power transistor.

Technical Description

On our phone system, the potential on the phone line is about 50 VDC. The ringing signal is a 20 Hz sinewave at 200V peak-to-peak riding on the 50 VDC. When the phone is picked up, the voltage drops to about 6 volts.

An opto-isolator looking at the phone line is used to trigger a one-shot. The opto-isolator I used is a General Electric H11L1. The input LED is connected to the phone line in series with a current-limiting resistor and a reverse polarity protection diode in parallel with the LED.

The output side of the isolator is an open collector Schmitt trigger. The output is normally high (there being an external pull-up resistor). When the LED current exceeds the turn-on threshold (approximately 1mA), the output goes low. This output triggers the one-shot. The period of the one-shot is approximately 100ms. This is longer than the period of the phone's ringing signal. Therefore, the output of the retriggerable one-shot stays high during the presence of the ringing signal.

The output of the one-shot drives the base of the TIP120 power Darlington. The bell is connected between the positive of the power supply and the collector of the TIP120. There is a catch diode across the bell.

Circuit Description

A 10K 1/2-watt resistor goes to pin 1 of the H11L1. A 1N4004 diode has its cathode connected to pin1 and its anode connected to pin 2 of the H11L1. Pin 6 goes to plus V, and pin 4 is grounded. A 10K 1/4-watt resistor goes from pin 5 to plus V.

Pin 5 of the H11L1 is connected to pin 5 of a CD4538. Pin 2 of the CD4538 goes through a 100K 1/4-watt resistor to plus V. Also connected to pin 5 is the plus lead of a 1uF tantalum capacitor; the negative lead goes to ground. Pins 3 and 16 go to plus V, and pins 1, 4, and 8 go to ground.

From pin 6 of the CD4538 there is a 220-ohm 1/4-watt resistor connected to the base of the TIP120. The emitter is connected to ground. One lead of the bell is connected to the collector, and the other to the positive of a power supply. A 1N4004 diode is connected in parallel with the bell. The anode is at the collector of the TIP120.

The opto-isolator and one-shot can be powered from a supply ranging from plus 5 to plus 15 VDC. The power for the bell (or any other load) may be the same or different. For my application, I found that because of the nature of the load (large pulsed currents) either heavy-duty decoupling or a separate power supply was necessary to prevent false triggering of the one-shot.

Phone Line Connections

The positive phone lead should be connected to pin 2 of the H11L1, and the other lead to the free end of the 10K 1/2-watt resistor. With this connection, the negative swing of the ringing signal will turn on the LED in the opto-isolator. This will trigger (or retrigger) the one-shot with each cycle. If the polarity of the connections is reversed, the one-shot will be triggered once when the phone is hung up, being triggered by the ringing signal. Do not connect the phone lines to any other part of the circuit.

Parts List

Resistors (fixed 1/2-watt 5%):

  • 1--10K

Resistors (fixed 1/4-watt 5%):

  • 1--220-ohm
  • 1--10K
  • 1--100K


  • 1--1uF tantalum


  • 2--1N4004
  • 1--Texas Instruments TIP120
  • 1--G.E. H11L1
  • 1--CD4538


by T.V. (Tim) Cranmer, Chairman Research and Development Committee

National Federation of the Blind

[Editor's Note: Since this article was written, the Fluke company has issued a new version. Originally, this article applied to the PM2525/613; now it is a PM2525/623. The sales people at Fluke assure me that the same commands apply, but that the capabilities of the RS232 port have been expanded.]

In recent months, the R&D Committee of NFB has actively pursued information about laboratory and industrial instruments that blind workers can access without the necessity of modifying the instruments. We examined tools like laboratory balances, industrial parts counters and various chemical monitoring systems that may have vocational significance for blind job applicants. This work is continuing.

More and more meter measurements are taken and stored by computers. Extensive data bases consisting of measurements made over a long period of time would become very labor intensive if each measurement were to be taken manually and written in a log. Consider the apparently simple task of plotting the voltage drop of a battery under constant load over a period of several hours. You could (I couldn't) slavishly read the voltage across the load every minute, every five minutes, or whatever. Alternatively, you could write a simple program for a computer to automatically take the readings for you and place them in a disk file for later study.

Laboratories in universities and industries routinely collect data on electrical, physical and chemical measurements. Such activities stimulate demand for instruments with computer interfaces. It was while searching for various machine-readable instruments that we came upon the Fluke PM2525/623 multimeter--the 623 at the end of the model number specifies an optional RS232-C computer interface.

This is a very versatile multimeter. Functions include autoranging measurements of AC and DC volts or amps, 2-wire and 4-wire resistance (the editor will explain the 4-wire resistors), diode ratio, dBm, capacitance time (as small as 10 microseconds) and frequency. The accuracy is 5-1/2 digits. Oh yes, with plug-in thermometer probes, you can extend the range of the instrument or measure temperature with a resolution of a tenth degree Centigrade or Fahrenheit.

[Four-wire resistors indeed. What this means is that there are four test leads, one pair to supply the resistor with a known current, and the other pair to sense the voltage drop across the resistor. In this way, any voltage drop in the test leads is not included in the measurement.]

The Fluke PM2525/623 is entirely accessible without any modification using a Braille'n'Speak, Pocket Braille, VersaBraille, or your talking computer. I personally find the Braille'n'Speak most convenient. If you don't have one, let's hope there's one in your future. Meanwhile, use whatever terminal you have. Here's how easy it is:

Set the baud rate and other communication parameters on the Fluke by flipping the DIP switches conveniently located next to the RS232C D-connector on the rear panel. Connect the cable between the Braille'n'Speak (or whatever) and the Fluke. Power up both devices--remember to enable the serial port on the Braille'n'Speak.

Now, to put the Fluke in remote mode, send escape2. That is, transmit to the Fluke the escape character followed by the digit 2. Now the Fluke is in remote mode and you are in control! Now, "enable measurement output automatic" by sending EMO A and a carriage return. The Fluke will respond by continually (automatically) sending back readings from the visual display on the front panel. With the Braille'n'Speak or terminal in interactive mode, you should now hear continually repeated VDC plus 000.000 E dash 3.

Volts DC is the default power-up function. The zeros following the VDC are, as you would guess, telling you that there is no voltage impressed on the instrument's input terminals. Actually, there will probably be one or two significant digits following the decimal; however, since there is a negative exponent of dash three (minus 3), the value represents noise and jitter.

To change functions, just type on your Braille'n'Speak vac for volts a.c., rtw for resistance two wire, tdc for temperature degrees centigrade, or any other three-letter function select code. If you type con for continuity, the Braille'n'Speak will continually repeat "open" until you touch the probes together, at which time you hear "closed" and a beep directly from the Fluke.

There are lots of ways you can make the instrument behave the way you want. You can type dmp to get it to dump details of the current setup, including function, range, resolution, etc. If you don't want to hear VDC or other function identifier repeated before every reading, just type "out" to tell the Fluke to output numbers only. Of course with the Braille'n'Speak, as with some other terminals, you can set it to say the value of the number rather than pronounce each digit. If you don't want to have the Fluke continuously output measurements, you can type EMO followed by a number from 1 to 99, and the Fluke will oblige by reading just that many measurements.

There's more that could be said about this versatile precision instrument, but this is enough to get you started. No doubt you will want to know that the Fluke with the interface costs $995--whether you are blind or sighted. Discounts are available to educational institutions. Call me at (502) 695-2388 between 8:00 a.m. and 7:00 p.m. East Coast time. Calls at other times may elicit a petulant response devoid of information.

For more specifics on functions, refer to the Fluke catalog, free on request. For a sales representative in your area, call the John Fluke Manufacturing Co., (800) 443-5853.


Blind technicians get multimeters by two routes: One kind of product comes about by modifying commercial meters with either speech output or some sort of audio-tone output; Science Products has done both kinds. The other approach is to design meters from the ground up that have features particularly suited to blind technicians. The latter approach has led to some extraordinary instruments; Robert Gunderson and Dave Heavner designed the best tone-output ones in the 1950s and 1960s. In collaboration with AFB, the Franklin Institute of Philadelphia designed a talking meter in the early 1970s.

A sound reason for designing our own is that we blind technicians use our test instruments more than sighted technicians do; we use instruments to identify components, to find our way around circuits, and to troubleshoot our solder connections. (Sighted technicians don't use their meters for any of those jobs.) To "fill the bill," our meters need to be accurate sometimes (for identifying components and for making occasional static measurements), we need quick relative indicators to tell us what we've got (for identifying types of components and for finding our way around circuits), and we need an output that screeches when something changes (for locating bad solder connections and for noting the immediate effect of adjustments).

This author, your Editor, has had experience with the various auditory meters--going back to the late 1950s--and I have combined, in one instrument, features that answer our needs (and I'm not very modest about this job). A sensitive VCO designed for medical use was redesigned by Tom Fowle into a crackerjack dynamic readout. Susan Fowle's system of storing speech, coupled with Al Alden's low-cost delta modulator, gives us a talking readout that is faster than any before. Tom and Susan's work interfacing the Nattering RAM with the National Semiconductor voltmeter chip provides the basis for an accurate digital instrument. My own love of meter networks evolved a fancy layout of binding posts and switches that allows the meter to remain in two circuits at once, thus avoiding unnecessary disconnecting and reconnecting of probes.

I'm proud to say that the ideas used to create the Smith-Kettlewell Talk-&-Tones multimeter are all borrowed. The features are pilfered from the best of evolution. Using this meter fills me with boyish glee.


[It is only fair to say that the accuracy of this meter would be enhanced by using resistors of 0.1% tolerance in critical places. The 1% units specified here were chosen for their ready availability; the meter's resultant accuracy is perhaps within 3%. If you can get better ones, the resistors that would improve the meter are listed below.

Of course, the resistors on all the range switches should be as good as you can get. The calibration of the digital portion would be more straightforward if you used a 0.1% 1meg and a 0.1% 100K resistor off pin 12 of the National voltmeter chip. Finally, the accuracy of the ohmmeter would be improved if resistors associated with IC2, the LM358, were improved: 0.1% resistors would serve well for the 1K unit off pin 7 of IC2 and the 20K resistor from pin 6 to ground.

In the voltage and current functions, both DC and AC measurements can be made. AC measurement taken by our instrument gives you either the "peak" or the "average" of the positive-going half cycle. (It is the author's feeling that sinusoids are a rare case. Except for measuring the outputs of transformers, most measurements nowadays will be of digital signal lines, and a phony RMS reading derived from an averaging circuit--which is how most meters do it--obscures valuable information.) The RMS value of a sinewave can be calculated by multiplying the meter's half-wave average reading by 2.22. A spare op-amp in the meter will allow the builder to include that RMS scaling as a permanent feature, but this author doesn't want it.

As for frequency response, measured with a sinusoidal input, the "averaging" circuit starts dropping off at about 700 Hz, while the "peak-reading" circuit is good up to perhaps 2500 Hz.

Voltmeter Input Network

The voltmeter in this instrument has five ranges; the lowest is 0 to 199.9 millivolts, and the highest is 1999 volts. The input resistance is 11 megohms on all ranges except the lowest one. In the 200mV position, the input goes directly to the voltmeter chip and the input resistance can be considered infinite (although there is a healthy 0.47uF capacitor on the far side of 100K while in the DC mode).

When the "Voltage" mode is not selected, the input network is floating, so it can remain connected to the circuit under test while other functions of the instrument are being used.

Protection on the National voltmeter chip allows up to 200 volts to appear at its inputs without damage. You are only given direct access to this chip when on the lowest range; otherwise, a 10-megohm resistor protects the unit even more in higher voltage positions. It is conceivable that the National chip would survive application of 2000 volts while selecting the 2-volt range. The National chip's protection may not extend to the CA3240 op-amp, IC4, of the tone readout; however, this op-amp is garden-variety, and if a socket is used, it will be readily replaceable.

In AC, a CA3130 sees the input network, not the National chip. Although this op-amp has input protection, it is not known how it will fare in a drastic over-voltage situation. Its input has a 100K resistor in series with it, so it will surely survive being off two ranges or so.

The 200mV range has its own binding post, separate from the main network. This assures that 2000 volts at the main binding posts will never get to the solid-state circuit; the technician would have to both select the lowest range and select the wrong binding post in order to misapply high voltage.

Current Input Network

The current meter in this instrument has six ranges; the lowest is 0 to 19.99 microamps, and the highest is 0 to 1.999 amps. The resistance of the current meter is unusually low; the 2-amp range puts 0.1 ohms in series with the test circuit, and the 20-microamp position presents the test circuit with 10K.

The high-current shunt is a 2-watt resistor, so it is good for a steady-state current of 4 amps. However, the switch we used is only good for 3 amps; you could choose a healthier switch. Anyhow, a 5-amp quick-blow fuse is provided.

When the "Current Mode" is not selected, this network is floating, so the test circuit can be run through it at all times while other functions of the instrument are being used. Caution! The voltage on the circuit under current measurement should not be more than perhaps 500 volts over other connection points to the meter. Current measurement of high-voltage circuits would require use of "double-insulating" techniques in constructing the Talk-&-Tones.

Ohmmeter Section

The ohmmeter in this instrument has five ranges; the lowest one measures up to 199.9 ohms, and the highest one reads up to 1.999 megohms. The current source supplying the unknown resistor applies only 1 milliamp on the 200-ohm range, 100 microamps on the 2K-ohm range, and so on down to 100 nanoamps on the 2-megohm range.

When the resistance function is selected, the current and voltage input networks are floating, so an in-circuit resistance measurement is possible without disconnecting other probes from the circuit. By the same token, the meter can be used to measure an external resistor at any time without disturbing other metering connections. However, since the ohmmeter's negative binding post is common to "ground" of the internal circuit, the ohmmeter probes must be free of a live circuit before selecting voltage or current functions.

Speech Readout

The speech system used is the addressable Nattering RAM (see SKTF, Winter 1989). The speech is "recorded" by the builder or the user in any language. As implemented here, the instrument speaks thirteen words: "zero" through "nine," "plus," "minus," and some catchy word for "overflow" (the author's meter barks).

Though the readout doesn't insert the word "point" before a decimal fraction, all the digits of the decimal fraction are raised in pitch. For this reason, the numbers should be "sung" at a constant pitch by the programmer in order that the shift in pitch be easily hearable. (Proper placement of the decimal point is set by extra poles on the range switches.)

Speech can be elicited by pressing a button, or a toggle switch can be set so that the readout speaks over and over again automatically. Another toggle switch suppresses all leading zeros; the numbers spoken in this mode will begin at the first non-zero figure. The circuit shows four toggle switches which can be installed to provide hard suppression of any selected digit, although we did not include these in our instrument.

Most important of all is that each reading is "latched" before it is spoken. There is no possibility of the meter changing its mind in the middle of a reading and speaking an erroneous value. (This has been a problem with some digital instrument implementations. An example of the problem would arise when measuring a 39-ohm resistor whose value is so close to 40 ohms that the A/D converter reads those two values alternately. If the spoken reading were allowed to change in midstream, erroneous readings of 49 and 30 would be commonplace.)

Tone Readout

The VCO-type tone-output system provides for dynamic readings, and serves as an aid in making adjustments. A full-scale reading causes an oscillation of about 2kHz. Past full-scale, this tone will continue to rise to about 5kHz corresponding to 2.5 times the input range.

Working as an advantage for the ohmmeter, the oscillator quits at grossly over-range inputs (25 times the selected range). When the ohmmeter probes are open, the oscillator chirps and dies. Applying the probes to a resistor within the selected range causes the oscillator to chirp and become active again. Thus, the tone output can serve as an audible continuity tester whose frequency is a relative indicator of the resistance value. It is backwards from other continuity testers in that the higher the resistance, the higher the pitch of the tone.

The tone output is also fitted with Tom Fowle's Braille-calibrated potentiometer and pulsator system. A precision linear pot, fitted with a pointer knob and Braille scale, forms a calibrated voltage standard with which the unknown at the input of the National voltmeter chip is compared. Unknown voltages above the setting of the pointer knob will cause the VCO to pulse on and off about 15 times per second. Even when pulsating, the pitch of the VCO is easy to identify, so its "relative indication" never gets lost.

A "Mute" switch is provided to silence the VCO if only speech is desired. In consideration of the speech section, a transistor at the output of the VCO inserts 10dB of attenuation when speech is invoked. The VCO can still be heard during speech, but it is not loud enough to impair intelligibility.

The Braille dial can be used in two ways: First, it can be preset to a desired reading; the user will know when that reading has been exceeded by the fact that the tone begins pulsating. Second, readings can be taken with this system; the dial can be manipulated in search of the "boundary" between a continuous versus pulsating VCO tone.Since the ranges of the meter are all 2 times some power of 10 (0 to 2 volts, 0 to 20 volts, etc.), it is logical to make the Braille dial with twenty divisions. Triple dots mark zero, mid-scale, and full-scale. Considering the 20-volt range as an example, single dots mark 1V, 3V, etc.; double dots mark 2V, 4V, etc. As a frill, widely spaced 2-dot marks--resembling the Braille letter K--are used at 1/4-scale and 3/4-scale (corresponding to 5V and 15V in our 20V example).The tone readout is a wonderful convenience when selecting proper ranges. With the pointer knob set to full-scale, the lowest range that produces a continuous tone is selected; a setting lower than this will make the VCO emit a pulsating tone.Finally, as could be done with a moving needle, you can use the tone readout to make fairly accurate measurements of undulating unknowns. For example, you can use the averaging AC ammeter to examine the battery drain of an audio power amplifier which is driving a speaker. A few test settings on the Braille dial will result in a place at which maxima of the undulating current barely causes the readout to pulsate. This kind of measurement is simply not feasible with a talking readout. Besides the slowness of speech getting in the way, you have no clear indication of when the D/A conversion takes place; it is not necessarily done at the instant the "Talk" button is pressed.

Physical Description

The Talk-&-Tones is built into an aluminum chassis measuring 8 by 12 by 3 inches. All the controls are on the 8- by 12-inch top surface, so you don't have to bend your wrists to operate them.

The wire-wrapped circuit boards take up a 6-1/2-inch square space at the left rear; this leaves a 1-1/2-inch strip along the front on which all the switches are mounted and a large panel space at the right in which the Braille dial is centered. A 1-3/4-inch by 5-1/2-inch strip at the right rear is reserved for the binding posts, as well as providing mounting space for a couple of tie strips (insulated terminal lugs on phenolic strips) used to carry the ohmmeter's FET and the free end of the 10-megohm resistor in the voltmeter network.

The speaker is mounted on the right-hand apron of the chassis. On the front apron, the earphone and remote jacks are available. These jacks are insulated from the chassis; they are mounted on a small plastic panel, and an oblong hole is cut in the chassis to allow ample clearance around them.

Holes on the rear and left-hand aprons allow access to all the trim pots, even those on the Nattering RAM board.


The "Read" button is at the left front corner. Just above it is the "Auto-Read" toggle. Beside the button is the "Leading Zero Suppression" toggle switch. The next item--moving to the right--is the "On-Off/Volume" control.

The right-hand two-thirds of this strip across the front is taken up by all the meter function and range switches. To the right of the volume control is the 3-position "AC/DC/Ohms" rotary switch. The next three knobs across the front are: "Ammeter Range," "Voltmeter Range," and "Ohmmeter Range" switches. The "Amps/Volts" and "Average/Peak" toggle switches are positioned above, and centered between, the range switches. Finally, somewhere near the right edge, is the "Mute" switch that silences the tone readout.

Binding Posts

As mentioned, these are at the rear edge of the top panel above the Braille dial. Three-fourths of an inch from the right apron, mounting holes are drilled for a 4-lug tie strip that carries the ohmmeter's FET and gate resistor.

Moving to the left, the ohmmeter binding posts are mounted--one above the other, with the one toward you being the negative binding post. (To accommodate standard banana-plug assemblies, the spacing between "centers" of these posts is exactly 3/4-inch.)

Two inches to the left of the ohmmeter terminals is an array of three binding posts for the voltmeter. The positive one is farthest from you with the negative one just below it. To the left of the negative binding post is the 200mV positive post. (The resultant configuration looks like a backwards L, with exactly 3/4-inch spacing between the center of the negative binding post and the centers of each of its neighbors.)

Finally, two inches to the left of the main voltmeter terminals are the ammeter ones. Mounted vertically, the negative binding post is toward you. (To accommodate standard banana-plug assemblies, the spacing is exactly at 3/4-inch centers.)

As room permits, there are three more items which logically belong in this section: At the back-right corner, there is a binding post mounted without insulator to the chassis; it is called "Chassis Ground." In case you are in a high-noise environment, grounding the chassis to an actual "earth ground" may help. Between the positive posts for voltage and current is the fuse holder for the current network. Finally, between the ohmmeter and voltmeter terminals is a banana jack that goes to the precision 2V output; this test point is provided so that calibration can be done entirely from the outside.

Construction and Board Layout

The switches, controls and binding posts are mounted in the usual way. However, as a safety measure, the first resistor (the 10meg unit) in the voltage multiplier string is run from the main positive binding post over to a nearby insulated terminal lug--an insulated tie point mounted between the "Volts" and "Ohms" binding posts. Thus, if 2000 volts is being applied, there will be a 10-megohm resistor at all times between it and circuitry in the meter. This positive binding post for the voltmeter is kept clear of other wires.

Because the National voltmeter chip is so fussy about keeping its "V-In plus" and "V-In minus" pins floating, it was decided not to "ground" the chassis to any part of the circuit. Thus, the foot-pedal and earphone jacks were insulated by mounting them on a square of plastic which, in turn, is held by two screws to the chassis. A rectangular or oval hole in the front panel of the chassis permits about 1/8-inch clearance around the mounting nuts of these jacks.

The memory-backup and ohmmeter bias batteries are expected to last a long time, so they were made by soldering assemblies of AAA cells to twistedpair leads from the circuit boards. (Mark the positive lead with a knot or something.) These batteries (both 4.5V) could be strapped to the back panel, but we just stuck them to the chassis with double-sided foam tape.

On the circuit boards, the wire-wrap sockets used have short pins; they are said to be of the "two-wrap type" instead of accommodating three wraps on each pin. By choosing these, the boards can be mounted in a sandwich with 1-inch spacers between. (The pins of the ribbon connector were cut short to match the length of the wire-wrap pins of the sockets.)

To permit access to the Nattering RAM (in case the recording is to be changed), the Nattering RAM board is the one seen when the bottom cover plate of the chassis is removed. In other words, with the chassis upside-down, the meter board is mounted (wiring side down) on 5/8-inch spacers; then, perched atop 1-inch spacers that are carried by the meter board is the Nattering RAM board. To keep wire-wrap pins of the Nattering RAM from damaging wires on the board underneath, a piece of Braille paper is fashioned to fit between them.

All of the trim pots for calibration and offset correction are mounted so as to be accessible from the top edge of the meter board; this allows access to them through holes in the rear panel of the chassis. (Given the layout of the Nattering RAM, see SKTF, Winter 1989, the delta modulator's bias adjustment can be accessed through a hole in the rear panel, while the clock speed peeks through a hole in the side--the left end panel when the meter is sitting upright.)

Layout of the Meter Board

The meter board contains the chips listed in Table II. It is different from the Nattering RAM board whose details, including layout, are found in "The Addressable Nattering RAM," SKTF, Winter 1989.

The board is a generous 6-inch square (cut along the 61st row and 61st column, so you will have 60 good holes in either direction). The top 3 inches are set aside for all the soldered components; the bottom half is where all the wire-wrapped ICs are placed.

Along the 29th row from the top, a plus 5V bus is run which adjoins a vertical 5V bus along one edge--the left edge with the wiring side up. Running across the board in the 31st row is a ground bus which adjoins a vertical one on the opposite edge. Across the top, in the first row of holes, is a 9V bus. (This arrangement of buses was done to match that of the Nattering RAM.) From now on in this description, we will state placement of items as viewed with the component side up. Therefore, with the wiring side down, there is a 5V bus under the right edge, a ground bus under the left edge, and so on.

Wire-Wrap Section

The chips are arranged in five columns. The spacing between chip columns leaves seven columns of holes. There should be five or six columns of free holes at the left and right edges of the board. The top chip in each column is two or three holes toward you from the horizontal ground bus. Leave two blank horizontal rows between chips within a column.

The right-most column has only one chip in it; IC7, the CD4017, is just under the horizontal ground bus. Moving to the left, the second column has two chips in it: the 74C193, IC8, is in the top position; while the first CD4013, IC12, is below it. The middle column has only one chip in it, the first 74C221, IC9. Below this is the ribbon connector for hooking the two boards together. The fourth column from the right has two chips in it; the upper one is the second 74C221, IC10, while the lower one is the other CD4013, IC13. The left-most column has IC11, the 74HC157 (or 74HC257), in the top position, and the CD4072, IC14, is below it.

Before you accuse me of random behavior, let me point out the order in which these digital chips fall if you go from right to left. Considering just the top-most chips in the wire-wrapped section, from right to left, they are: IC7, IC8, IC9, IC10, and IC11. Seen from right to left across the bottom, we find: a blank space, IC12, the ribbon connector, IC13, and IC14.

The ribbon connector is oriented horizontally, so that the ribbon cable can emanate from the bottom edge. A tie strip of terminal lugs is mounted on the component side along the very bottom edge, whereupon it is then bent toward you so that solder connections can be made to the lugs after the "sandwich" of boards has been assembled. This strip of tie points carries the "digit flags" from the CD4017, and stranded wires run from these lugs to the various range switches.

The trim pot for the auto-read timing is just to the left of IC11, vertically oriented at the left edge of the board. This is the only trim pot which cannot be accessed through a hole in a panel of the chassis.

Point-to-Point Wiring Section

About six holes from the right edge and three holes down from the top, the calibration trim pot for the Braille pot is placed. Below this is a column of two 8-pin chips; the upper one (just under the pot) is IC2, the LM358, while IC3, the first 3240, is just below this. IC1, the voltmeter chip, is located to the left of these (with perhaps nine blank columns left between). Directly above IC1 are two trimmers, the precision 2V adjustment and IC1's offset correction pot. With this spacing, there will be room enough to put the LM336 and all of its paraphernalia between the first trimmer and the latter two.

Leaving nine blank columns to the left of IC1, IC6, the 7556 tone-output chip, is placed near the top edge of the board; IC5, the associated 3240, is just below IC6. Having this near the top edge leaves room for all those crazy resistors and large capacitors associated with IC1. To the left of ICs 5 and 6, you can place the two transistors used by the tone-output system.

Finally, two more trim pots--the offset corrections for the two CA3130s--are placed next to each other, six holes to the right of the left edge. Below these is IC15, the follower for the ohmmeter, and below that is IC4, with plenty of room to its right for the large sampling capacitor (and some of the larger parts that IC1 calls for).

Recording the Speech in the Nattering RAM

The advantage of using a detachable ribbon cable is that the meter board can be unplugged so as not to prevent the Nattering RAM from responding to its own start button. Also, any external circuitry connected to the Nattering RAM stands a good chance of shifting its pitch (as if a decimal fraction were being presented), and this will create problems with the recording. Therefore, don't forget to unplug it.

The clock of the Nattering RAM should be set for a message length that will accommodate the longest words ("minus" and "seven" being the longest English words in the character set). If you can speak at a good clip, a third of a second should be enough message time.

The delta modulator bias should be adjusted for minimum noise before recording is done. With the microphone unplugged, repeatedly push the button while turning the 50K trim pot until a point of near silence is reached. Grounding pin 4 of the 4013 in the Nattering RAM will cause it to run continuously, if you like.

Once the recording has been completed, the clock speed can be turned up to make the speech faster. This will make it sound like you've been breathing helium, but you will want the speech to be as fast as you can get it.

The following table lists the addresses and their associated words:

Table I

Word Addresses

  • 0000--"zero"
  • 0001--"one"
  • 0010--"two"
  • 0011--"three"
  • 0100--"four"
  • 0101--"five"
  • 0110--"six"
  • 0111--"seven"
  • 1000--"eight"
  • 1001--"nine"
  • 1101--"minus"
  • 1110--"arf"
  • 1111--"plus"

Table II

Reference List of ICs for the Multimeter Board

Note: No IC designation numbers have been assigned to chips of the Nattering RAM. For all its complexity, there are no duplicate chips in the Nattering RAM's circuit. However, there are duplicates in the multimeter circuit; rather than calling a given CA3240 "the second 3240" and a package of one-shots "the first 74C221," each chip on this board has a number like IC5 and IC9.

  • IC1--ADC3511 Voltmeter
  • IC2--LM358 Dual Op-Amp for Reference Voltages
  • IC3--CA3240 Dual Op-Amp for Ohmmeter and Optional RMS AC
  • IC4--CA3130 Single Op-Amp for AC Peak/Average Circuit
  • IC5--CA3240 Dual Op-Amp for Tone-Output Section
  • IC6--ICM7556 Dual Timer for Tone-Output Section
  • IC7--CD4017 One-of-Ten Decimal Counter Providing "digit flags" for "Point" Information
  • IC8--74C193 Programmable Up/Down Counter for Addressing IC1
  • IC9--"First" 74C221 Dual One-Shot for Starting A/D Conversion, and for Starting the Nattering RAM
  • IC10--"Second" 74C221 for Clocking the 74C193, and for Auto-Read Function
  • IC11--74HC257 (or 74HC157) Multiplexer for BCD/Sign Readings
  • IC12--"First" CD4013 Dual D Flip-Flop for Initiating Sign Information, and Containing the "Stop" Flip-Flop
  • IC13--"Second" CD4013 Dual D Flip-Flop for Shifting Pitch of Voice in Accordance with Digit Flags, and for Suppression of Leading Zeros
  • IC14--CD4072 Dual 4-Input OR Gate for Detecting Zeros, and for Suppression of Unwanted Zeros
  • IC15--Second CA3130 Op-amp Used as a Follower in the Ohmmeter

(Notice that last number is a little out of place? Well, that's how trips "back to the drawing board" can make themselves known.)

Multimeter Circuitry

Power Supplies and Standard Voltages

Except for the fact that DC voltage and DC current binding posts must be floating, the "circuit common" for this section is shared with "ground" of the Nattering RAM. Two positive supply lines also come from the Nattering Ram, the 5V logic VCC and the unregulated 7.5V to 12V line (which will be called the 9V line from here on). Each of these two lines has a 10uF bypass (negative of these caps at ground).

The ADC3511 voltmeter chip (designated IC1) has its pins 13 and 22 grounded. Pin 1 goes to 5V. Pin 2 is bypassed to ground by 10uF (negative of the cap at ground).

An LM336 precision 2.5V standard has its "anode" grounded. Its "cathode" goes through 680 ohms to the 5V line. A 10K trim pot has its arm going to the 336's "adjust" lead. One end of this pot goes to the anode of a 1N914 diode; the cathode of this diode is grounded. The other end of this pot goes to the cathode of another 1N914; the anode of this goes to the junction of the 336's "cathode" and the 680-ohm resistor. The LM336's "cathode" is the precision 2.5V point.

IC1, the 3511, requires a corrective offset. The 2.5V point goes through 100K, then through a 50K trim pot to ground. The arm of this pot goes through 22 megohms to pin 12 of IC1.

A voltage divider is used to establish 2 volts. The 2.5V point goes through a 25K resistor (made up of four 100K 1% units in parallel), then through a 100K 1% resistor to ground. The junction of these resistors is bypassed to ground by a 1uF tantalum electrolytic (negative at ground).

The above voltage divider is buffered by an op-amp in IC2, an LM358 dual op-amp. A1 in this IC2 package is made into a follower by tying pins 1 and 2 together. Pin 3 goes to the junction of the contrived 25K resistor and its 100K companion. Pin 1 of IC2 is the precision 2V standard; this goes directly to pin 16 of IC1 (the 3511), as well as going to a banana jack test point (see text).

Pin 4 of IC2, the 358, is grounded. Pin 8 goes to the 9V line.

A2 of IC2 is used to create a 5.1V supply and an elevated 0.100V standard for the ohmmeter circuit. Pin 5 of IC2 goes to pin 1, the 2V standard. The inverting input of A2, pin 6, goes through a 20K resistor to ground (made from two 10K 1% units in series). The output, pin 7, goes through a 1K 1% resistor, then through a 30K 5% unit back to pin 6. Pin 7, the output of A2, is an approximate 5.1V source; the junction of the 1K and 30K resistors is a minus 100mV precision standard hanging from the 5.1V.!

Clock for IC1

On IC1 (the voltmeter chip), there is a 10K 5% resistor going between pins 17 and 18. Pin 17 also goes through 220pF to ground.

Setup for a 200mV Range on the Metering Circuitry

Pins 14 and 15 of the ADC3511 are tied together and go through a 1 megohm 1% resistor to pin 12. Pin 12 goes through a 111,110 ohm resistor to ground; this is made up of the series string of 100K (1%) plus 10K (1%) plus 1K (1%) plus a 110 ohm 5% unit. Pin 12 also goes through 0.47uF (good-quality Mylar) to ground.

Tone Readout System

IC5, a CA3240 dual op-amp, and IC6, a CMOS Intersil ICM7556 dual timer, are used. On the 7556, pin 7 is grounded. Pins 10 and 14 are tied together and go through 220 ohms to the plus 9V line. Pin 14 is bypassed to ground by 10uF (negative side of this cap at ground). On IC5, the 3240, pin 4 is grounded, while pin 8 goes to plus 9V.

Pins 3 and 5 of IC5, both non-inverting inputs, are tied together and go to the "V-filter" pin of IC1, pin 9 of the 3511.

A 10K linear Braille-calibrated pot has its counterclockwise end grounded. Its clockwise end goes through a 100K rheostat, then through a 10K 1% resistor to pin 1 of IC2, the precision 2V point. The arm of the Brailled pot goes to pin 6 of IC5, the inverting input of A2.

Pin 2, the inverting input of A1, goes to the emitter of a 2N2222; pin 2 also goes through 4.7K to ground. The output of A1, pin 1, goes to the base.

The collector of the 2N2222 goes through 22K, then through 10K to pins 2 and 6 of IC6, the 7556. Pins 2 and 6, which are tied together, also go through 0.0047uF to ground. The output of the 7556, pin 5, goes to the anode of a 1N914 diode; the cathode of this goes to the junction of the 22K and 10K resistors.

Pins 8 and 12 of IC6 are tied together and go through 2.2uF to ground (negative end of the cap at ground). Pins 8 and 12 go through 10K to pin 13. Pins 8 and 12 also go through 22K to pin 7 of IC5, the output of A2. Pin 9 of the 7556 goes to pin 4 on the same chip, thus connecting the output of one section to the "Enable" of the other.

The output of the audio oscillator, pin 5 of the 7556, goes through 100K, then through 470K to the junction of the two 22K resistors off the top of the volume control; i.e., audio from the 7556 is coupled into the middle of the audio filter in the Nattering RAM. An SPST "mute" switch shorts out this audio; one side of the switch is grounded, while the other switch terminal goes to the junction of the 100K and 470K resistors.

The junction of those coupling resistors and the switch also goes through 47K to the collector of another 2N2222. The emitter of this 2222 is grounded. Its base goes through 47K to the "Word-Active" line of the Nattering RAM (which can be found on this board at IC1 pin 19, IC9 pin 9, and IC10 pin 2).

Ohmmeter Circuit

Because it is correctable for offset, a CA3130, IC15, is used as a follower. Pin 4 of this 3130 is grounded; pin 7 goes to plus 9V. Between pins 1 and 8 is a 100pF disc. One end of a 100K trim pot goes to pin 1, while the other end of the pot goes to pin 5. The arm of this pot is grounded.

The output and inverting input, pins 6 and 2, are tied together and go to position 3 on pole C of the "AC/DC/Ohms" selector (the ohms position). Pin 3, the non-inverting input, goes to the positive ohmmeter binding post.

The negative binding post is grounded. The positive one goes to the drain of a P-channel JFET (Siliconix J176).

One half of a dual op-amp, A2 in a CA3240 (IC3), is used as well. Pin 4 is grounded, while pin 8 goes to the plus 9V line. If the meter is not made to read RMS, pins 1 and 2 are tied together. Pin 3 is grounded.

The JFET and A2 of IC3 comprise a current source. The non-inverting input, pin 5of the 3240, goes to the junction of the 1K and 30K feedback resistors on A2 of IC2 (the elevated minus 100mV standard). Pin 7 of IC3, the output of A2, goes to the negative side of a 4.5V battery; the positive of this battery goes through 47K to the gate of the JFET.

Between pins 6 and 7 of IC3 is 100pF. Pin 6, the inverting input, also goes to the source of the JFET. The source goes through the switchable standards to 5.1V as follows:

The arm of pole A of the "ohms range" switch goes to pin 7 of IC2 (5.1V). Each position goes to one end of a 1% resistor as follows: Position 1 has a 100-ohm resistor, position 2 has 1K, position 3 has 10K, position 4 has 100K, and position 5 has 1 megohm. The free ends of all these resistors are tied together and go to the source of the JFET.

AC Rectifier Circuits

A CA3130 single op-amp, IC4, is used. Pin 4 is grounded, while pin 7 goes to plus 9V. Between pins 1 and 8 of IC4 is a 30pF disc. Pin 1 also goes to one end of a 100K trim pot, the other end of which goes to pin 5. The arm of this pot is grounded.

Pin 3 of IC4, the non-inverting input, goes through 100K (5%) to position 1 (the "AC" position) of pole A on the AC/DC/Ohms selector switch. Position 1 of pole B is grounded.

Peak/Average AC Input Circuit

Pin 6 of IC4, the op-amp's output, goes to the anode of a 1N914 diode. The cathode of this goes to pin 2, the inverting input. Pin 2 goes through 10 megohms (5%) to ground; pin 2 also goes through 100K (5%) to the top of the "storage capacitor," with the bottom end of this cap (0.47uF Mylar) being grounded. On the AC/DC/Ohms switch, position 1 of pole C goes to the top of the storage capacitor.

Both arms of the "peak/average" selector switch (DPDT) are tied together and go to pin 2 of IC4. The "peak" position of pole A goes to the junction of the 100K and the storage cap. The "average" position of pole B goes to the anode of the diode and to pin 6 of IC4.

RMS-Reading Circuit

To simplify switching, the "Peak/Average" switch is eliminated. Pins 6 and 2 of IC4, the output and inverting input, are tied together. Pins 6 and 2 go through 10 megohms (5%) to ground; pins 6 and 2 also go through 100K (5%) to the top of the "storage capacitor," with the bottom end of this cap (0.47uF Mylar) being grounded. The top of this storage cap goes to pin 3 of IC3, the non-inverting input of A1 in this CA3240.

A1 of IC3 is set up to have a gain of 2.2 as follows: Pin 1, the output, goes through a 12K 1% feedback resistor to pin 2, the inverting input; pin 2 also goes through a 10K 1% resistor to ground. The output, pin 1, goes to position 1 of pole C on the AC/DC/Ohms switch.

Finishing Up the Input Networks

On IC1, the 3511 voltmeter, pin 9 (its "V-filter" pin) is bypassed to ground by a 0.47uF Mylar capacitor (this cap matching, in brand, the one on pin 12). This pin 9 also goes through a 100K 5% resistor to the arm of pole C on the AC/DC/Ohms selector.

Pin 10 (the 3511's "V-in minus") goes through 51K to position 2 of pole B on the AC/DC/Ohms selector. Pin 11 ("V-in plus") goes through 51K to position 2 of pole A. (Position 2 is the DC position.) On both poles A and B, position 3 is left blank. The arm of pole A of this AC/DC/Ohms selector goes to the arm of pole A on the "Volts/Amps" switch (a 3-pole double-throw toggle). The arm of pole B of the AC/DC/Ohms selector goes to the arm of pole B on the Volts/Amps switch.

Current Range Switch

A 2-pole 6-position switch good for at least 3 amps is required. On pole A, each position goes to one end of a precision resistor as follows: Position 1 has a 10K resistor going to it, position 2 has 1K, position 3 has 100 ohms, position 4 has 10 ohms, position 5 has 1 ohm, and position 6 is connected by a heavy lead to a 0.1 ohm 2-watt resistor. The far ends of these resistors are tied together and are connected via a heavy lead to the negative binding post. The arm of pole A goes through a heavy lead to a 5-amp quick-blow fuse, with the other side of this fuse going to the positive binding post.

The arm of pole A on the above range switch also goes through a light wire to the "amps" position of pole A on the Volts/Amps switch. The "amps" position of pole B goes to the negative binding post (via a light wire).

Voltage Range Switch

In theory, the input terminals are shunted by a string of resistors totaling 11,111,111 ohms; the range is selected by choosing a tap on this divider. In practice, the circuit is slightly variant (see text). (Because there is a 2000-volt position, the layout is critical, and should be done in accordance with the text.)

Three binding posts are used: a low-voltage positive, a main positive, and a negative common to both. A 2-pole 5-position switch is required. (We used a 2-pole 6-position switch--see text.)

The low-voltage (200mV) binding post goes to position 1 of pole A on this switch. The main positive binding post goes through a 10meg 1% resistor to position 2 of pole A. Position 2 goes through 1meg (1%) to position 3, which goes through 100K (1%) to position 4, which goes through 10K to position 5. Position 5 goes through the series combination of: a 1K 1% resistor, a 100-ohm 1% resistor, then an 11-ohm 5% resistor; the far end of this string goes to the negative binding post. (If your switch has a position 6, it can go to the junction of this string and the negative binding post.)

The arm of pole A on the above range switch goes to the "Volts" position on pole A of the "Volt/Amp" switch. The "volts" position of pole B on the Volt/Amp switch goes to the negative binding post.

Decimal Point Selection

The arm of pole D of the AC/DC/ohms switch goes to pin 6, the "Set" input, of IC13 (a CD4013 appearing in the wire-wrap table to follow). Positions 1 and 2 of pole D are tied together and go to the arm of pole C on the volts/amps switch.

The "amps" position of pole C on the volts/amps toggle goes to the arm of pole B on the current-range switch. The "volts" position of pole C on the toggle goes to the arm of pole B on the voltage-range switch. The arm of pole B on the resistance-range switch goes to position 3 of pole D on the AC/DC/ohms selector.

The various positions of the "B" poles on the range switches go to outputs of the CD4017 digit counter (IC7), and are tabulated as follows:

Table III

Assignment of Digit Flags

[Note: The actual output labels on the 4017 refer to the clock pulses that trigger them; i.e., the "1" output goes high with the first pulse (coincident with the plus/minus sign); the "2" output comes up with the second 4017 clock pulse, which happens to be the first spoken digit. Therefore, in this table, I have "relabeled" the 4017 pins 4, 7, 10, and 1 as "digit flags" 1, 2, 3, and 4, respectively (instead of listing the proper output labels "2," "3," "4," and "5," respectively).]

Current Range Switch--Positions of Pole B:

  • Position 1 to Digit Flag 3; 4017-10 (reads 19.99uA full-scale)
  • Position 2 to Digit Flag 4; 4017-1 (reads 199.9uA full-scale)
  • Position 3 to Digit Flag 2; 4017-7 (reads 1.999mA full-scale)
  • Position 4 to Digit Flag 3; 4017-10 (reads 19.99mA full-scale)
  • Position 5 to Digit Flag 4; 4017-1 (reads 199.9mA full-scale)
  • Position 6 to Digit Flag 2; 4017-7 (reads 1.999 amps)
  • Voltage Range Switch--Positions of Pole B:
  • Position 1 to Digit Flag 4; 4017-1 (reads 199.9mV full-scale)
  • Position 2 to Digit Flag 2; 4017-7 (reads 1.999V full-scale)
  • Position 3 to Digit Flag 3; 4017-10 (reads 19.99V full-scale)
  • Position 4 to Digit Flag 4; 4017-1 (reads 199.9V full-scale)
  • Position 5 left open (reads 1999V full-scale)
  • Resistance Range Switch--Positions of Pole B:
  • Position 1 to Digit Flag 4; 4017-1 (reads 199.9 ohms full-scale)
  • Position 2 to Digit Flag 2; 4017-7 (reads 1.999K full-scale)
  • Position 3 to Digit Flag 3; 4017-10 (reads 19.99K full-scale)
  • Position 4 to Digit Flag 4; 4017-1 (reads 199.9K full-scale)
  • Position 5 to Digit Flag 2; 4017-7 (reads 1.999 megohms full-scale)

Table IV

Wire-Wrap Table for the Voltmeter Board

Note: We could have just given IC numbers for the chips--those numbers taken from the "Reference List of ICs." However, it is the editor's opinion that such a practice would separate the builder from the circuit and would promote errors. Thus, in most cases, I was content to leave the IC designation number off. But, there are two sets of duplicate chips in this wire-wrap table--two 74C221s and two CD4013s. For those, I preceded the actual part number with designations such as "IC9," or whichever.

IC1, the ADC3511 Voltmeter Chip

  • 3 to 74HC257-10
  • 4 to 74HC257-13
  • 5 NC
  • 6 to IC9 74C221-10
  • 7 to 4017-15; IC9 74C221-13; IC12 4013-4-10; IC13 4013-4
  • 8 to 74HC257-5
  • 20 to 74C193-3
  • 21 to 74C193-2
  • 23 to 74HC257-3
  • 24 to 74HC257-6

IC9, 74C221 Dual One-shot

  • 1 to IC10 74C221-5
  • 2 through 47K to VCC; to tip contact of "Read" foot-pedal jack; through "Read" button to ground. The sleeve of the jack is grounded.
  • 3 to IC12 4013-3; "NOT Word Active"
  • 4 to 74C193-11
  • 5 to the "Start Input" (on Nattering RAM)
  • 7 through 0.1uF to first 74C221-6; through 470K to VCC
  • 8 to ground
  • 9 to IC10 74C221-2; 3511-19; "Word Active"
  • 11 to IC12 4013-12
  • 12 NC
  • 15 through 0.2uF to first 74C221-14: through 470K to VCC
  • 16 to VCC

IC10, 74C221 Dual One-shot:

  • 1-8 to ground
  • 3-11-16 to VCC
  • 4 to 74C193-4; 4017-14
  • 7 through 1uF to second 74C221-6 (negative end at pin 6); through series combination of 330K and 1meg rheostat to VCC
  • 9 through 47K to VCC; through the "Auto-Read" toggle to ground
  • 10 to IC12 4013-13
  • 12-13 NC
  • 15 through 0.1uF to second 74C221-14; through 470K to VCC

IC8, 74C193 or CD40193 4-Bit Up/Down Counter:

  • 1-8-9-14-15 to ground
  • 5-10-16 to VCC
  • 6-12-13 NC
  • 7 to IC12 4013-8

IC12, CD4013 Dual D Flip-Flop (designated "first 4013" in earlier papers)

  • 1 to 74HC257-1
  • 2 to 4017-13
  • 5-14 to VCC
  • 6-7-9-11 to ground

IC11, 74HC257 or 74HC157 Multiplexer:

  • 2-11-14-16 to VCC
  • 4 to "'1' Address"
  • 7 to "'2' Address"
  • 9 to "'4' Address"
  • 12 to "'8' Address"
  • 8-15 to ground

IC13, CD4013 Dual D Flip-Flop:

  • 1-12 NC
  • 2 to "Pitch Shift" Input
  • 3-5-7-9-11 to ground
  • 6 through 47K to ground; triggered for decimal information (see "Decimal Point Selection" above)
  • 8 to 4017-2
  • 10 through "Leading-Zero Suppression" Toggle to VCC; through 47K to 4072-13
  • 13 to anode of a 1N914 whose cathode goes to the "Abort" line
  • 14 to VCC

IC14, CD4072 Dual 4-Input OR Gate:

  • 1 to the anode of a 1N914 whose cathode goes to the "Abort" line
  • 2 through 47K to ground; through SPST "least-significant" toggle to 4017-1
  • 3 through 47K to ground; through SPST toggle to 4017-10
  • 4 through 47K to ground; through SPST toggle to 4017-7
  • 5 through 47K to ground; through SPST "most-significant" toggle to 4017-4
  • 6-8 NC
  • 7 to ground
  • 9 to 74HC257-4
  • 10 to 74HC257-7
  • 11 to 74HC257-9
  • 12 to 74HC257-12
  • 14 to VCC

IC7, CD4017 Digit Counter

  • 1 ("5" Output) flags fourth digit
  • 5-6-9-11-12 NC
  • 3 ("0" Output) NC
  • 4 ("2" Output) flags first digit
  • 7 ("3" Output) flags second digit
  • 8 to ground
  • 10 ("4" Output) flags third digit

Review of Nattering RAM Terminals

Note: the part numbers referred to here are on the Nattering RAM board, not in the circuit of the above multimeter board. Both circuits have CD4013s and 74HC257s, so stay on your toes.

Nattering RAM Inputs:

  • "1" Address--74HC257-3
  • "2" Address--74HC257-6
  • "4" Address--74HC257-10
  • "8" Address--74HC257-13
  • Start Input--Anode of a diode whose cathode goes to 4013-6 (pull high to start; but speech begins after it goes low again)
  • Two Abort Inputs--Anodes of two diodes whose cathodes go to 4013-4 (pull high to abort)
  • Pitch Shift--Second 4053-9 (pull low to raise pitch)
  • Audio Tone Input--The junction of the 22K resistors off the top of the volume control
  • Nattering RAM Outputs
  • Word Active--4013-1 (goes high when Start line goes high; thereafter, stays high until word is complete)
  • NOT Word Active--4013-2 (goes low when Start line is brought high; thereafter, stays low until word is complete)


Unfortunately, to calibrate such a digital instrument, you need a meter which is more accurate than the one you are calibrating. You can find such meters in school electronics programs, and you can use students as meter readers.

There are two adjustments which affect the ADC3511, IC1. First, it requires a 2-volt standard; this can be measured between the banana test jack and the negative ohmmeter binding post. This is set by the 10K trim pot associated with the LM336. Second, offset of comparators in IC1 can be corrected for by adjusting the 50K trim pot whose arm goes through 22 megohms to pin 12. We chose to use this as a kind of "fine tuning" of the meter's basic calibration. We set the meter up for DC volts, applied 19.999 volts to its inputs, and adjusted the 50K trimmer to the point where half the readings were numbers and the other half were overflow barks. (There is always jitter that causes the meter to read a non-zero value, so you can't adjust this 50K trimmer for an output of zero.)

Using a variable power supply to check several points within a selected range, there may be some advantage to juggling these two adjustments slightly to achieve "good tracking" (a good correlation of readings throughout a given range). It is interesting to note that, since the offsets you are correcting follow the polarity sensor, you won't get plus or minus indications that are meaningful when adjusting the 50K trimmer.

With a known voltage still applied and with the Braille readout set to this value, the 100K trimmer off the top end of the Brailled pot is set to the point where the tone output is uncertain as to whether it should be pulsed or smooth.

Next, switch to AC and select the averaging circuit. Short the input terminals and adjust the 100K pot associated with IC4 until the meter reads zero--or near zero. Then, apply the unknown voltage and switch back and forth between AC and DC to see if the "average" of the DC signal agrees in value with the direct DC reading. If not, readjust the trimmer until they do agree, and then test to see if a shorted input still reads zero. You may end up striking a compromise, but ours worked out just fine, even near the lower end of the range.

Finally, select the resistance function. Short the ohmmeter terminals and select one of the middle ranges--20K or 200K. Adjust the trimmer associated with IC15 so that the meter reads zero.

Notes on Measuring AC Waveforms

On AC, this meter is either measuring the peak amplitude reached by a positive excursion, or the arithmetic mean (the "DC level") of the positive-going portion on a waveform. Remember that the AC circuit is a half-wave rectifier, so only positive-going excursions are being noted. Half-wave rectification is a good feature; you can look at each "half" of a waveform separately by changing the polarity of the probes.

Rectangular Waveforms

By looking at the ratio between peak and average readings, you can tell what the duty cycle is. In the extreme case where the duty cycle is 100%--a DC voltage--the peak and average readings are the same. Where the squarewave is positive-going for half the time, the average reading will be half the peak. If narrow pulses are being metered, the ratio of peak to average will be the ratio of "on time" to "total time" (the width of the pulse divided by the period).

Triangles and Sawteeth

Taken over the whole wave shape, the arithmetic mean of a triangle is half its height. If you are sure that you have such a waveform, and if you are confident that its sides are straight, you can calculate the percentage that is positive-going; i.e., if the average is 1/4th the peak value, only half of this waveform is positive-going and the other half is below 0 volts. If the waveform were entirely positive-going and if it filled its period, the average reading would be half of the peak reading.

Other Waveforms

If the sides of a waveform are not straight--if you have a sawtooth whose slope is noticeably exponential--the relation of average to peak readings gets more complicated. If the waveform is known to be "complete" (positive-going and such as to fill the period), you can make a guess as to what you have. If the average is more than half the peak, you may have a convex rising curve (such as that from a capacitor charging); if the average is less than half the peak, you may have a concave descending curve (such as the discharge of a capacitor).

Parts List

Resistors (1/4-watt 5%):

  • 1--11 ohm
  • 1--110 ohm
  • 1--220 ohm
  • 1--680 ohm
  • 1--4.7K
  • 3--10K
  • 2--22K
  • 1--30K
  • 11--47K
  • 2--51K
  • 5--100K
  • 1--330K
  • 4--470K
  • 1--10 megohms
  • 1--22 megohms

Resistors (1%, low power rating unless otherwise specified):

  • 1--0.1 ohms 2-watt (Newark 46F7583-.1, a 3-watt unit)
  • 1--1 ohm 2-watt (Newark 46F7584-1, a 3-watt unit)
  • 1--10 ohm 1/2-watt (Newark 46F7581-10, a 1-watt unit)
  • 4--100 ohm
  • 5--1K
  • 8--10K
  • 1--12K (optional)
  • 8--100K
  • 3--1 megohm
  • 1--10 megohm


  • 1--10K semi-precision linear, such as the Clarostat 58C1-10K
  • 1--10K trimmer (at least 10-turn)
  • 1--50K trimmer (not critical
  • 3--100K 10-turn trimmers
  • 1--1 megohm trimmer (not critical)


  • 4--10uF 16V electrolytic
  • 1--2.2uF 16V electrolytic
  • 2--1uF 6V tantalum
  • 3--0.47uF Mylar (should match in brand)
  • 1--0.2uF disc or Mylar, probably two 0.1uF units in parallel
  • 2--0.1uF disc or Mylar
  • 1--0.0047uF disc or plastic
  • 1--220pF disc
  • 2--100pF disc
  • 1--30pF disc


Note: Any of the 74C series could be replaced with 74HC components. Also, I left RCA's "CD" prefix off all the "4000" chips, as any will do; the 4072 could be an RCA CD4072 or a Motorola MC14072.

  • 6--1N914
  • 2--2N2222
  • 1--Siliconix P-channel JFET J176
  • 1--LM336
  • 1--LM358
  • 2--CA3130
  • 2--CA3240
  • 1--ICM7556
  • 1--National ADC3511
  • 2--4013
  • 1--4017
  • 1--4072
  • 1--74C157 or 74C257
  • 1--74C193
  • 2--74C221


Note: It is assumed that the on-off switch is on the volume control of the Nattering RAM; if not, add one more SPST unit. Also, individual digit-suppression switches are not listed here; if you want those, add four SPST toggles.

  • 3--SPST toggle
  • 1--DPDT toggle (without center-off)
  • 1--SPST normally open pushbutton (as big as your hat)
  • 1--3-pole double-throw toggle (without center-off)
  • 1--4-pole 3-position non-shorting rotary (Mouser ME105-14574)
  • 2--2-pole 5-position non-shorting rotary (Mouser ME105--14572)
  • 1--2-pole 6-position rotary "Make-before-break" for the ammeter (Mouser ME105-13572)

Connectors and Sockets:

  • 4--black binding posts, one not insulated from the chassis
  • 4--red binding posts
  • 1--red banana jack
  • 1--single-terminal insulated tie point
  • 2--4-terminal insulated tie strips
  • 1--6-inch ribbon cable with connectors (Digi-Key C3AAG-1406M-ND) 2--wire-wrap ribbon-cable connectors (Digi-Key CHW14G-ND) 3--14-pin "2-level" wire-wrap sockets (Jameco 14WW)
  • 5--16-pin "2-level" wire-wrap sockets (Jameco 16WW)
  • 5--8-pin solder-type sockets
  • 1--14-pin solder-type socket
  • 1--24-pin solder-type socket

[Editor's Note: Those pesky ribbon wire-wrap things have under-designed screw holes to hold them down; there is no room for the head. I made very narrow No. 2-56 nuts by filing opposite edges nearly through to the threads, and those worked. With that system, the heads are on the wiring side of the board and the goofy nuts are on top of the connector. You need locking forceps to hold those nuts in place.]


  • 1--1/8-inch closed-circuit earphone jack
  • 1--1/16-inch open-circuit foot-pedal jack
  • 1--small plastic panel for the above jacks
  • 1--5-amp quick-blow fuse with holder
  • 2--4.5V light-duty alkaline batteries (soldered together AAA cells)
  • 1--8- by 12- by 3-inch chassis (Newark 90F957)
  • 1--cover plate for chassis (Newark 91F1425)

Address List

Jameco Electronics, 1355 Shoreway Road, Belmont, CA 94002

Tel: (415) 592-8097

Mouser Electronics, 11433 Woodside Avenue, Lakeside, CA 92040

Tel: (619) 449-2222

Newark: Their headquarters are in Cleveland, Ohio. They want you to call and find the Newark outlet nearest you; (216) 391-8300. For export to foreign lands, call (312) 784-5100.


The following appeared in The Matilda Ziegler Magazine, Vol. 84, No. 3, April 1990.

"Memorial Fund Donations--The Alumni Association of the New York Institute for Special Education has started a fund in memory of Robert W. Gunderson, who taught amateur radio and electronics at the school for many years, and died in 1987. The fund will be used by the alumni at each June graduation to award a small scholarship to either a visually impaired or deaf-blind student most deserving scholastically. The amount awarded will depend on the level of donations to the fund. You may donate by sending checks or money orders to the Robert W. Gunderson Memorial Fund, c/o the Alumni Association of the New York Institute for Special Education, Att: Joseph Bruno, Treasurer, 420 West 261st St., Bronx, NY 10473.

"Please make checks payable to: The Robert W. Gunderson Memorial Fund, and in your check memo, please put c/o The Alumni Association Fund, N.Y.I."

Thank you, Alumni. The sentiment is proper and the goal is laudable. He loved the New York Institute, and he would do anything to help young folks along.

I owe that fine fellow my career. It is nicer today; with all the disability awareness that is now commonplace, it shocks me to think of the isolation that was the backdrop behind disabled people choosing unusual careers in the 1950s. Robert Gunderson brought us together on the air, in his "brain trust" of alumni, and through the distribution of the only hard reading you could get back then, The Braille Technical Press.

When other disability groups were forming in the early 1970s, I would hear people talk of "role models" as if they were a new thing. I always had them, and it was not because I was somehow blessed. The image of my success was handed to me in a packet of Braille that arrived every month.

Well then, here's a chance to ante up, and I'm glad of it. I sent a chunk of money, and I'd like to encourage you to do the same, please. Thanks.


Well, here I am again, and glad to be thus. To save you from tuning in late, I send it later; seems like a good system to me.

I've got some articles from you fine folks that will make a partial next issue--thanks. However, I've temporarily turned off my soldering iron, and I have been working on research projects that don't lead to buildable things. Thus, I can only hope that you will put up with a search through the pockets of my Boy Scout's uniform to find toys of interest for this magazine. I can describe the things I've worked on. You might get a kick out of them:

One project has been to evaluate "Descriptive Video Services"--describing TV movies to blind people over the "second audio program" channel (SAP). I now know tons of stuff about how TV transmission and network distribution is done. New tape machines and editing stations are pretty nifty as well, and perhaps you'd like to "follow me through my technical tour of the TV industry."

The other project is to define the technical stumbling blocks to using fax (facsimile) machines in a centralized reading service. Blind people with fax machines would call a toll-free number with readers, "fax" their print stuff over the phone, and get it read back to them: "Yeah, that looks like a court summons sure enough, buddy." We're about to start a trial project here locally.

Since I've got a fax machine temporarily sitting on my work bench, I've been sending silly poems to unsuspecting recipients; I could publish some of those. (Such great works as: "Oh I'd go Down the Aisle with Any Facsimile of You Dear," and "My Golden Retriever is an Over-Achiever, and He Left Me Cryin' in the Rain." Well, maybe I won't waste this magazine on those; write with requests.

Anyhow, the world keeps precessing, and it'll take us somewhere interesting--I'm sure of it. Until next time, have a mighty 1990. I always loved it when Gunderson would get sentimental and give us all New Year greetings. With that spirit, I give you my hope for the best.

P.S. It is time to subscribe again, and I need your support. Notice that the prices haven't changed--$18 for Braille, $16 for diskette, and $14 for talking book cassette. Those who, out of unflagging loyalty, have already sent me money are duly noted down, and you will not get the cute notices--with death threats in 'em--that we send around.

P.P.S. Hey, I almost forgot; we had an earthquake. That's why it's late! Everyone else uses that excuse, so why can't I? Most of us are fine, by the way. My piano didn't go for a walk, and the cuckoo clock didn't stop. There are a lot of new sets of dishes being bought out here, though.