Fullscreen HTML5Med Res (??MB)HTML4

Vintage Radio Experimental one-valve superhet radio By Fred Lever For commercial superhets, specific dual valves were designed to combine the functions. For example: • An RF amplifier integrated with the local oscillator (a ‘converter’ like the 6AN7) • An IF amplifier with diodes for detection and AGC (eg, the 6N8) • An audio preamplifier and output driver in the same envelope (an ‘output valve’ like the 6GV8) That gave the designers some scope for clever circuit arrangements. The 6Y9 was one of the last twin pentodes of the valve era and was used in TV sets. It seemed like an excellent valve to build the single-valve superhet radio. Concept and techniques I have built many superhets with traditional circuit techniques, using three or more valves. However, I was intrigued when I saw a suggestion that the 6Y9 dual valve for TV sets could perform the active functions required to make a complete superhet radio from antenna to speaker. I googled “one valve superhet” and, sure enough, many people have been there before me. However, each arrangement I found did not quite meet the requirements of a practical home radio set, or they used an uncommon valve. I won’t go into much superhet theory here as it has been covered extensively in these pages. A study of Wikipedia’s “Superheterodyne receiver” entry (https://w.wiki/8DYV) will fill a reader in on the concepts and explain some of the acronyms used. 80 Silicon Chip The basic principle is to mix the incoming signal with a signal say 455kHz above or below it, then filter out everything except the 455kHz component from the mixer. After that, we can amplify and demodulate that much lower (and fixed) frequency signal. A superhet AM radio can be easily built using three valves: an RF amplifier/mixer/oscillator, an IF amplifier/ detector and an audio amplifier/loudspeaker driver. That is about six functions jammed into those three valves. Australia's electronics magazine I took up the challenge, starting with a draft circuit originated by Ian Robertson. The resulting radio, described in this article, nearly met all the criteria. I mainly used junk-box parts and modified the theoretical circuit to suit the parts I had. This completed radio sits on a shelf and, with an indoor aerial wire, produces a couple of watts of sound through a five-inch (~127mm) speaker and tunes in all the local AM radio stations. The circuit, shown in Fig.1, uses every technique possible to provide the functions mentioned above from the single valve, including autodyne, reflexing, neutralisation and negative feedback. Negative feedback is a commonly used technique these days, but the others may not be that well known. Reflexing is a method of passing the radio signal multiple times through one valve at different frequencies. In this case, the RF amplification, local oscillator and mixing are handled in the first valve section, while the IF, AGC and audio functions in the second. Neutralisation is a form of positive (regenerative) feedback that cancels out unwanted, inherent negative siliconchip.com.au feedback to get more gain from a valve or transistor. ‘Autodyne’ is a very old superheterodyne single-valve technique used in the 1930s, subsequently displaced by the dual-purpose converter valves. Essentially the incoming signal is fed to a valve set up to oscillate at a different frequency, so it acts as both the oscillator and mixer. The theory behind these techniques can be studied by consulting the textbooks of the era, such as the Radiotron Designer’s Handbook. Did I cheat? I cheated a little bit in some people’s eyes by including some solid-state diodes in the circuit. I elected to use diodes for the power supply and the detector functions. The main components in the rest of the set are from the 1960s era or modern equivalents. Still, I think I got away with it because it’s still true to say that the only active devices in the circuit are within that sole valve envelope. The diodes (bridge rectifier) in the power supply are only needed because it’s a mains-powered set; had I elected to make it battery-powered, they could have been eliminated. That leaves the detector diode (D5) as the only part that might have needed another envelope back in the valve era, although other types of rectifiers were available back then, like selenium rectifiers. I used two 1960s commercial IF transformers but scramble-wound the tuning coils on repurposed coil formers. Other parts came from my junk box or the Jaycar stock bin. I certainly used new capacitors and resistors! Practical difficulties The aim of any radio set is to gather radio waves at microvolt (μV) levels out of the air, then select and amplify the signals in a particular frequency range to drive a loudspeaker coil with a few volts at audio frequencies. That implies a level of voltage amplification of thousands of times or more. That amplification is usually spread over a chain of tuned circuits, with amplifying valves interposed at strategic points to keep boosting the signal level. The standard practice is to keep each circuit input wiring well away from the output wiring, to minimise the chance of uncontrolled feedback turning into instability. However, in this set, we surround one valve with those series of tuned circuits, but keep feeding signals back into the same valve position for another trip! It is a fact that, by necessity, the input and output signals of each ‘stage’ are in close proximity. Circuit details We have two pentode sections, V1A and V1B. V1A combines the signals from the aerial coil/transformer and the oscillator coil/transformer. The aerial coil is connected to the control grid input at pin 1, while the oscillator coil is connected to the cathode at pin 2. The valve output at pin 4 has two loads stacked in series. The first intermediate frequency transformer (IFT1) load is tuned to respond only to 455kHz, while the second load, the oscillator coil, only responds to oscillator frequencies (around 1-2MHz). Fig.1: this radio circuit I developed utilises the first half of the dual pentode, V1A, as an Autodyne mixer/oscillator, while V1B is reflexed to act as an IF amplifier as well as an audio signal amplifier to drive the speaker transformer. The only ‘cheat’ is silicon diode D5 as the detector. siliconchip.com.au Australia's electronics magazine July 2024  81 V1A receives a tuned carrier signal from the aerial coil into pin 1, which appears on the plate at pin 4. The plate is also connected to the oscillator coil, which is phased as a positive feedback and is resonant. Feedback goes to the pentode cathode at pin 2. That input signal change accelerates the feedback through the oscillator coil, and the valve bursts into oscillation at the frequency determined by the resonance of the oscillator coil with its tuning capacitors. That oscillator signal also appears at plate pin 4. The plate circuit has a combination of station carrier sine waves and oscillator sine waves, the differences between those two, plus any modulation present. The signal thus looks like an unresolved blur on an oscilloscope, but by sweeping slowly, you can get an idea of the multiple RF waves with an audio modulation sitting on top. Once past the 455kHz trap, the IFT signal Scope 1: the yellow trace is the 455kHz IF signal, while the cyan trace shows the recovered 440Hz audio modulation. resolves a bit better. In Scope 1, the yellow trace is the 455kHz IF signal modulated at 440Hz (the blue signal). Consider a tuned signal carrier at 1MHz being fed into pin 1. An amplified version of this signal appears at the plate, pin 4. The oscillator coil is also connected to the plate through the IFT1 primary. As the oscillator coil acts as a feed-forward from the output (plate) to the input pin 2 (cathode), the circuit oscillates at around 1455kHz, which also appears at the plate, pin 4. There is a difference (beat) frequency of 455kHz (1455kHz – 1000kHz). As IFT1 is a 455kHz resonant trap, any other frequency at the plate of the valve is rejected, and only the 455kHz ‘beat’ modulated by the original audio program content gets through. It therefore arrives at the input control grid of the second section, at pin 8. IF and AF amplifiers Scope 2: the signal delivered to the speaker without the gimmick capacitor; it is distorted and full of RF due to the second high-gain stage oscillating uncontrollably. Scope 3: with the gimmick capacitor added, a couple of picofarads of extra Miller capacitance have increased stability to the point where the set is only oscillating at the desired frequency (455kHz above the tuned frequency), and the detected audio signal is clean. 82 Silicon Chip Australia's electronics magazine The second pentode, V1B, also has two loads stacked in its output plate at pin 10. The top load is a second 455kHz IFT that passes only 455kHz signals and ignores anything else. The amplified 455kHz signal from pin 10 is trapped by IFT2 and passed to the 1N4148 detector diode, D5. The conducting action of the diode clamps the positive half-cycle of the 455kHz carrier, leaving the negative half-cycle of the carrier wave and the audio-frequency (AF) modulation. That signal half-cycle passes through a low-pass RC filter (100kW/270pF) into a 1MW load. The filter removes intermediate frequency 455kHz signals but not the AF modulation nor the negative DC component. The negative DC level is fed via a 1MW isolating resistor to the AGC line that goes back to the input control grid of V1A at pin 2. This acts as a level control, reducing the set’s gain for stronger stations. The audio modulation is fed forward to the pentode grid input at pin 8 via the volume control, VR1, and IFT1’s primary. This time, V1B amplifies the AF signal (at the same time it is amplifying the IF signal!), and that appears at the plate output, pin 10. This AF signal is ignored by the top load IFT2 (acting like a small RF choke only) and develops across the output transformer’s primary. It matches the low impedance of the speaker (4W) to the high impedance of siliconchip.com.au Photo 1: the routing of the wiring under the chassis is critical since so many different signals meet at the valve base. The ‘gimmick capacitor’ formed by the green and black wires twisted together at lower middle provides a bit of extra feedback to the second stage (V1B) so it doesn’t burst into oscillation. the pentode plate (~10kW), and the AF signal is fed to the speaker. That is the basics of the circuit, where V1A amplifies frequencies that are pretty close together, while V1B handles signals that differ significantly in frequency. The gain of the first section is very low; certainly less than 10 times. The rest of the gain is in the second section, where near-heroic measures have to be implemented to keep the gain high and the stage stable. That is where the neutralisation comes in. Stability and neutralisation Overall stability with fair performance was first reached by a combination of shielding and bypassing. Then, when it became unstable with more gain, I implemented the magic neutralisation by deliberately bringing some output and input leads together to remove the instability. The latter technique was new to me and seemed like witchcraft until I studied relevant technical texts. They described what happens when a careful portion of the output energy is fed back to the input, with the promise that the stage gain could be raised siliconchip.com.au without instability. I did not believe it until I had the screaming unstable IF/ AF reflexed stage go quiet and docile simply by twisting two wires together to form a very small amount of capacitance from output to input! Editor’s note: “Neutralisation” refers to adding positive feedback around an amplifying device to cancel out its inherent negative feedback due to Miller capacitance, thus enhancing its bandwidth. While the added ‘gimmick’ capacitor in this case is similar to a neutralising capacitor, its purpose is slightly different. Here, due to reflexing, the Miller capacitance couples signals between the two IF transformers, one connected to pin 8 and one to pin 10. As they are both resonant at 455kHz, feedback can lead to unwanted oscillation. The gimmick capacitor reduces that coupling by partially cancelling the Miller capacitance, increasing stability. Normally, neutralisation would reduce stability due to the added positive feedback. In Fig.1, the neutralisation is shown diagrammatically by the wire connecting to pin 10 of valve V1B being capacitively coupled to the wire connecting Australia's electronics magazine to the volume control, VR1. They are the green and black wires that run up the middle of the chassis in Photo 1. In a typical set, the green wire would be kept short and well away from any valves, and thoroughly shielded to prevent unwanted coupling! A long run of a sensitive input wire inside the chassis over the valves can provoke the amplifying stage into regenerative instability, particularly in this case where both IF and AF signals are being handled. Scope 2 shows the signal going to the speaker without the wires twisted together, while Scope 3 shows the same waveform with them in close proximity, achieving stability. The result was an epiphany to me, having been brought up in the school of keeping output leads well away from input leads. The wild oscillations began to clear up as the wires were brought adjacent, with several twists being enough to remove all bad behaviour. If too many turns were made, the instability reappeared, there being a “Goldilocks” amount. Other stability components Some negative feedback is implemented for audio-frequency signals to July 2024  83 Photo 2: you can see the internal structure of the 6Y9 dual pentode in this photo. The right-hand quarter or so is the first pentode, V1A; the power pentode, V1B, occupies a much larger portion of the structure. Photo 3: the top side of the finished chassis. The HT is pretty low at 175V, generated from a 140V winding on the transformer, as the valve’s maximum anode voltage rating is 190V. roll off the supersonic response. A feed is taken from the speaker to the bottom end of the volume control potentiometer. The 5.6kW resistor in series with V1B’s grid and the capacitors bypassing the cathode resistor, all mounted directly on the valve socket, also improve stability by attenuating signals above the intermediate frequency. The circuit notes components that had to be mounted directly at the socket for maximum stability with asterisks. The process of achieving stable running was actually a long journey and hard fought. can be operated in an autodyne oscillator/mixer configuration, combining the tuning coil circuits. The power section can be employed in the reflex configuration, combining the IF amplifier and the AF amplifier/AF output. The key to its success is the colossal gain of the power section. Even though it is not being used in the intended application, which was for TV video amplification and CRT driving up to 5MHz, the gain and bandwidth are well-suited for use at 455kHz (IF) and 2MHz (upper end of the oscillator range). Without a separate triode to act as a local oscillator, the pentode V1A must be arranged as an autodyne converter. This is the weakest part of the set as the RF gain in this section, by virtue of the dual use, is relatively low. I was not successful in using an internal ferrite loop stick or loop antenna with this front end, so I settled on using conventional tuning and oscillator coils. Without the gain of a loop stick antenna, the set needs an external wire antenna to give good reception. The 6Y9 valve While the 6Y9 is a dual pentode, its two pentodes are quite different. The first section is a medium-gain signal amplifier, making it suitable for mixer/ oscillator duty. The second is a highgain power amplifier that can drive the speaker transformer. In Photo 2, the signal section at the right of the picture uses about 25% of the structure, while the power section is the remainder. The base of the valve has 10 pins that allow the electrodes of each valve to be accessed while keeping them separate. Because of this, the signal section 84 Silicon Chip Finishing the set I had to put a bit of ingenuity into obtaining or making the parts. I made Australia's electronics magazine the chassis from pieces of scrap sheet metal bent and pop-riveted together. I drilled holes where I thought parts should go, plus more, just in case. The front panel was part of a base plate from something with vent slots spaced just right to bolt the speaker onto and let the sound out. The larger parts you can see in Photo 3 are a motley crew of new, old and modified devices. The speaker transformer is a Jaycar MM1900, using the 0.5W tap. The power transformer is a Jaycar MM2011 rewound with 140V and 6.3V AC secondaries. The speaker is a Jaycar AS3008 4W unit. The tuning gang is a dual 500pF unit from my junk boxes. The tuning dial is a reduction type, also from the junk box. The IFTs were both from my junk box as well. IFT1 was from an Astor chassis and is marked 7872, while IFT2 was made by HMV and is marked 906 0062. I verified that both resonated at 455kHz before using them. I used these types as they came from valve radio chassis, so they should be happy with valve currents and voltages. The larger HMV unit for IFT2 has quite thick wire in it; I was mindful siliconchip.com.au Photo 4: the finished radio fitted into its case. IFT1 is on the right, while the beefy coil for IFT2 is in the middle. of the plate current of the 6Y9 possibly frying any miniature IFT. The important thing with IFT1 is that the primary winding impedance does not inhibit the oscillator frequency feedback. With some later experience making other autodyne sets, I feel that any valve-type IFT with ferrite adjusting cores and large resonating capacitors will work well. Under the chassis, the rest of the parts (except for the tuning coils) are what you have in stock or buy from Jaycar etc. I selected the components with reference to an article called “Radio Therapy” from Radio and Hobbies, November 1943 that gives a run down on autodyne radio sets and suitable parts. With its 140V AC HT winding, the power transformer output 175V from a bridge rectifier. I was mindful of the manufacturer’s maximum rating of 190V for the 6Y9, as well as advice from TV-era service techs that exceeding that voltage can cause valve failure. a shielding plate between them to remove weird whistles due to field interference. I rewound the coils several times during development, so like the rest of the set, they look a little messy with taps and bits of tape hanging out. Both primaries eventually tracked the necessary frequency ranges to suit the 500pF gang I used. I moved the secondaries several times to change the amount of coupling. The aerial coil resonates from 600kHz to 1800kHz, while the oscillator coil resonates from 955kHz to 2255kHz. The oscillator coil primary has a 430pF padder in series with the gang to make the ratio of frequency change nearer to 2:1, to suit the aerial coil ratio of 3:1. By fitting trimmers to the gangs and ferrite cores in the coils, I was able to tweak the tuning to get good tracking, and a near-constant 455kHz difference beat to feed the 455kHz IFTs. The tuning coils At this point, I had a chassis that worked as a usable radio. Still, to make the set truly practical, there had to be some sort of cabinet to house the chassis. I cobbled the tuning coils together from discarded plastic formers with the original windings removed. They are mounted side-by-side with siliconchip.com.au The cabinet Australia's electronics magazine I simply ran a tape measure around the chassis and, with scraps of Bunnings 5-ply, concocted a “kennel” cabinet for the set to live in. I nailed the bits of ply together and also Aqua glued them. Once set, I sprayed the wood with enough coats of waterbased white paint until it looked shiny. Conclusion The experience of making this single-­ valve autodyne-mixer practical receiver opened my eyes to the technology of the era. The process took several workshop months and resulted in many pages of tests and experiments, far too long to reproduce in this magazine. From the lessons learned from this project, I have made several other Autodyne radio sets with 1960s miniature and 1940s octal metal valves. The latter are the most well-­ developed as my understanding of the techniques improved. This process also answered the query: why did the autodyne die? For more details on this project, see my Vintage Radio forum posts at https://vintage-radio.com.au/default. asp?f=12&th=130 SC July 2024  85