Smith-Kettlewell TECHNICAL FILE
A QUARTERLY TECHNICAL JOURNAL FOR BLIND AND VISUALLY IMPAIRED READERS
Sponsored by REHABILITATION ENGINEERING CENTER SMITH-KETTLEWELL INSTITUTE OF VISUAL SCIENCES
2318 Fillmore Street, San Francisco, CA 94115
Winter 1982, VOL. 3, NO. 1
TABLE OF CONTENTS
To begin with, a couple of small typo corrections are in order:
On page 1, "phone" appears in the beginning of the second paragraph, this should actually be phono.
On page 3, "prreparation", at the beginning of the second paragraph should read preparation, Secondly, I would like to point out that the 276-1780 Top Hyphen chip octave chip, mentioned on page 23 in the singing chips article, is no longer available from Radio Shack, the equivalent is the 50240 made by AMI and MOSTEK.
My thanks to Mike Bhagwandas of Kobe, Japan, for unearthing a comedy of errors in the read-out article.
To begin with, it is important to note that the 4068 is an 8-input NAND gate, not an NOR gate as stated in the article. For future reference, a 4078 is an 8-input NOR gate with the same pin-out.
At first glance, there seems to be a logic error in how the input gates are combined (stated to be done with 8-input NOR gates,) since "off" inputs to the read-out box are low, the outputs of the front-end NAND gates are normally high--thus, they cannot be combined with a NOR function. By luck or by design they are not, since the 4068's are NAND gates as required.
In Circuit II, an actual logic error was committed. The outputs of the 4068's, which are normally low, were stated to be combined in a spare NAND gate, which is not possible. We need an OR function to detect when either of the 4068's goes high. For Circuit II, there are three possible solutions:
You can add an OR or a NOR gate chip (it doesn't matter which), remembering to tie all unused gate outputs to ground. You can use the NAND gate as stated if you invert the signals from the 4068's, which still requires another chip. My solution is to use the 4073's in place of the 4023's (these are 3-input and gates with the same pin-out), and to replace the 4068's with their 4078 NOR equivalents. With this latter solution, the circuit will work as described without any wiring changes. Clever, non?
In trying to be helpful (?), the Editor committed a logic error in circuit I, by connecting an inverter (one of the 4002 gates; pins 5 and 6 in, pin 4 out) behind the final gate to make it into an OR gate as incorrectly stated. Instead, tie all unused inputs to ground (5, 6, 8, 9, 12 and 13), and connect pin 3 of the 4001 through 47k to the oscillator transistor. Mr. Swail, I did not, as you think, read your manuscript with my feet--I read braille with my thumbs, as does everyone from my planet, Binarus.
Having been a ham for two decades, and having worked in a lab full of test equipment on which many of these connectors are used, the editor has accumulated a bag of tricks for installing RF connectors. Besides making the job of installation easier, these tricks are intended to address two objectives; one is to prevent damage to the cable in preparing it, and the other is to give the blind technician positive assurance that the connector is properly fitted and soldered to the cable.
RCA phone plugs shall not be called RF connectors (in my presence), and are taken up in part V, as is their cousin the Motorola plug.
Cable-hyphen mounted male connectors are the ones addressed in this discussion. Cable-mounted female units are used less frequently, and the principles used in working with them are exemplified in the discussion of their male counterparts. Chassis-mounted females often have conventional solder terminals and warrant no special mention.
The Standard UHF Connector (PL259)
The so-called "UHF" connector (whose name goes back to a time when F's were not very H and were not at all U) has been widely accepted by radio amateurs and CB equipment manufacturers. In a way, this is unfortunate, since installation is time consuming, and proper soldering of the shield braid is difficult. What's worse is that this connector is poorly understood even by veterans of radio work who often invent short-cuts to proper installation, making it imperative that knowledgeable people (such as we) oversee the preparation of cables at antenna parties.
The male center-conductor pin is tubular so that the cable's center conductor can be passed through it. The overall length of this tube is 5/8 in. from its tip to the bottom of the well inside the connector shell.
In the connector shell directly behind the center-conductor pin is a chamber whose inner diameter is just large enough to accept the braided portion of the cable. This chamber is about 1/2 in. deep and has an inner diameter of .350 in. Four holes around the perimeter of this chamber allow access to the braid through the shell of the connector; it is intended that the braid be soldered to the shell through these holes.
Further back, the inner diameter of the connector is large enough to accept the plastic-covered portion of large coaxial cable such as RG8/U. Internal threads at the rear of the connector are provided to accommodate "reducing adaptors" to be used with small-sized cable.
Preparation of Large Cable (RG8/U, RG11/U)
About 3/4 in. from the end, cut through the outer sheath, the shield braid, and the diaelectric insulator, and remove these pieces to expose the center conductor. Check with a fingernail to see that none of the strands of the center conductor were badly damaged by the knife. If obvious damage has been done, it will save you future grief to start over.
Using the procedures outlined in "Soldering Part III", smooth out and rebundle the strands of the center conductor and generously wrap them in solder. Using a heavy-duty soldering iron or gun, tin this conductor with the end of the cable pointing slightly downward to prevent a bead of solder from collecting at the base of the lead; such a solder bead can prevent insertion of the wire all the way into its tubular pin. Before continuing, insert the lead into the tubular pin from the front end of the connector to make sure that it still fits after being tinned.
Cut through the outer sheath 5/8 in. back from its end (5/8 in. back from the previous cut). Make no sawing motions with the knife that might cut into the braid; gently press the knife through the sheath on all sides, being careful not to knick the braid. Carefully remove this short length of outer covering and assess the damage. If whiskers around the knife cut indicate that a half dozen braid wires have been severed, it would be wise to start over.
Essential to installation of the connector is that the braid be completely tinned so as to make soldering inside the connector shell possible. Furthermore, the braid must be smoothed down so as to make it small enough to fit into its chamber. What makes tinning the braid difficult is that the dielectric insulator melts at a comparatively low temperature. Bring the braid to soldering temperature is often very injurious to the center insulation, and melting of the plastic can even contaminate portions of the braid, making it unsolderable.
I find that a thin strip of paper wrapped around the center insulator under the braid can protect the plastic from the heat. The braid can be flared out temporarily while the paper is wrapped around the plastic insulator beneath, then the braid can be closed over the paper. You may wish to "train" the paper by first wrapping it around a pencil to make it more manageable. The strip should be at least 1/2 in. wide and about an inch long. One-half of the slip from a Chinese fortune cookie works very well; not only is it precut to the right width, but this paper is very thin and supple.
It is nearly impossible to keep the braid smoothed down enough to fit into its intended space, and you will more than likely have to file the tinned braid down to size before the connector can be installed. As long as filing will have to be done anyhow, a piece of thin bare wire can be used to tie the braid into place. (Without the extra paper insulator, this trick can get you into trouble, since the wire will assure firm contact between the hot braid and the dielectric insulator.) Loop the wire around the end of the braid, and twist its ends together, tightening it just enough to keep it in place around the braid.
To keep the tip of the iron from contacting the plastic outer sheath, you may wish to wrap the cable in gummed paper or masking tape just behind the exposed braid. Whatever you use, it should be of a material that will not readily contaminate the tip of the iron.
Wrap a coil of solder around the exposed braid, using perhaps as much as 6 inches of solder. Using a heavy duty iron or gun, proceed in tinning the braid on all sides of the cable. By starting on the under side, you will prevent large segments of the solder coil from dropping off as you go. Each time you contact the braid, linger just long enough to feel the squeakiness of clean solder wet metals, then draw the iron off the end of the braid. After you have done this on all sides, let the cable cool for a couple of minutes and then check your work for large globs or unmelted pieces of solder. Approach these soldered globs specifically with the iron and draw them off the end of the braid.
On the one hand, you must take long enough with this process to assure that the braid is thoroughly tinned. On the other hand, the intense heat inside the braid cannot help but damage the cable. Your objective should be to reach a good compromise.
The braid can now be filed to size. Wrap a piece of tape around the center conductor to protect it from being marred by the file. File down the jagged end of the braid until it is even with the end of the dielectric insulator. File around the outside of your tinning job until it is smooth and cylindrical. (For those of you who own a micrometer, the tinned braid should measure 349/1000 of an inch.) Concentrate your efforts on the 1/2 in. of braid nearest the end, since this is the portion which must fit into the soldering chamber.
Testing for Proper Fit
Unfortunately, the adaptor threads inside the back end of the connector make testing for a fit difficult. This threaded portion of the connector is small enough in diameter to bite into the outer sheath of large sized cable; screwing and unscrewing the connector during the trial-and-error process of fitting is laborious and damaging to sheath. A slip-on test jig can be built by cutting off the rear-most 1/2 in. of a connector, thus doing away with these internal threads. Slipping this test jig onto the cable will quickly indicate if the tinned braid is too large for the chamber, or if the tinned center conductor is too large for the tubular pin.
Two indications as to whether or not the cable can be fully inserted into a connector are:
- The center conductor can become a "measuring stick" if you cut it to the exact length of 5/8 in. When the braid and dielectric insulator hit bottom in the connector shell, the end of the center conductor will be even with the end of the tubular pin. (You will recall that advice was given to cut away 3/4 in. of material to expose this center conductor. Depending on the shape of the end of the cable initially, you may need this extra length to insure that a good clean 5/8 in. long piece of this lead is available.)
- By tapping through the soldering holes with the point of a braille stylus, you will be able to tell whether or not the braid is present underneath. The surface beneath the holes should feel and sound metallic when tapped, and the holes should be completely obstructed by the braid. If the holes are only half covered, either the soldering chamber or the tubular pin is unable to accept full insertion. If the surface beneath the holes feels rubbery, the braid has been forced back by the entrance of the chamber, and your stylus is contacting the center insulator.
Preparation of Small Sized Cable
Slide a reducing adaptor of appropriate size onto the cable. Carefully one inch of outer sheath from the end of the cable. Position the adaptor so that its front end is even with end of the sheath and wrap tape around the cable directly behind the adaptor to keep it from slipping backward. Fold the braid back over the small end of the adaptor and tie it into position with a wire loop. Wrap the braid in solder and tin it; this time, draw the iron in front to back motions as if to smooth out the braid. Solder droplets will appear to stick to the adaptor, but the fact that the metal of the adaptor never reaches soldering temperature makes picking these droplets off possible after the project has cooled.
About 1/16 in. ahead of the adaptor, and braid strip off the dielectric insulator to expose the center conductor. Tin this lead, even if it is solid wire, since this will make soldering easier.
File the braid to size and test its fit in the connector as before.
Installing and Soldering of the UHF Connector
The connector comes with a screw-on ring to be used in securing it to its mating socket. Be sure that this ring has been passed down the cable before installing the connector. Screw the connector onto the cable and repeat the above tests to ascertain that it is properly in place.
Especially with large sized cable, the center conductor should be soldered first to hold the connector in place. If this is not done, bringing the body of the connector up to soldering temperature will soften the outer sheath, thus allowing the connector to slip forward. Your subsequent unseemly remarks will have the neighbors concerned for you.
Clamp the connector in a vise with the tubular pin pointing upward. Some technicians heat the pin from the side and hope that the conductor inside reaches soldering temperature. For this to work, the lead must be wet with solder; it helps if there is room in the pin to slip thin solder in alongside the conductor. These connectors, however, have a spoon-shaped end on the tubular pin so that the iron can be brought into contact with both the pin and the lead.
Hold a straight piece of thick solder vertically and position it on top of the pin. You can either position the tip of a soldering gun at the opening of the spoon, or you can slide a hot iron up the pin until the tip finds the spoon. When the solder melts, feed a half inch of it straight down into the pin. Very often, you will lose the pin with the solder. Let the project cool off and try again. Solder spilled outside the pin can be wiped off with the iron later; do this with the pin pointing downward. Also, it is a good idea to check for spillage around the base of the pin with a stylus or other small probe. This spillage can be removed with a knife.
The braid must now be soldered through the holes in the shell. If the connector is to be used outdoors, soldering all four holes will help protect the cable from moisture.
Orient the cable horizontally and clamp it in a vise about an inch behind the connector. (Clamping the connector in a vise directly will make heating it impossible.) Before securing the cable in the vise, turn it so that a solder hole is facing upward. It is easy to lose track of which holes you have soldered. You may wish to lay a strip of tape adjacent to the first hole to serve as a reminder.
I use a gun for soldering the holes, since positioning the tip of a hot iron can be tricky. In any case, hold the solder at an angle and insert it into the hole. Place the tip of the iron on the hole, or if possible, orient it so that a corner of the tip seats in the hole. The solder which melts immediately (disconnecting the solder in your hand from the connector) will serve a to create a column of molten metal under the iron in order to improve the efficiency of heat transfer.
Keep trying to find the edge of the hole with the solder. Spilling solder elsewhere on the connector will not hurt anything; just try to see that some of it spills into the hole. Do not leave the connector with the iron until rocking of the iron feels squeaky; this is your only true indication that wetting of the connector has occurred.
After the connector has cooled, clear some of the larger droplets out of the way with a knife blade and feel in the hole with a stylus. The hole should either not be there, or should feel more like a "dent" than a hole.
As a final test, try unscrewing the connector from the cable. Even with the center conductor soldered, the connector will rotate an eighth of a turn if no adhesion to the braid has occurred.
With knives, picks, files, grinders, hatchets, and soldering irons, remove enough spillage from the connector so that the outer ring can be attached as intended.
Salvaging UHF Connectors?
Don't! Even if you are successful in getting the cable free of the connector, it is impossible to assure that little whiskers of braid are not lurking in the shelter of the soldering chamber ready to arc to the center pin when power is applied.
I always buy more units than I need in case something goes wrong in soldering them, or in case I forget that pesky outer ring. Always start over with new connectors.
BNC and N Connectors
These connectors are much different from the "UHF" units discussed above. Their connection to the shield braid is solderless. The only soldering to be done is in attaching a tiny male connecting pin to the center conductor of the cable. Without effective holding clamps, attaching this pin can be difficult.
These connectors come with several small parts whose dimensions are specific to a given size of cable. A connector which has been purchased for one size of cable cannot effectively be used on another. (A list of connector numbers at the end of this article should aid in ordering appropriate units for standard sized cables.) Care must be taken so as not to lose any of the small parts, since all of them are necessary for proper installation. A list of these parts is given below:
- A very tiny pin which tapers to a point is intended for soldering to the cable's center conductor. A hole drilled axially into the rear of this pin accepts the center lead of the cable; the diameter of this hole closely fits the lead of its intended cable. Soldering is done through a hole in the side of the pin.
- A tapered washer is intended to fit into a matching counterbore inside the shell of the connector. The braid, which is folded back against the tapered washer, is sandwiched between the washer and the bottom of the counterbore.
- A rubber ring fits around the cable behind the tapered washer and is held in compression; it clamps the braid firmly between the washer and the connector shell, and it grips the outer sheath of the cable to act as a strain relief device.
- A thin metal washer behind the rubber ring prevents damage to the rubber as it is being compressed.
- A threaded insert screws into the shell of the connector behind all of the above pieces and is firmly tightened with a wrench to compress the rubber ring.
The threaded insert, the thin flat washer, and the rubber ring are slid onto the cable in that specific order. Next, strip off some of the outer sheath to expose the braid. (The amount of braid to be exposed will depend on the connector and the size of cable. Basically, the braid should flair out and fold back to cover the front surface of the tapered washer.) After the sheath is removed to expose the braid, slip the tapered washer over the braid (with the tapered side toward the end of the cable; press this washer back against the sheath and fold the braid back over it.
About 1/8 in. ahead of this braid assembly, remove the dielectric insulator to expose the center conductor. Generously tin this lead and cut it off to a length of 1/8 in. (beyond the end of its insulator).
The tiny center-conductor pin can be held in locking forceps, with one of the forceps' handles being held in a vise. You may wish to line the jaws of the forceps with bits of braille paper so as not to mar the pin with their serrations, and to keep the forceps from acting as a heat sink. Orient the pin horizontally with the soldering hole facing upward.
In order to keep the cable in position, it can be held in another vise located a few inches away from the pin. The center conductor should comfortably rest in the pin, i.e., there should be no sideways force on this lead which will cause it to cut through the softening dielectric insulator as the connection is heated.
Hitting the tiny pin with a soldering iron, without touching the dielectric insulator, is no easy task. You may wish to protect the insulator by poking the center lead through a small bit of braille paper, thus creating a barrier between the iron and the plastic. Hold a very straight piece of solder vertically and insert its end into the soldering hole in the side of the pin. Carefully follow the forceps over to the pin with the tip of the iron and position the iron on the under side the pin. Since the pin is machined to fit its center conductor, you do not need to apply very much solder to the connection. This is fortunate for us, since feeding the solder straight down into the small hole is often a matter of luck. You will often lose the pin with the solder, whereupon you should let it cool off and try again. Perhaps 1/8 in. of solder should be fed to the connector. Spillage onto the sides of the pin can be filed away after it has been secured to the center lead.
List of Connectors
- PL259--UHF connector for large cable, RG-8/U and RG-11/U.
- UG-175/U--Reducing adaptor for RG58/U cable.
- UG-176/U--Reducing adaptor for RG59/U cable.
- UG-21/U--Type N connector for large sized cable, RG-11/U.
- UG-88/U--Type BNC connector for RG58/U cable.
- UG-260/U--Type BNC connector for RG-59/U cable.
- Amphenol #69475--Type BNC connector for RG-174/U and other miniature coax.
by David Plumlee
I hope to make this the first in a series of articles with the purpose of introducing some musical applications of IC chips. Rather than focus on a specific project, I will use the approach of discussing several chips and some principles of their application in this area. Of course, I will draw heavily from some musical projects I am presently building to illustrate these principles.
For example, one of my projects is an attempt to approximate the sound of bagpipes and that of a dulcimer using digital techniques. (I call my instrument the "Ethereal Nachthorn", and a few additional strange sounds have been incorporated into it.) I assure you that I did not find the schematic diagrams of my projects anywhere. Rather, the circuits have come about as ideas occur to me to apply the chips as I shall seek to explain.
You will notice that I did not use the term "electronic music" in my introductory paragraphs - the omission was deliberate. Many people might think of electronic music as being some far-out bleeps, sweeping tones and noise, and assume that they have no interest in this series. I regard this type of electronic music as being a valid art form which must be appreciated on its own terms, although I am not too far into it myself. My hope here is that the material presented under the series "Singing Chips" will spark your imagination to apply IC's in the creation of your favorite type of music, all the way from the most traditional to the most "far-out". To borrow a phrase from a popular song, use the chips to "make your own kind of music".
SINGING CHIPS (Continued)
In this article, let us concentrate on generating an essential part of our traditional music, the scale of equal temperament. For those of you who gasp "What is that?", we set forth the following relationships:
An octave involves a two to one relationship. As concert A is defined as 440 Hz, then an A one octave above this is 440 times two, or 880 Hz; A one octave below would be 440/2, or 220 Hz.
What about other musical intervals? Semi-tones for example, the interval from concert A to B flat, are related by the 12th root of 2. The frequency of B flat can be gotten by multiplying A--440 by the 12th root of 2. The 12th root of 2 approximately equals 1.0594631. 440 Hz times 1.0594631 equals 446.16376 Hz. If you were to take 440 Hz and multiply it by the 12th root of 2, a total of 12 times (the mathematician would say, "440 times the quantity 2 to the 1/12th power raised to the 12th power"), we would get 880 Hz, the octave above concert A--440.
Implicit in the above is that any note in the equally tempered scale is related to its neighboring semi-tone by a common ratio. In fact, any given interval within this temperament (an interval being comprised of a given number of semi-tones) will have a specific ratio of frequencies, no matter where in the scale you play this interval. For example, the interval known as a fifth (Do Re Mi Fa So) spans seven semi-tones. Whether the fifth is from A to E or from C to G, the ratio of frequencies is the same. Seven successive multiplications of the 12th root of 2 equal 1.5874; the ratio of A to E is 1.5874, and the ratio of C to G is also 1.5874. A table at the end of this article lists the equally tempered pitches and ratios.
Synthesizing music with digital electronic circuitry would be trivial if we could do as Pythagoras did many centuries ago - define a temperament based on whole number ratios. He defined a "perfect fifth" as having a frequency ratio of 3 to 2 (1.5 to 1). We could easily produce the true Pythagorean intervals by using counters. As an example, if we clock a programmable divide-by-N counter at a frequency f and set it up to divide by 2, we will get the note exactly one octave lower than f, f/2. Now, if we set up another counter to divide the same clock frequency by 3, we will get the note exactly an octave and a fifth below the clock; its frequency will be f/3. The two counter output frequencies will be in a ratio of 3/2 (f/2 divided by f/3), and the sound heard through a linear audio amplifier would be a "perfect fifth".
(The term "counter" is used to denote a string of flip-flops in cascade. As discussed in "Flip-Flops", SKTF Summer 1981, gates can be set up to terminate or reset a counter after any given number of clock pulses. A programmable counter chip contains a string of flip-flops as well as gates whose connections can be orchestrated by the user in order to make it divide the clock by a particular divisor, N.)
We could insist that certain intervals be tuned perfectly to the Pythagorean ratios, but we would quickly reach a point where a note could not be tuned to satisfy all related intervals. In Western civilization, we have learned to accept small deviations from the Pythagorean ratios for each interval. If these deviations (compromises) are spread over the entire octave they are small enough so as not to be disconcerting. In other words, our ears have become so accustomed to the equal temperament that we do not notice the small errors distributed among the notes in the scale. (Any text on piano tuning indicates that intervals are either made wide or narrow by so many "cents", hundredths of a semi-tone; each interval on the keyboard, except the octave, is tuned with a very small error.)
Now that we have all those decimal numbers and ratios spinning around in our minds; let's consider how we might generate the equal temperament by digital techniques. We have heard that in digital logic, switches are either off or on, so how do we synthesize the imperfect frequencies of the equal temperament using digital hardware? Let's connect 12 divide-by-N counters so that they are all clocked by a common signal in the RI band. With careful calculation, we can choose a clock frequency and 12 rather large values of N (12 divisors) so that each counter will have an output that closely approximates a pitch in the equally tempered scale. Fortunately, the arithmetic and fabrication has already been done for us by way of the Radio Shack 266-1780 Top-Octave Synthesizer.
Table of Divisors
For our comparison, two sets of clock frequency divisors are listed, and will appear after each note's name in the following respective order:
The first will be the "On Chip Divisor" as taken from the Radio Shack Data Manual. The second will be the "Ideal Divisor" which would obtain the actual equal temperament pitch as calculated on a Radio Shack scientific calculator. The manufacturer starts out with a clock frequency of 2.0024 MHz and a divisor of 478 in order to get "low e" (which is the top key on the piano). I calculated successive ideal divisors by repeatedly multiplying 478 by the reciprocal of the 12th root of 2. The first note listed is the e at the top of the piano, approximately 4187 Hz.
- Low C--478, 478.0 (given)
- C sharp--451, 451.17192
- D--426, 425.84959
- D sharp--402, 401.94849
- E--379, 379.38885
- F--358, 358.09539
- F sharp--338, 337.99704
- G--319, 319.02673
- G sharp--301, 301.11213
- A--284, 284.22050
- A sharp--268, 268.26843
- B--253, 253.21168
- C (one octave above the piano)--239, 239.0.
All this chip needs is a clock signal at pin 2, VSS at pin 1, and VDD at Pin 3 (in the case of this chip, VDD is negative with respect to VSS). The remaining pins will provide the equally tempered pitches within a tolerance of 1/2 cent, according to my calculations. The outputs of the chip are buffered and produce square waves. Of course, the overall pitch accuracy is only as stable as the clock which drives the chip, but if wiring is done with reasonable care there will be no problem. The 276-1780 Top-Octave Synthesizer is a CMOS device selling for about $7.00, so treat it kindly, folks. We will clock it with another CMOS chip, the 4011 Quad 2input NAND gate, priced at about $1.00.
(Two gates in this package are used to comprise an oscillator.)
Pay careful attention to the voltage polarities on these two devices.
In fact, let me spend a moment on this subject for the sake of the newcomer, especially since we are dealing with a moderately expensive chip. Such devices as these are often produced as components to be included into commercial products; their specifications and voltage reference points are chosen so as to be most appropriate for the product design. These interface specifications and voltage reference points may differ from the norm - a notable example is the power supply arrangements required to operate the TSI Speechboard. VSS is usually the point of reference for all voltages listed in manufacturers' data sheets.
Looking at the 4011 we see that VDD (pin 14) must be positive with respect to VSS, the reference (pin 7). We must connect pin 14 to the plus line, pin 7 to the minus line, and we shall henceforth take the negative line as ground. On the other hand, the packaging card for the Top-Octave chip indicates that VDD must be negative with respect to VSS: we must connect VSS (pin 1) to the plus line and VDD (pin 3) to the minus line.
Circuit for Operation of the Top-Octave Synthesizer.
(This circuit is shown on the back of the card to which the Top-Octave Synthesizer chip is attached).
This circuit can be powered from a supply of 11 to 16V; I used 12V regulated in my projects. Pin 7 of the 4011 is taken as ground and goes to the negative terminal of the power supply. This pin 7 also goes to pin 3 of the Top-Octave Synthesizer. Pin 14 of the 4011 goes to pin 1 of the Top-Octave chip and to the positive terminal of the supply.
Pins 1 and 2 (inputs to one gate) on the 4011 are tied together and go through a 10K rheostat, thence through a fixed resistor of 2.2K to pin 3, the output of the same gate. (More on this 2.2K resistor later). Pin 3 of the 4011 goes to pins 5 and 6 of the 4011 which are tied together (inputs to another gate). Pin 4 of the 4011, the output of the second gate, goes through a capacitor of 100 pF back to pins 1 and 2, inputs to the first gate.
Pin 4 of the 4011 is the output of the clock, and it is labeled on the package to be at a frequency of 2.0024 MHz. This clock output goes to pin 2 of the Top-Octave Synthesizer.
While pins 8 through 13 of the 4011 are not mentioned above, we must remember that this is a CMOS device, and proper treatment dictates that no input must be left uncommitted to a logic level. Connect pins 8, 9, 12, and 13 to ground. I suggest making these connections in such a way that you can later release these pins easily from ground to commit their connection to other components. As the project expands, you may think up a use for these two unassigned gates. The important point is that with CMOS, you don't leave inputs open-- doing so invites noise pick-up and other circuit malfunctions. With TTL, on the other hand, you often leave open inputs--they assume a condition of logic 1 when they are not being pulled down.
Remember that if you add TTL chips to your project, the logic levels are incompatible, 5V versus 11 to 16V. Appropriate interface chips, such as the 4049, must be used to convert the 12V level of this project to the 5V level required by TTL devices. (The 4049 would be powered from the 5v supply, thus restricting its output swing to 5V, but its inputs can be pulled up to 16V without damage being done.)
Now the information I promised you on the 2.2K resistor. In some instances, this resistor may not be quite small enough to allow tuning of the clock to 2.0024MHz. If this turns out to be the case, temporarily insert a SK rheostat in place of the 2.2K resistor and set it to about mid-range. Turn the 10K rheostat on your board to its minimum resistance. Connect a signal tracer (audio amplifier) to pin 16 of the Top-Octave chip (low C in the scale) and adjust the temporary rheostat until the pitch matches high C on the piano (approximately 4187Hz.). You now have the chip tuned to generate that octave which is one octave above the piano (you will later use counters or simple dividers to generate pitches of lower octaves).
For a little extra margin, you can now "tweak" the temporary rheostat down in value so as to bring the pitch up to C sharp or at most a D. Don't go any higher; some of the dividers in the chip will not behave properly, and you will get incorrect notes. Now, turn off the generator, remove the SK rheostat and replace it with a fixed resistor of the value to which it is being experimentally set.
By the same token, you can directly generate an octave considerably lower in frequency by increasing the value of the 10K rheostat and/or the fixed resistor. If your application were, for example, a game or toy in which you need only the range around middle C, merely tune the chip to that range by setting the rheostat, increasing its maximum value as necessary. You can also provide fine and coarse tuning by placing two rheostats in series. Remember, however, that there must be a minimum resistance in the clock circuit so that the frequency of the oscillator cannot go too high.
One quirk of this circuit is that when the oscillator is set at the top of its range, the chip may fail to generate pitches when power is first applied. While I know of no fool-proof way to eliminate this problem, I know that the circuit can be started by merely increasing, for a moment, the value of the rheostat (lowering the frequency). The system begins to function immediately, and can then be returned to the desired frequency. One solution might be to include the on-off switch on the pitch control (the 10K rheostat), so that when the circuit is first turned on, the rheostat would be at maximum. Another approach might be to insert an additional parallel combination of a 15K resistor and a normally-closed push button switch, a "start" switch. Momentary opening of this switch would temporarily lower the clock frequency and permit the system to start.
Those of you who are not seasoned in electronic logic might be interested in the fact that a 4001 quad 2-input NOR gate can be used in the oscillator, and would you believe that you can plug it in and have it work perfectly with the same wiring given above? Why?
You will recall that we have both inputs at the assigned gates joined together. These gates are connected in the so-called "inverter configuration", and we are merely using them as inverters. If you were to write truth tables for the NAND and NOR gates, you would see that the response of their outputs is identical under the two conditions where the inputs are equal (both high or both low). The difference between the two gates will occur only when their two inputs can go to different logic levels. Suppose we were to free one of the inputs--for example, pin 2 of the 4011 will stay connected as shown in the schematic, but pin 1 will now go to the arm of a single-pole, double-throw switch. We will ground one position of the switch, and connect the other position to 12V. If the switch is set so that the arm is grounded, the AND function of the NAND gate will never occur, and the output will go to "1" and stay there. If we set the switch in the other position, its arm at 12V, the output of the gate will depend on the state of pin 2, and the gate will become an inverter as before.
If a NOR gate were put into the same service (still retaining our switch arrangement), the oscillator would only work when the arm of the switch is at ground, and not when the switch takes pin 1 high. The switch positions would be reversed.
In the previous two paragraphs, we have created a means of "enabling" or "disabling" the clock. We could now remove the switches and connect pin 1 of the oscillator to a JK flip-flap. Clocked from another source, say at 1Hz, we could repeatedly key a note--this might be just the thing in an electronic game application. Also, keep this principle of using a gate to control a train of pulses in mind, it may often come in handy.
The Top-Octave chip provides 13 pitches as follows:
- Pin 4--C-sharp
- Pin 5--D
- Pin 6--D-sharp
- Pin 7--E
- Pin 8--F
- Pin 9--F-sharp
- Pin 10--G
- Pin 11--G-sharp
- Pin 12--A
- Pin 13--A-sharp
- Pin 14--B
- Pin 15--High C
- Pin 16--Low C.
You thus have one complete octave--all 12 notes of the equally-tempered scale are presented plus the 13th note which completes the octave.
Since the equal temperament is based on ratios rather than absolute frequencies, you can freely choose a clock frequency to provide a different range than that listed above. If, instead of C to C, you need a range of F to F or G to G, simply adjust the clock frequency to suit your application. If you set the clock so that pin 16 is a G, pin 15 would be the G one octave higher; you can rename the rest of the pins, and the other pitches will be correct as to their position in the scale. For instance, if pin 16 is a G (whereas it is listed above as a C), pin 4 (which is listed as a C-sharp) must be renamed G-sharp, etc.
Once you assemble one of these temperament generators, the number of applications will be limited only by your imagination. (Henceforth we shall use the term "temperament generator" when referring to the basic circuit described earlier. The term "Top-Octave" will be used when referring to the 276-1780 chip.) As one simple application, you might consider making a toy organ. Connect each pitch (perhaps excluding sharps and flats) to one end of a normally-open push button. The other end of all these push button switches can then go to a common bus which is the input of an amplifier. You can then tune the clock to produce pleasing pitches for a child.
Note that the data card specifies a minimum voltage of 11V for the Top-Octave Synthesizer. However, in a pinch one evening, I got my chip to work on a 9V battery, although I wouldn't count on such a set up for reliable operation. I would consider using two 9V batteries in series with a couple of silicon diodes in the Line to drop about 1.2V off the supply voltage. These diodes would also protect the device from polarity reversal should your youngster inadvertently touch the batteries to the clips backwards. You should also by-pass the supply with a parallel combination of 0.1mF and perhaps 100mF for suppression of digital noise on the line.
some amplifier chips, such as the LM386, will not operate properly above 12V. Choose your amplifier with this in mind, or perhaps run the audio amplifier from a 7812 voltage regulator.]
Project Ideas (Continued)
In developing a project, we might first decide whether we want the basic system to be "monophonic" (you can play only one note at a time) or "polyphonic" (you can play any number of notes simultaneously). While my basic system is monophonic, let me drop a hint to those who should desire polyphonic operations. Consider acquiring twelve CMOS 4040 12-stage binary counters and connecting their respective clock inputs to pitches on the Top-Octave chip. On the 4040 pin 16 is plus VDD, pin 8 is VSS, pin 10 is "clock", and pin 11 is "reset" (If high, the counter is reset to all "zeros" and held there as long as pin 11 remains high). The other pins would provide octaves below the Top-Octave pitches.
I have not taken a polyphonic approach with my project. Rather, I have one 4040 which is switched to the desired Top-Octave output. With this approach, you may send only one pitch at a time to the counters' clock input--otherwise, you will get garbage as an output. (Remember that outputs should not be shorted together; if you want to combine pitches from the Top-Octave chip, do so using a gate or other multi-input logic block.)
Don't let the lowly monophonic project fool you. It can be pleasant to listen to, and there are things you can do to make it interesting. (You can make multiple recordings to get the effect of polyphony, harmony.) Though you can only key one note at a time, you can wire some drones of appropriate pitches on the project to run continuously, and that's where my simulated bagpipe project began.
My next article might show how we can use two 74148 priority encoders, a 4051 eight-line multi-plexer, a 4066 quad bilateral switch, and a few other bits and pieces to generate two monophonic octaves with programmable timbre. As a preview, if you want to create a different timbre (tone quality) by digital means, you can AND two pitches together which are one or more octaves apart. For instance, if you combine the frequencies f and 2f, you will have an output at frequency f which will be a rectangular wave with a 25% duty cycle. To predict this, all you need is to tabulate each point in the wave form based on the changes of the input conditions and apply the rules for an AND gate (see examples in "Gabbing about Gates", SKTF, Summer 1981).
We could also take four outputs from the 4040, for instance f, 2f, 4f, and 8f, and run each of them to the top of the potentiometer. Each arm could go through a suitable isolation resistor to the amplifier input. By adjusting the pots, we can regulate the amplitude of each frequency and create many interesting effects. The "f" pot will carry with it an infinite number of odd harmonics: f, 3f, 5f, 7f, etc. The 2f pot will give you 2f, 6f (three times 2f), 10f, etc. Similarly, each frequency has its own series of odd harmonics, so you will have a lot of interesting "knobbing" at your fingertips. Simple RC filtering and active filters using op-amps can also be used to get different tonal characteristics from a square wave.
Some Notes on Digital Wave Forms
In electronics, we have at our disposal a capability which does not exist in the mechanical world. We take for granted the ability to generate a tone which has a virtually instant rise and instant fall time when viewed on a scope. To get a perspective on this point, compare such a wave form to the situation of a vibrating string. In order for the string to generate a square wave, it would have to move immediately between extreme positions of its travel in zero time. Of course, a string doesn't behave that way, and neither does a vibrating air column. However, if you listen to a square wave whose pitch is around middle C on the piano, you will note that it sounds somewhat clarinet-like. This is true because a square wave contains an infinite number of odd harmonics; a clarinet's tone is also rich in odd harmonics. We can regard a square wave as making instantaneous transitions between two steady DC-like states (high and low). As mentioned earlier, our temperament generator's outputs give us square waves.
Incidentally, let's clarify any possible confusion between the terms "rectangular wave" and "square wave". To qualify as a rectangular wave form, the signal needs to have instant rise and fall times, and the tops and bottoms of the wave form must be straight horizontal lines. No requirement is made regarding "duty cycle", i.e., the signal could still qualify if it were high for 90% of the time and low for 10% of the time, a "90% duty cycle". A square wave, on the other hand, must be a rectangular wave with a 50% duty cycle. Once a signal deviates from the condition of having equal "on" and "off" times, a 50% duty cycle, it becomes classified as a rectangular wave. Note that a square wave is not necessarily square, since the amplitude and period are independent parameters. You can have a very tall square wave (having a large amplitude) who?s "on" and "off" times are very short (thus making the pulses very skinny).
While the outputs of the Top-Octave Synthesizer produce square waves, the clock oscillator does not. Some of the quirkiness in getting the system started could result from a problem-with the duty cycle of the clock. Let me use what is probably a gross exaggeration to illustrate my point.
Suppose when the clock is run up to around 2MHz, it reaches a point where the duty cycle is 99%. If this were the case, the output would be off for only 100th of its period, 1/2,000,000 divided by 100, seconds. Its off time would be only 5 nanoseconds, which is unusable with CMOS logic.
As an experiment to improve triggering of the Top-Octave chip, one might try triggering it through a flip-flop (inserting a flip-flop between the data between the clock output and pin 2 of the Top-Octave chip). One would of course have to double the clock frequency (the flip-flop would divide it by 2), and the oscillator's feedback capacitor would undoubtedly have to be changed. The flip-flop would have to be capable of responding to a drive signal of better than 4MHz. Perhaps with such experiments, one could find out what is limiting the maximum frequency of pitches the system will reliably generate.
It is not that I strongly advocate doing the above experiments--rather, I feel that issues like the matter of duty cycle in connection with digital circuits should at least be placed somewhere in the back of your mind as a point to think of when your circuit misbehaves.
Finally, let me close on a note (pun intended). As encouragement to all of you, whether you are a seasoned electronics builder or a virtual beginner, you don't have to spend several hundred dollars all at once on a kit or a batch of parts to have fun and/or get your feet wet with a few of these "singing chips" you will soon be creating your own electronic music.
Tailor your project goals to your experience. If you are a beginner, start small and keep it simple. Set up a board that lends itself to ANDing and a variety of other connections to see what you get. Who knows? Your project might take a course somewhat like mine.
I started out with a basic temperament generator using only the two chips in the Radio Shack schematic. For a while, I keyed the system by touching a jumper wire to the desired points on my board. I then graduated to a keyboard using an upside-down pie pan with some inexpensive push-buttons mounted on it (I still have that project around).
Then I said, "If I could add a flip-flop, or maybe a dual flip-flop...I did that, and a few days later my thought was, "Hum-mm-mm, if I got a multi-stage binary counter..."
And the story goes on until you wake up one evening to count the chips on your board and find that there are perhaps 20. "Did I count right?"
Since your project may grow into a multi-chip board, I strongly encourage you to learn the wire-wrap technique. This takes a little diligent practice and some care in doing the work, but it has several advantages with respect to flexibility. No heat is used in wrapping, so you can make changes with the chips still in their sockets. Although the small wires are somewhat difficult to trace in a complicated project, you can make changes without much trouble while the layout is fresh in your mind. If you want to make an experimental connection on a wire-wrap pin already fitted with leads that you know will stay there, wrap your questionable wire near the top of the pin; you can easily spot this high wrap when and if you wish to remove it.
If the mechanics of constructing a keyboard worries you as it does me, there are, as the proverbial saying goes, more ways than one to skin a cat. Hint--you might mount a group of reed switches in a row and use a magnet to activate them. You could just slide a magnet along the row to key the system, which is what I'm planning to do in my simulated bagpipe project.
Bill Loughborough, who has worked as an inventor at our Institute from time to time, developed a portable organ using the above temperament generator and 12 4040's. As a keyboard on the first model, he connected the tone outputs to banks of round-headed tacks which were placed in close proximity to horizontal bus-bars. These bus-bars all went directly to the input of an audio amplifier. When the player's finger "bridged" the tone to the bus-bar, it was heard through the amplifier. Furthermore, the harder you pressed, the louder would be the tone, since coupling to the amplifier was a function of the firmness of this connection. In other words, the finger was used as a pressure sensitive resistor. Hints on building such a keyboard can be gotten from Swaib's "Tactile Readout" of the next article.
A later model, the "Spongiphone", had little pieces of conductive foam such as that used for storing CMOS chips which were taped on top of printed circuit traces arranged in a keyboard configuration.
Mr. Loughborough's instrument had tiers of "keys", each of which fell just short of being a "tenth interval" in length. In this way, tenths were placed adjacent to one another in a column, and could be played in multiples with the fingers of one hand.
Because of distributed capacitants, all the notes found their way into the amplifier to some extent. To suppress some of this noise, it would be wise to run the amplifier input through a parallel RC combination to ground--perhaps starting with 47k in parallel with 0.001uF.]
Some Notes on Digital Wave Forms (Continued)
Once again, your own interest in gradually learning about digital IC's will set the pace at which your instrument grows. For beginners I repeat: start small. So get some of the chips, play with them, and have fun making music.
Table of Frequencies and Ratios
Starting at middle C on the piano, this table shows the frequency for each pitch in one octave of keys. These frequencies are 1/16th of the value produced by the Radio Shack 276-1780 when clocked at the recommended frequency of 2.0024MHz. The table also shows the equally tempered ratios as reference to C; these hold true for any octave.
Following the name of the note will be two entries. The first is the actual frequency of the note in Hz, and the second is the ratio of the given note to C, the first note in the table.
- Middle C--261.62556Hz, 1.0
- C-sharp--277.18263HZ, 1.0594631
- D -- 293.66477Hz, 1.22460
- D-sharp--311.12698Hz, 1.1892071
- E--329.62756Hz, 1.259120
- F--349.222823Hz, 1.3348399
- F-sharp--369.99442 Hz, 1.4142136
- G--391.99543Hz, 1.4983071
- G-sharp--415.30470Hz, 1.5874010
- A--440.0Hz (defined), 1.6817928
- A-sharp--466.16376Hz, 1.7817974
- B--493.8830Hz, 1.8877486
- C--523.25113Hz, 2.0.
by J. C. Swail
In my work at the National Research Council of Canada on the development of special instrumentation for the blind we have tried a number of systems to permit reading of digital instruments. The main interest is in frequency counters, digital volt-ohm-millimeters, and ham transceivers.
Our first approach was to interface the Telesensory Synthetic Speech Board to such devices. This works very well, and talking frequency counters and ham transceiver read-outs are commercially available from J.C.U. Electronics* of Calgary, Alberta. (*JCU Electronics (Mr. James C. Upright, VE6ATA), 7007 Huntridge Hill North-East, Calgary, Alberta, Canada T2K 4A4.)
We have found, however, that while the speech works very well for many general instrumentation applications, it is often unsatisfactory as a read-out on amateur transceivers. The speech tells you where you are presently tuned, but you must listen to a series of announcements when adjusting the equipment to a desired frequency. Finding a net or sket frequency with such a system is simply a grand pain where you sit down.
We thus looked for a tactile solution. Braille was the natural first choice, but it does require a code conversion from the code used in most counters to that of six dots. This is no trick where one has facilities to burn PROMs (Programmable Read-Only Memories), but it is not the sort of thing that the average home builder is equipped to deal with. After some experimentation, I discovered that binary-coded decimal (BCD), which is the language of the counter, can be read in tactile form just as easily as Braille.
If we refer to the article "Counting in Base 2", SKTF, Summer 1981, we will find that the numbers 1 through 10 are as follows:
Now, if we represent a number by a vertical column of dots having digital values assigned to them (from top to bottom: 1, 2, 4, and 8), we have a simple tactile BCD read-out system. With this system, for example, a 9 would be represented by the first and last dots being on; a 5 would be represented by turning on the second and fourth dots in the column (counting from the bottom).
Our next problem is how to generate the dots. Solenoids or Piezo-electric crystals can be used to push pins up, but again, these techniques are expensive and out of the question for home constructors.
A system which has been adopted with great success is to form the dots as insulated islands surrounded by a copper plane. When your finger touches a dot, a high-resistance path is established from the dot to the plane. Acting as a switch, this high-resistance path closes the gate and sounds an alarm --provided that this dot is an active component of the digit being displayed. Thus, as you run your finger along a column of dots, those which sound a tone are considered as 1's, and those which are silent represent 0's.
The units built here have been constructed in chassis 5 inches wide by 6 inches deep and 2 inches high. The read-out board is on the top surface, with the circuitry being contained inside. A connector on the rear brings data from the instrument being read. There is room enough for a built-in power supply, although the necessary BCD can be stolen from the instrument in most instances.
The read-out itself, as built here, is on a printed circuit board of the double-sided variety. The top side is a continuous copper plane, except for insulated islands which will be the dots. Columns of dots are spaced 0.75 inches apart; the space being dot centers within a column is 0.5 inches. A hole is drilled in the center of each dot. Directly on the other side of the board and leading away from these holes are traces which afford connection of the dots to the circuitry inside the chassis. A very short piece of wire (slightly longer than the thickness of the board) is inserted into each hole. Each piece of the wire serves two functions; it connects its dot to the trace on the under side of the board, and it provides a structure around which a cone or solder may be built up to form a nice dot.
For the home constructor, it may be simpler to build a read-out on a piece of perforated board (vector board). Dots can be made using round-headed pins or brass (brass ones are easier to solder to). For the back plane, the rest of the board can be covered with aluminum foil, or, heavy bus wires can be run alongside the columns of dots. (Perforated flat copper bus strips, vector, No. T107, could be used; this would afford securing it in several places along its, length with hairpin loops of wire pushed up from the bottom and soldered).
Generally in our units, we have limited the number of columns of digits to four. This is sufficient in the case of ham transceivers, since by noting the position of the band switch, the first two digits can be inferred.
The circuit for the read-out takes either of two forms, depending on the design of the counter to which it is attached. These two variations are listed as follows: some counters have a separate BCD out for each digit. Each of these BCD's goes to a BCD-to-7-segment driver, which in turn, drives one digit on the visual display. In this case, we require a two-input AND gate for each dot, one input for the dot and the other for the corresponding connection to the counter's BCD point.
The second case, unfortunately, is much more common. Only a single BCD signal is available, and it goes to all of the 7-segment drivers in parallel. The BCD information is "time-shared", and is changing continuously in accordance with a group of drive signals. A particular driver and its visual digit only responds to the BCD information when its specific multiplex signal is present (changes from 0 to 1). Thus, in this instance, we require triple-input AND gates--the user is signaled when a "1" is present and THE MULTIPLEX SIGNAL IS PRESENT and THE DOT IS BEING TOUCHED.
Connections are made directly to the counter board of the instrument to be read. In the first case, non-multiplex and a four digit hook-up, an 18-wire ribbon cable is run to the tactile unit: 16 wires for the BCD signals, 1 for common ground, and the last for BDCC.
The circuit is for a four-digit read-out to be used where all four BCD outputs from the counter are available.
All 16 dots go through 22 MEG resistors to ground, with the back plane being connected to BCC. An AND gate combines each dot signal with its respective output from the BCD lines. The required 16 input gates are contained in four packages, Type 4011 Quad 2-input and an AND gate chip; one package is used for each column or digit. Pin 14 of all four 4011's go to BCC, and pin 7 of all four goes to ground. Dot 1, the "1" digit, goes to pin 1 of a 4011; pin 2 of this gate goes to the "1" output of the corresponding BCD line. Dot 2 goes to pin 5, with pin 6 going to the "2" connection on the BCD. Dot 3 goes to pin 12, with pin 13 going to "4" on the BCD. Dot 4, the "8" digit, goes to pin 8, with pin 9 going to "8" on the BCD. The above arrangement is repeated for all four digits.
Now, to combine all 16 of the above gates we use the 8-input NOR gates which are further combined in a 2-input OR gate. The 8-input NOR gates chosen are contained in type 4068 packages. Pin 7 of both 4068's is grounded, and pin 14 of both goes to BCC.
Pins 3, 4, 10, and 11 are the "1", "2", "4", and "8" output pins in each 4011. These pins on package 1 go respectively to pins 2, 3, 4, and 5 on the first 4068. These inputs on the second 4011 go to pins 9, 10, 11 and 12, respectively, on this same 4068. The same procedure is followed in connecting the outputs of the third and fourth 4001 to the second 4068. Finally, pin 13 of the first 4068 goes to pin 1 on a 4001, while pin 13 on the second 4068 goes to pin 2 of the 4001. (The 4001 is actually quad No.2-input NOR gate). Pin 3, the output of the first gate, is the 4001, goes to pin 5 and 6 on the same chip, with pin 4 being considered as the output of the final "OR" gate.
Pin 4 of the 4001 goes through 47K to the base of general-purpose PMP transistor whose collector is grounded. The emitter goes to center tap of a transformer (500 ohm CT primary to 8 ohm secondary). One end of the transformer goes to BCC, while the other end goes to 0.1uF to the base of the transistor. The transformer secondary to the voice coil of a small speaker. An alternative might be to form an oscillator using the two remaining gates on the 4001 package, although this would necessitate building an additional audio amplifier. Otherwise, pins 8, 9, 12, and 13 are unused inputs which are grounded.
This read-out system is for use with counters where the BCD line is multiplex. In this type of counter, a single BCD output (4-wire line) is available and the digits of the visual display "time share" the information over this BCD line. In other words, in the course of a full cycle of the BCD drive signal, information about all the digits is presented sequentially. The multiplex rate is usually in the order of 300Hz or so. Getting the correct BCD information to the right display digit is controlled by multiplex drive signals.
Data connections to our read-out unit include four wires for the BCD line and four multiplex signal leads. Thus, a ten wire ribbon cable is run from the counter to the tactile read-out counter. Eight for the signal leads, one for ground, and one for BCC.
In this case, we need to have three signals appearing simultaneously for a dot to sound its tone: a multiplex signal, a BCD wire, and of course, a dot signal. To combine these signals, we use triple-input gates (Type 4023 Triple 3-input NAND gates). We still need 16 such gates for our dots, and as we usually have three to a package, we will need six 4023's. However, one of the leftover gates will serve to combine the two 4068's. Finally, because the multiplex signals are at a good audible frequency, we do not use a gimmick type oscillator at the output; we run the final gate directly into an audio power amplifier.
While this circuit, in concept, is quite simple, it is somewhat more difficult to put into words. Instead of giving pin numbers as I go along, which would lead to absolute confusion, I shall assign designation numbers to the columns and dots, and let the builder pick which gate in what package he wishes to use. Since the dot represents "1", "2", "4", and "8", they shall bear these respective signals, and these designations will also apply to the corresponding BCD leads from the instrument. The columns will be designated A, B, C, and D from left to right (in case of a transceiver, these would represent hundreds, tens, units, and tenths, respectively). The multiplex lines which also carry these same letter designations. Viewing the tactile display panel as a matrix, the gate associated with the top left dot will have the designation 1A, while the gate for the bottom right dot will be 8D, and so on.
As before, the back plane goes to VCC. Each dot goes through 22 meg ohms to ground, as well as going to the first input of its respective gate. The second input of all the gates associated with dots "1" (the top row or the top of all four columns are tied together and go to "1" on the BCD line). All four second inputs of the number "2" gates to the "2" of the BCD line, and this is similarly done for the "4" and "8" gates. All gates associated with the A column have their third input going to the multiplex line. Similarly, the B column, C column and D multiplex lines go to the third inputs of their appropriate gates which are so designated.
Eight outputs of the gates in columns A and B go to the 8 inputs of the first 4068, while the outputs of the gates for columns C and D go to the inputs of the second 4068. Now, we still have to combine the outputs of the two 4068's. Two of the 4023 gates are, as yet, unused. The outputs of the 4068 each go to an input of one of the free 4023 gates; the third input of this latter gate is unused and goes to VCC. The final gate (4023) is not needed, and has all three of its inputs grounded.
The output of the gate which combines the 4068's goes directly to pin 3 of a LM386 power amplifier. Pin 5, the output, goes to the positive end of a 25uF 6V capacitor, with the negative end going through the speaker to ground. Pins 2 and 4 of the 386 go to ground, and pin 6 goes to VCC. Pin 7 of all the logic chips are grounded, while their pins 14 go to VCC.
4001 Quad 2-input NOR, 4011 Quad 2-input NAND:
- Pin 7--ground
- Pin 14--VCC
- Pin 1 , 2--IN 1
- Pin 3--OUT 1
- Pin 5 , 6--IN 2
- Pin 4--OUT 2
- Pin 8 , 9--IN 3
- Pin 10--OUT 3
- Pin 12, 13--IN 4
- Pin 11--QUT 4.
4023 Triple 3-input NAND Gate:
- Pin 7--ground
- Pin 14--VCC
- Pin 1, 2, 8--IN 1
- Pin 9--OUT 1
- Pin 3, 4, 5--IN 2
- Pin 6--OUT 2
- Pin 13, 12, 11--IN 3
- Pin 10--OUT 3.
4068 8-input NOR Gate:
- Pin 7--ground
- Pin 14--VCC
- Pin 2, 3, 4, 5, 9, 10, 11, 12--IN
- Pin 13--OUT.
Portable entertainment systems are very popular nowadays. I, like most blind folks, spend an inordinate amount of time either on public transportation or going by "Ankle Express". I deplore this colossal waste of time during which I could be edifying my cultured little self. I set out to find a monaural earphone which I could use in an active mobility situation: i.e. one which would not occlude my hearing.
A class of phone is available which uses an efficient moving-coil loudspeaker with a comparatively large diaphragm. Being fairly large, the diaphragm can move enough air to be heard without a path of tight acoustical coupling between the unit and the ear canal. Admittedly, best base response is gotten with good coupling (creating a chamber around the ear and treating the speaker baffle as infinite). However, opening up the ear for business (sacrificing the coupling chamber) does not entirely ruin the earphone for reproduction of music, and the intelligibility of speech remains excellent.
So far, I have unearthed two candidates. One is the Sony ME100, priced at about $6.00 and available from IRTI* (IRTI, 375 Distel Circle, Suite A11, Los Altos, CA 94022; Telephone: (415) 965-8102.) The second is the Radio Shack "Novaphone", No. 33-1036, priced at $7.00.
The Sony ME100
This unit houses the "speaker" element in a snap-together poly ethylene capsule. An outer ring of the same plastic loops over the ear of the wearer. Finally, a foam rubber sleeve is stretch-fitted over a bezel on the front of the speaker capsule.
Modification for my application is easy; slip the foam rubber ring off and discard it. In fact, if you were to want the foam rubber to stay attached, you had better secure it in place with a dab or two of rubber cement.
With the back of the unit simply being snapped on, this device lends itself to some interesting modifications; for example, a series resistor and/or low-pass filter can be "haywired" right on to the back of the speaker element.
Pad Circuits for the ME100
When this earphone is used with equipment whose output stage is noisy or whose volume control is very touchy, you can significantly increase listening comfort by inserting an 82 ohm 1/4 watt resistor in series with the hot lead. If this amount of attenuation is objectionable, try 33 ohms.
I find it painful to monitor recordings being made on some cassette machines, especially the G.E. units marketed by APR. Not only is the level exceedingly loud, but the equalization pre-emphasis is present on the monitor jack; the high frequency content makes the program material unlistenable. You can make an earphone especially for these machines as follows:
The hot lead goes through two 43 ohm resistors in series to the speaker element, with the junction between these two resistors going through 1 UF (small tantalum) to the cold side of the element. This gives you a drop of about 20 DB with a roll-off frequency of about 7500 HZ.
This earphone is a plastic cup (Lexan) which is draped over the ear of the wearer. It-has roughly the shape of a rectangle whose top end is rounded off; it is arch-shaped. A regular one-inch loudspeaker is housed in a plastic capsule on the back of the phone. A pulley-shaped nub along one side serves as a cable strain-relief device. To improve acoustics within the cup, a piece of thin foam rubber is cut so as to conform to the inside. Finally, perspiration vents are cut around the top of the "arch".
The foam lining is easily extracted and discarded. Modification then consists of cutting away all of one side of the cup.
With a coping saw, start cutting at the third perspiration vent ahead of the center line. Partially undercut the arch, then make a U-turn with the saw and follow the speaker capsule around to the cord's entrance. From this point on, saw straight toward the bottom of the rectangle. (Note that if this is to become a left-eared device, the side opposite the strain-relief nubbin is cut away. However, if a right-eared version is desired arrangements will have to be made for replacing the strain-relief on the other side).
With a rat-tail file, further undercut the arch by following the outline of the speaker capsule; take care not to significantly mar the capsule and/or the rim of the arch which hangs the device over your ear with other appropriate files and emory cloth, round off all corners and edges.
The plastic cap which holds the speaker to the phone is glued in place. However, if done carefully, the cap can be pried off to repair the cord and/or to insert a pad network as described earlier. The cap must then be reglued into place.
After using both units for over six months, I still cannot make up my mind as to which I prefer. They both have good and bad points as follows:
The ME100 stays in place very well on the ear; it's just the thing for mobile coffee drinkers. However, at difficult street crossings, you need a completely free hand to remove it. On the other hand, the Novaphone can be toppled off quickly and easily in a perilous traffic situation.
The ME100 makes a surprisingly efficient low-profile pillow speaker. Novaphone is the wrong shape for this application.
If you use your ears extensively for echo location of buildings and obstacles, keep the program material turned down to a low volume, or limit its frequency range to the absolute minimum. High frequencies are those most used in echo location, and high frequency components in the program material will "jam" this band and put you out of commission. If you listen to FM or other hi-fi material, keep it level as low as is usable.
These may not be the best earphones around; please let me know if you come across any better ones. Perhaps you can even make your own. For example, a very intelligible low-fi earphone can be made by mounting a hearing aid phone on spectacle frames and aiming the sound at your ear via a curved piece of tubing.
Good reading, current event listening, dancing, or whatever your pleasure might be.
The Rehabilitation Engineering Center at the Smith-Kettlewell Institute has just published a catalogue of vocational and educational aids which have been developed by the Institute over the past six years. Most of the aids have been tested in the field by blind consumers and many of their comments appear under the device headings. Information on how to obtain each device is included together with approximate costs of parts. The catalogue is available in both braille and standard-size print form and is free of charge to Technical File subscribers.
For your copy, write or call Liz Auer, Smith-Kettlewell Institute, 2232 Webster St. San Francisco, Ca. 94115. (415) 561-1619. Please specify whether you would like to receive a braille or print copy.
Correction--"Waiter! What's this needle doing in my soup?!"
"No sir, that's a noodle; it was a misprint."
Making career choices is not easy for anyone. For the young blind person, making hobby and vocational decisions often grows into a contest of wills. Every proposed activity is challenged by somebody--a relative, a teacher, a counselor, a college dean, and every employer will perennially be asking pointed questions, "How can you possibly expect to ...?" Eventually, you ask these questions yourself, and the answers may not immediately be apparent.
When the choices being made are major, there are two things you can do; you can change your interests so as to take a path of less resistance (we have all done this to some degree), or you can put off the questions in hopes that answers will become apparent later on and arm yourself as best you can for a battle of wits. Both the above alternatives are frustrating and exhausting. Where can the necessary energy be gotten?
You are lucky indeed if pioneers have come before you to blaze a trail in your field. They have answers to some of the critical questions, and their success lends legitimacy to your aspirations. In my case, there were questions like, "How can you do electronics without an oscilloscope?" Provisional answers could be gotten from the Braille Technical Press in "Information for the Radio Service Man" by Bob Gunderson and the test equipment articles (distortion analyzers and the like by Jim Swail. But there were other stabbing questions like, "How can you possibly keep up with the reading?" and -"How do you intend to communicate your design information to other members of the work team?" To my critics and to myself, all I could say was, "look" there's a guy in New York who builds and repairs equipment all the time, in addition to which he runs a magazine, and there's a full-fledged engineer in Canada; surely these people have worked out some of these things."
Fifteen years later, reading is still a problem, communicating with colleagues is done verbally as it was in school, and we now have a few more measurement instruments. What saves my bacon is that in a real-time work situation, these issues come up one at a time. Now I am a full-fledged engineer is San Francisco, and I edit a magazine. Had it not been for Gunderson and Swail, I would have been less well-armed for a battle of wits, and I would have faired badly in a contest of wills.
It is not without emotion that I take stock in the fact that the two people who were most influential in my choosing engineering as my career have, as of this issue, both written material for my magazine. I am deeply moved by this honor.
I am not directly involved in the preparation of address labels and the mailing process (which is lucky for you). However, as the boxes of magazines wait to be whisked out the door, I cannot resist button-holing someone to read a few of the envelopes. To me, the mailing list reads like a "who's who"; anyone who comes anywhere near a technical magazine has, in some way, survived a contest of wills and has put up a fair fight in defending the pursuit of his interests. I recognize most of your names; I have either gotten letters from you, talked to you on the phone, read your name in an old list of blind hams, or read your articles in the BTP.
For me, running this magazine is like reliving a favorite all-star ball game, only this time, I'm in it.