1996 - 2016
BSEE - KB3UON
RF Cafe began life in 1996 as "RF Tools" in an AOL screen name web space totaling 2 MB. Its primary purpose was to provide me with ready access to commonly needed formulas and reference material while performing my work as an RF system and circuit design engineer. The Internet was still largely an unknown entity at the time and not much was available in the form of WYSIWYG ...
All trademarks, copyrights, patents, and other rights of ownership to images and text used on the RF Cafe website are hereby acknowledged.
My Hobby Website:
November 1929 Radio-Craft[Table of Contents]
People old and young enjoy waxing nostalgic about and learning some of the history of early electronics. Radio-Craft was published from 1929 through 1953. All copyrights are hereby acknowledged. See all articles from Radio-Craft.
Vreeland Corporation was an early radio manufacturer located in Hoboken, New Jersey, with multiple patents on file for innovative circuits. The Vreeland band selector system mentioned here was originally filed in the U.S. Patent and Trademark Office in August of 1927 and had not been awarded by the time of this November 1929 article in Radio-Craft magazine. In fact, it wasn't until five years later, in 1932, that the patent was finally assigned. The official description reads in part, "The general purpose of the invention is to receive the component frequencies of such a band with such uniformity as to avoid material distortion of the modulated wave, and to exclude frequencies outside of the band which the system is designed to receive. Another purpose of the invention is to provide means for shifting the position of the band in the frequency scale at will, by a simple adjustment, so that the system may be readily adapted to receive modulated waves of any desired carrier frequency, including the side bands of such modulated waves." That sounds to me like a standard heterodyne system with selective filters. It seems the filter characteristic with a wider inband region and sharp cutoff is what make it unique. One line in the patent states, "... a radio receiving system embodying a plurality of my [emphasis added] band selector units, one being associated with a collector ..." Note the use of the possessive "my," by Frederick K. Vreeland, which I doubt would ever appear in a contemporary patent filing. Mr. Vreeland authored books such as Maxwell's Theory and Wireless Telegraphy as well as Camp Buildings and Scout Shelters for the Boy Scouts of America.
A super-sensitive set of simple construction incorporating the latest principles in radio design.
By Clyde J. Fitch
The modern factory-built radio set is so far advanced in design that one hesitates now to construct his own; yet with a little care and patience a commercial-looking job can be turned out at home that will excel the latest commercial models. The receiver shown in the accompanying illustrations was designed for sensitivity, tone quality, and ease of operation; while it incorporates new features not found in other receivers. It was built for installation on the Florin Lenox estate in the Adirondack Mountains, a location which is immune from interference of any kind except an occasional thunderstorm; consequently unusual selectivity was not required. However this set compares with the best in selectivity.
It comprises a Vreeland band-selector system which feeds into a three-stage untuned R.F. amplifier using A.C. screen-grid tubes and iron-core untuned radio-frequency transformers. A screen-grid detector is used with one stage of resistance-coupled audio amplification and one of push-pull using type '10 power tubes. This makes seven tubes in all, four screen-grid type '24, one type '27, and two '10's; not counting the '81 rectifier and '74 voltage regulator used in the power pack.
The band selector, the theory of which is explained below, passes a 10-kilocycle band of wavelengths at any location in the broadcast spectrum, depending upon the adjustment of the tuning dial. This allows the carrier-wave with its accompanying "sidebands" (which represent the music), to pass through from the antenna to the amplifier uniformly, giving distortionless reception in this part of the set. This is true also of the radio-frequency amplifier, which employs the old Acme iron-core transformers, used several years ago in that company's reflex sets. These transformers give very efficient amplification over the entire broadcast band, and associated circuits do not oscillate ("'whistle," howl, etc.) at any point on the dial when used with screen-grid tubes. Two type-R3 transformers, (RFT1 and RFT2 in Fig. 1) with one type-R4 (RFT3), gave the best results; the type-R2 was found unsatisfactory. No doubt there are other iron-core radio-frequency transformers that may be used instead, in case the constructor has difficulty in finding the Acmes.
The use of a screen-grid detector with a stage of resistance-coupled audio amplification is another feature that gives high, distortionless amplification.
Here (in New York City) the set picks up locals with an indoor antenna two feet long; with no antenna it picks up nothing, since it is well shielded. Quality is unusually good, for the reasons mentioned above, and operation is simple, as the illustration (Fig. A) shows; one knob (C2-C3) being used for tuning, and the other (R5) for a volume control (which is easy to regulate because it does not detune the set or cause it to oscillate and squeal or howl).
In the illustrations, Fig. A shows the front view; Fig. B the top with the covers of the shield cans removed; Fig. C the rear, and Fig. D the power pack. The panel was cut to the size shown, merely to fit a special cabinet.
Construction of the Set
The metal base should be made first; that illustrated here was a sheet of monel metal 13 1/2 x 29 1/2, inches and 1/32-inch thick with the corners cut out to a depth of an inch each way and the sides bent down and soldered at the corners, making a "pan"11 x 27 x 1 1/4·inch deep. This work should be done by a tinsmith. Aluminum may be used instead, but it should be at least 1/16-inch thick, as it is mechanically weaker than monel metal; since aluminum cannot be soldered easily, the corners should be strengthened with brass angles bolted in place. Even a good wooden baseboard may be substituted.
Dimensions for drilling the base are not given, as these will depend upon the particular parts used. The general layout shown in Fig. B can be followed without difficulty. The sockets are mounted with their terminals underneath; so that practically all the wiring is below, out of sight and well protected. With the exception of the socket holes, the drilling of the base may proceed as the set is being assembled, by the use of a hand drill.
The Band Selector
The band-selector coils L1 and L2 consist each of 85 turns of No. 28 wire wound 2 inches in diameter, with a center tap, as detailed in Fig. 2. (Hammarlund space-wound coils were used.) They are clamped by two strips of Bakelite, which carry the three terminals for the two end connections and center tap. Brass angles are used to mount them in the proper relation to the metal base, inside the shield cans as shown.
The two 0.00035-mf. tuning condensers C2, C3 are mounted on the shield cans as illustrated; both on the same shaft and with a single-dial control.
The 23-plate "midget" antenna coupling condenser C1 is mounted on an insulating strip of Bakelite which is fastened to the upper left-hand corner of the right-hand can, (Fig. B). The 9-plate "midget" trimming condenser C4 is mounted directly on the other shield can, as shown in the same figure.
The coupling coil L3 consists of 4 1/2 turns of wire wound on a vacuum-tube base. This is plugged into a standard socket, mounted on the set base as shown. The two ends of the coil are connected to the filament prongs of the tube base. This completes the band-selector parts; wiring and adjustment will be described later.
Many explanations have been published about this new method of radio tuning, but the practical man wants more practical data and a clearer understanding of the theory without wading deeply into mathematics; therefore, let us begin at the beginning. and explain why bandpass tuning is necessary for undistorted reception.
Suppose we start at the transmitter, but forget the old idea of modulated carrier waves and look at the situation from a different viewpoint; a viewpoint where even the layman can get a clear insight into radio transmission. From our method of heterodyne reception (as used in the supetheterodyne receiver) we know that when two alternating currents of different frequencies, F and F1, are combined, two other frequencies are produced, equal to (F + F1) and (F -- F1); making four distinct frequencies in all. (Harmonics of these will also appear, but they are useless for our purpose.)
"Modulation" Simply Explained
Since radiation of electric energy from the transmitting antenna must take place at very high frequencies to be efficient, a high-frequency generator is used at the transmitter. For broadcasting, frequencies from 500,000 to 1,500,000 cycles per second are used, each station having its own assigned operating frequency. This frequency is called the carrier-frequency, or "carrier wave." We will call this frequency f. Now, suppose we combine with this carrier frequency, f, the sound-frequencies (or music and speech-frequencies produced in the studio) and see what happens; keeping in mind that the music- and speech-frequencies range from about 50 to 5,000 cycles per second. We will call these the audio-frequency band, or just "AB."
From the above heterodyning action, we learn that four distinct frequencies will result; namely, the carrier-frequency f, the audio-frequency band AB, and the bands (f + AB) and (f-AB).
The audio-frequency band AB will not be radiated from the antenna, because its frequencies (50 to 5000 cycles) are too low for efficient radiation. The carrier frequency, f, will be radiated as will also the frequencies (f + AB) and f - AB). These latter two are called the sidebands, and they contain all the music and speech, because they contain the audio-frequency-band component AB.
In explaining radio transmission to a layman, we can simplify this still further and merely state that radiation takes place at very high frequencies; therefore at the transmitter we add a high frequency, say a million cycles, to the sound frequencies, so that they can be radiated, and later remove the high frequency at the receiver. The ordinary person will grasp this heterodyne idea much quicker than he can grasp the modulated-carrier-wave idea.
We see from the above that a group of frequencies (namely, f, f + AB, and f - AB) are radiated from the broadcast transmitter having a maximum difference of I plus and minus 5000, or a total separation of 10,000 cycles, or 10 kilocycles. For example, using a carrier of 1,000,000 cycles, a band of frequencies from 995,000 to 1,005,000 cycles will be radiated. Our receiving set must therefore tune in all these frequencies at once in order to receive all the music; or the sopranos will be lost in the ether, and this might not always improve the programs. Hence, the use of a band-pass or selector.
The ordinary radio set must be selective in order to tune in a single station without interference from others. Tuning must be sharp. But sharp tuning cuts the sidebands; it weakens or eliminates the higher audio frequencies, and causes the music to sound deep and muffled. Instead of tuning in the complete ten-kilocycle band of waves, it covers a band of only about four or five. This is shown in the response curve of Fig. 5, which only has good response over a 4-kc. band. Each tuned circuit in the receiver has one definite "peak" where the response is maximum, and this is adjusted to be in resonance with the carrier-frequency.
Design of the Selector
The band-pass selector has two tuned circuits, loosely coupled together. The two circuits give two peaks, adjusted so as to be dose together, making a somewhat flat-topped graph 10 kilocycles wide, as shown in Fig. 6. The curve has two small humps at the top c used by the two peaks of the two circuits. The sides slope down about as in the curve of Fig. 5, but the width at the base, using carefully-designed circuits, is no greater than in the former curve. Therefore, selectivity is as good when using the band selector as it is with the ordinary method of tuning, and it responds to the entire band of transmitted frequencies more evenly.
Fig. 7 shows the circuit diagram of a band selector; this is the system used in our complete schematic, Fig. 1. The two tuned circuits are indicated at A and B, and are coupled together by the common inductance L. The circuits A and B have each the same sizes of coils and condensers, and are designed to cover the broadcast range; therefore any good broadcast coils and condensers may be used. Magnetic coupling must not exist between the coils; therefore they must be thoroughly shielded.
The degree of coupling is determined by the value of the inductance L; this, also, determines the spacing between the peaks of the two resonance curves, and consequently the width of the band of frequencies that it will pass. If L is increased, coupling is increased and the width of the curve, Fig. 6, is increased. If L is decreased, the width of the curve is decreased. The correct value of the coupling inductance L will give us the desired 10-kc. width. 4 1/2 turns, wound on a tube base, is about the correct number. Several such coils may be made, ranging from 2 1/2 to 6 turns, and are easily interchanged by mounting them in a standard socket. The smaller the coil, the greater the selectivity. You can therefore adjust the selectivity to suit your particular requirements.
The width of the curve (Fig. 6) does not remain constant over the entire broadcast band, but varies with the frequency. Using the circuit of Fig. 7 with inductive coupling, the width increases with increase of frequency; because the reactance of the inductance L is greater at the higher frequencies, and the degree of coupling is greater. Hence it may pass only the correct10-kc. band at the upper wavelength settings, but a 10-kc. band at the lower. As this will be objectionable in some localities, the system shown in Fig. 8 may be employed instead.
The only difference is that the two circuits are coupled by a common capacity C, instead of by an inductance. (When using the coupling condenser C, the grid circuit of the first tube in the R.F. amplifier will be open. The resistor R, of about 2 megohms, should be connected across the condenser C.) This gives just the opposite effect to that obtained with the inductance; the larger the capacity, the less the degree of coupling, and vice versa. A coupling condenser of 0.025-mf. is about the correct size. This also may be mounted into a tube base, together with the resistor R, and plugged in the socket in place of the coupling coil.
With capacitative coupling, the coupling decreases with increase of frequency. Therefore the selectivity is greater at the lower wavelengths than at the higher, which is just the opposite effect to that obtained with the inductance. Different sizes of condensers may be tried to get the correct balance. (Perhaps a combination of capacity and inductance may be found to give a 10-kc. curve through the entire broadcast band.)
Summing up, the receiving aerial picks up all the stations on the air, but the band selector allows only the band of frequencies transmitted by anyone selected station to pass through it and into the R.F. amplifier.
R.F. Amplifier and Detector
The radio-frequency amplifier and detector is a complete unit in itself and wired independently. The mounting of the apparatus is indicated in cross-section in Fig. 3, and the reproduced photographs clearly show the unit and its location on the base. The copper shield can (S3) of a Remler "Infradyne" intermediate amplifier was used by the writer, and was just the correct size for inclosing the parts; a metal top was required, however, measuring 4 by 16 inches. All the parts are mounted on this top, so that it can be removed for inspection or repair without removing the can from the base. The can is 2 3/4 inches deep inside, and can easily be made up of sheet copper.
The sockets are equally spaced on the metal top, as shown, while the radio-frequency transformers are mounted below them and away from the can. Below the detector socket is fastened a strip of bakelite supporting one of the by-pass condensers; the other condensers are soldered together and mounted between the sockets. Eleven one-microfarad condensers in all are used in this unit. Condensers of this capacity were selected to by-pass thoroughly all the connections in each stage, so there would be no tendency to oscillate due to coupling of the common connections. In ordinary sets this effect is negligible; in "high gain" receivers using screen-grid tubes "common coupling" becomes a serious problem.
The wiring of this amplifier is clearly shown in the complete schematic diagram, Fig. 1. The space enclosed by dotted lines S3 indicates the connections inside this unit.
The 100-ohm resistor R1 , between the R.F. cathodes and ground, gives a 1 1/2-volt "C" bias on the R.F. control grids. The 5000-ohm resistor R2, connected between the detector cathode and ground, gives 5 volts bias on the detector control grid. These resistors are mounted in the amplifier unit but are not shown in Fig. 3.
Connections to the amplifier unit are brought out through holes in the bottom. There will be two heavy leads for the 20 1/2-volt A.C. supply to the heaters. Large flexible wire should be used for these, as the current for the four tubes will be 7 amperes.
There are four other connections to the amplifier, colored preferably as indicated in the diagram; blue for volume control (R5) regulating the screen-voltage of the R.F. tubes; brown, for detector screen-voltage control (R6); green for the detector plate; and red for the 180 volt "B" supply. The shielded control-grid leads are brought out through the top, as shown in the photographic illustrations. The first one comes from the band-selector shield can, as will be seen from Fig. B.
Suitable untuned transformers for the R.F. amplifier can be constructed according to the illustration, Fig. 9. The core consists of strips of very thin silicon steel transformer laminations 1/2" wide by 2 3/8" long, stacked up to a thickness of 3/16", as shown. On each side of this is placed a wooden form 5/32" by 1/2" by 2 3/8" long, so that the finished core assembly is 1/2" square. It is bound together with a few wrappings of waxed paper.
The primary winding is "random wound" over one half of the core length. No. 38 S.C.C. wire is used. (This is indicated in the illustration for clearness as a single layer.) The secondary is also of No. 38 S.C.C. wire wound in ten sections, or "pies," equally spaced over the entire length of the core as shown. A layer of waxed paper is placed between the two windings.
The wavelength band which the transformer efficiently covers depends upon the number of turns in the primary and secondary. For the first and second stage units, RFT-1 and RFT-2, the primary should have 35 and the secondary 83 turns.
For the third stage, RFT-3, the primary should have 26 and the secondary 43 turns. As there will be some variation in these transformers due to different qualities of iron, it may be necessary to add or subtract wire from both primary and secondary to obtain the best results. This can easily be done experimentally after once operating the set to determine if it amplifies uniformly over the entire broadcast range.
The Audio Channel
The audio amplifier comprises one stage using resistance-capacity coupling (V5) and one with push-pull (V6-V7). Resistance-capacity coupling was selected for the first stage, because it is the most efficient method when used with a screen-grid detector. The plate-circuit resistor R3 used at a plate voltage of 180 is 250,000 ohms, which allows about one milliampere to pass through the detector with a grid bias of 5 volts and a "screen" voltage of 75. The screen voltage is adjusted by the 500,000-ohm potentiometer, R6, mounted on the base. When once set, this requires no further adjustment.
The plate resistor R3, grid leak R4, and 0.05-mf. coupling condenser C17 are mounted on the base, as shown. The locations of the '27-type A.C. tube, the two '10s, and the input push-pull transformer AFT are clearly illustrated.
Underneath the base are mounted the detector R.F. choke L4; the 0.0005-mf. by-pass condenser C16; the 1500-ohm "C" bias resistor R7, for the first audio tube; the two 50,000-ohm stabilizing resistors R8 and R9, in the push-pull grid returns; and the by-pass condensers. A 1-mf. condenser C18 shunts the 180-volt and ground terminals; 2-mf. (C19) is the by-pass capacity for the 90-volt output; and 1-mf. (C20) by-passes the first-audio "C" bias resistor.
The two plate terminals 1 and 2 of the output tubes are mounted on a Bakelite strip at the right end of the base, while the other terminals are along the back, arranged as indicated in Fig. B.
The center tapped output choke used with this set is mounted in the cabinet with the reproducer. It has a value of 400 henries and a carrying capacity of 60 ma. (To prevent parasitic oscillation the halves are balanced at the factory for absolute electrical symmetry on both sides of the center tap.)
It is fitting to observe at this moment that a primary fault in average push-pull circuits is that malady variously known as "feedback," "interference," "harmonics," etc.; in plain words, parasitic oscillation. It manifests itself in an above-normal plate current; and it may, or it may not, be accompanied by a high-pitch whistle, and distortion. It does not make any difference whether it causes distorted reception, the idea is that a fault exists which is going to cost the owner some money unless it is remedied. As exceptional precautions have been taken in the design of this receiver to prevent parasitic oscillation in the push. pull stage, there is almost no likelihood of trouble from this source. In addition to measures mentioned above, the author points out the use of two stabilizers or oscillation suppressors, R8 and R9, in the "grid return" lead of each power tube. This connection is possible only when the input transformer secondary has a balanced, two-section winding. (To by-pass these resistors would be to defeat their purpose. The "grid bias resistor" R14 serves an entirely different purpose and it is necessary that this unit be properly by-passed, - the purpose of C25.) A single resistor in a single return lead would be "common" to both tubes and the "isolation" desired would not be obtained.
By carefully following the diagram, Fig. 1, you should have no trouble in wiring the set. Be sure to use large cable connections for the 7 1/2-volt filament and the 2 1/2-volt heater supply. The practice of transposition is followed with all filament leads; that is, they are twisted to eliminate any possibility of an induction hum. The metal base is grounded, also the "B-" lead from the power pack. The band-selector connections should be as direct as possible. The leads should be well insulated and securely soldered. After the panel is put in place, the condenser shaft lined up and dial mounted, the set is ready for operation.
Type '45 or '10 Power Tubes?
During the early stages of the design, the question arose as whether type '45, type '10 or type '50 tubes should be used. Each has its advantages and disadvantages. (The constructor is recommended to experiment with the necessary power equipment for type '45 tubes, where the power-handling ability of the type '50 tube is not needed for driving dynamic reproducers.) Type '10 tubes, however, were selected because it was intended to use an electrostatic reproducer, the polarizing potential for which would be available with the operating voltages of type '10 tubes; whereas the use of type '45 tubes, which do nor operate at such high plate potentials, would necessitate a special "polarizing" unit for the electrostatic reproducer. (The "undistorted power output" of two '10s is about the same, at the voltages shown, as that of two '45s at 250 volts, plate.) The question of single or push-pull operation was settled almost instantly; as the push-pull arrangement, on many counts, is far superior to a single tube. It will be noticed that coupling condensers have been dispensed with; this arrangement is possible only when the output is push-pull, for the direct current (which would ordinarily circulate through the reproducer winding or matching transformer primary, when contact is made directly to the plate - in the case of "dynamic" and "electro-magnetic" reproducers), is balanced out by this connection. What the constructor is most interested to know is that the frequency-discrimination of coupling condensers has been eliminated and better reproduction results.
Where "electrostatic" reproducers are used, it will be an advantage to connect directly to the plate, as will be observed from consideration of the output connections of Fig. 1. Of course, the use of a dynamic reproducer is optional; its two leads being connected to output posts 1 and 2.
The Power Pack
The power pack (Fig. D) is also mounted on a metal base. This base (when shaped) is 10" x 18"x 1 1/2" deep. Only one illustration, besides the diagram (Fig. 4) is given as this unit is comparatively simple to build. The list of parts gives the details of the apparatus used in the power pack, all of which is mounted on top of the base.
Each of the filter condensers (C21-22, C23-24) consists of two 4-mf. units. Two of these are connected in series, giving a total capacity of 2 mf. across the highest-voltage side. One of the other 4-mf. sections is connected across the 425-volt lead and the fourth across the resistance bank.
The latter consists of one 3000-ohm, 100 watt, resistor R10; one 2000-ohm, 100-watt, resistor, R11; one 3000-ohm, 20-watt, R12; and one 4500-0hm, 20-watt resistor, R13; all connected in series as shown and mounted on a small Bakelite panel. A 1000-ohm, 100-watt, "C" bias resistor R14 is also mounted on this panel, and by-passed by the 1-mf. condenser C25. Be sure to use well-insulated wire for making all connections.
Resistor R10 functions as an over-all voltage control, "absorbing" the potential-difference between 180 and approximately 400 volts.
Choke L5 has a greater current carrying capacity than L6. Its rating is 120 ma., 20 Henries, and D.C. resistance of 210 ohms; L6 being rated at 60 ma., 50 Henries and 600 ohms. It will be seen that the power tubes derive their plate supply through L5, but not through L6. This connection results in absolutely humless operation (so far as the filter is concerned) when the output tubes are in push-pull; a little more filtration being desirable for a single-tube-output design.
The 110-volt A.C. connection to the power transformer passes through the switch SW on the panel of the set. The filament connections between the power pack and the set should be made with large wire; so that there will be no appreciable voltage drop. The pack should not be placed too close to the set; because A.C. hum may be picked up by the set if it is placed in the magnetic field of the power transformer.
This completes the assembly of the parts, with all the apparatus mounted, the radio-frequency amplifier unit all wired and its leads coming out the bottom, through holes in the base.
Adjustment and Operation
With everything hooked up properly, simply turn on the switch and the set is ready for operation. To adjust the band selector, tune in a station and turn the volume down. Then set the antenna-coupling condenser C1 at maximum and retune the station while changing the trimming condenser. For maximum sensitivity, the two tuned circuits must be in resonance; and this condition is obtained by carefully adjusting the trimming condenser C4. A screwdriver may be used to adjust the two "midget" condensers by cutting holes in the shields and slotting the shafts of the condensers,
If the set is broad in tuning, decrease the capacity of C1 and readjust C4. In this way, the correct capacity for your particular antenna can be found. A trimming condenser for the first tuned circuit has been found unnecessary; but one may be used to facilitate the adjustment. The adjustment should be tried at both upper and lower ends of the scale.
The tubes also may be shifted around, for they vary slightly, and some work better as a detector than others. The detector screen voltage should also be adjusted for maximum sensitivity.
Using a good electrostatic or dynamic reproducer, this set will give unusually good tone quality with very little A.C. hum; and you will be well pleased with the ease in operation and sensitivity. With the '10 push-pull amplifier you can obtain sufficient volume for any occasion. If preferred, '45 tubes may be substituted by making changes to suit in the power pack.
Acknowledgement is made here of the courtesy of Mr. R. H. Siemens, chief engineer of the Radio Construction Laboratories, who kindly provided laboratory facilities and technical aid during the design and construction of the "1930 Electric Receiver."
Parts Required for "1930 Electric" Receiver
Posted December 18, 2015