Fullscreen HTML5Med Res (??MB)HTML4

Vintage Radio Stromberg-Carlson “Air Hostess” Model 4A19 Stromberg-Carlson was known as a high-end radio producer, but this was one of their more inexpensive models. It had a pretty bare circuit, with just four valves, one IF transformer and a very basic volume control. The volume control, in particular, was its Achilles’ heel. A fter the Second World War, Stromberg-Carlson sold large numbers of their full-specification model 5A27, a medium-size mantel radio. The 5A27 cost £21/10, while the 4A19 radio featured here was priced at 18/7/6 (£sd). Even though Stromberg-Carlson heralded the Air Hostess as Australia’s newest and finest radio, they would have known that the publicity department had overstepped the mark considerably. The radio is cheaply made. It is adequate for the kitchen, and then only if it is left tuned to one station. To be fair, the five-inch (127mm) Rolla speaker baffled by the case provides good quality listening. You will not encounter the likes of the promotional text for this radio today, which you can read in the accompanying advertisement from The Australian Woman’s Weekly, June 26, 1948 (see Fig.2). The 4A19 is shown in the large photo on the right, while the more expensive 5A27 is below it and to the left. 102 Silicon Chip By Associate Professor Graham Parslow It may have proved too edgy even for that time because they changed the name to Air Queen in 1952. Perhaps the coronation of Queen Elizabeth II was an additional factor. Although the case was offered over several years, the circuitry inside varied considerably, with the single common factor of having four valves. If you encounter one of these radios, it may be quite different internally from the one featured here. Whatever components were in stock seem to have been adapted to make this series. In this radio, the only new-at-the-time miniature valve is the 6AV6 IF amplifier and detector. Circuit details The circuit diagram shown here (Fig.1) has been modified from the original to reflect what was inside this radio. The original circuit had a valve lineup of 6A8, 6AR7, KT61 and 6X5, reflecting the use of old valve stock. Unlike the circuit shown in this article, the grid bias to the KT61 output Australia's electronics magazine valve was set by a resistor between Earth and the cathode. In this radio, the 6V6 cathode is tied to Earth, and bias is created by a 240W resistor from the mains transformer centre tap to Earth. I measured the 6V6 output beam tetrode grid bias at -5.9V. That is relatively low, but the HT voltages across the π filter built around inductor L4 were also rather low at 140V at the rectifier end and 122V at the output end. Even so, the volume was more than adequate, and the set consumed a modest 24W, including a dial lamp not shown in the circuit diagram. Using a choke in the π filter is the only extravagance in the component complement. I expected the filtering to be marginal due to using electrolytic capacitors of just 8μF each. However, hum was acceptably low, so there was no need to add extra capacitance to the π filter. This radio was released before the common use of ferrite rods for antennas, so it has a conventional aerial coil, siliconchip.com.au Fig.1: the circuit diagram for the Model 4A19. As it was during the post WW2 period, the radio was manufactured with whatever components they could find, which was likely one of the reasons why the design is so simple. In the original circuit, the capacitor below the primary of L2 is listed as “300” (μF), the correct value should be 300pF. L1, with a tuned secondary for spanning the medium-wave (MW) broadcast band. A 5pF capacitor between the aerial coil primary and secondary boosts signal strength when tuning higher frequencies. The EK2 octode mixer valve was introduced by Philips in 1938 with a proprietary flat-pin C18 base (also known as a P-base). The EK2 and EK32 are electrically identical; most manufacturers preferred the conventional octal base on the EK32. The EK32 mixer was produced with and without a metallised coating that could be grounded to act as a shield; when present, the shield was usually painted red. In this radio, the EK32 made by Philips has no shield. The rubber insulated wire to the EK32 top cap (grid connection from the tuning capacitor) was perished, as were other wires that required replacing. The Armstrong-configuration local oscillator using transformer L2 is a conventional way of introducing the heterodyne frequency to generate siliconchip.com.au the 455kHz intermediate frequency (IF). The big surprise is finding only one IF transformer in the set. Before I acquired the circuit diagram, I was perplexed as to where to find the second IF coil. Other models in the same case could have had two IF transformers. Even with only one IF stage, the station selectivity is surprisingly good. What is not so good is the volume control, as it is actually an RF gain control. A 5kW potentiometer sets the grid bias voltage on the EK32 mixer. Gain is at maximum when the wiper of the potentiometer is connected to Earth. When the potentiometer creates resistance from the EK32 cathode to Earth, that raises the cathode voltage, reducing the effective grid bias. That’s because the EK32 grid is effectively at Earth potential by connection through the aerial coil. The result is that the volume potentiometer sets the negative bias to the control grid. Australia's electronics magazine Photo 1: the Philips EK32 mixer valve can be seen in the foreground of this photo. Often this valve has a metallised coating which acts as a shield, but there was none present. September 2024  103 This is not a particularly good way to achieve volume control. At any particular setting, while tuning through the broadcast band, strong stations blast in at a high distortion level, while weak stations are not audible. If the radio remains tuned to one station, that problem goes away. There is no reflexing of the audio through the 6AV6 because that would require a second IF transformer. The detected output from pin 5 of the 6AV6 is coupled to its grid for audio pre-­ amplification. There is no tone control. However, the 20nF capacitor from the 6V6 anode Fig.2: an advertisement for the Stromberg-Carlson Model 4A19 from Women’s Weekly Saturday June 26th, 1948. Source: https://trove.nla.gov.au/newspaper/ article/47221078 speaker). Initially, I did not notice that the output valve socket was empty. However, as you can deduce, my curiosity was piqued as to how many things can be wrong with such a radio. Once I started, there was no going back. Photo 2, showing the underside of the chassis, was taken during my preliminary assessment. I had not yet replaced the two-core mains lead in case the radio was unsalvageable. Several capacitors had been replaced, indicating that someone had restored it previously. As there was no output valve, I connected a signal tracer to the grid pin of the 6V6 socket. Nothing tuned in. The 20nF coupling capacitor between the 6AV6 and 6V6 had been previously replaced with a polypropylene type that is usually highly reliable. With little reason to expect a different result, I connected the signal tracer to the 6AV6 anode. I was rewarded with a good signal from stations that tuned across the spectrum by manually turning the tuning capacitor. That dud 10nF coupling capacitor hit the bin rather quickly, and my enthusiasm to continue was unabated. The enthusiasm even survived testing the speaker transformer primary by measuring its resistance between pins 3 and 4 of the 6V6 socket – it was open circuit. The speaker transformer was riveted to a bracket on the Rola 5C speaker. Drilling through the rivets and removing the transformer allowed me to confirm the open-circuit primary. The next step was to clean the chassis by brushing residue off using mineral turpentine and then blowing it with compressed air. A spray of green paint on the mains transformer made a significant visual improvement. Next, it was time to restore the stringing to the tuning knob. That initially appeared to be impossibly difficult, but removing the dial backing sheet revealed that it was really rather simple, as shown in Photo 4. The small dial drum has a broad rim with a single hole that exposes the grub screw binding the drum to the shaft of the tuning capacitor. The dial cord is a single piece with a hook at one end and a spring at the other. These can be anchored to the hole in the drum, and three loops of cord can pass over the tuning knob shaft. Australia's electronics magazine siliconchip.com.au 104 Silicon Chip to Earth acts as a top-cut audio filter, as well as filtering out any remaining IF signal. There are circuits from other manufacturers with as few or fewer components, such as the Astor DLP described by Ian Batty in the October 2016 issue (siliconchip.au/Article/10333). Restoration I nearly put the radio aside after a first inspection. It had perished rubber wiring, a two-core mains flex to replace, was very dirty, the tuning system was broken, and the radio did not work (complete silence from the Photo 2 (above): this photo of the underside of the chassis was taken before restoration. Some capacitors had been replaced by the previous owner(s). Photo 3 (below): similarly, this photo of the top of the chassis was taken during the early stages of restoration, A few of the components had been removed to make room for replacements. siliconchip.com.au Australia's electronics magazine September 2024  105 Photos 4 & 5: the front of the chassis with the dial cursor backing sheet removed (above) and the new one in place (below). When the dial backing was reinstalled, the cursor was easily rotated to span the tuning range. An odd thing you may have noticed in the lead photo is the presence of wobbly dial calibration lines on this radio. That is not how it left the factory, as shown by many other photos of the model. It is also not unique because I have seen this on other glass dials. I cannot be certain how station identifiers wander and lines distort, but a combination of heat and moisture are likely contributors. The case was in good condition and brought to a sparkling sheen with Meguiar’s Ultimate Liquid Wax featuring pure synthetic polymers. It has a hefty price tag repaid by the outstanding virtue of leaving no white residue. Used on cars, it produces a finish that is good for a year between polishes. With a new cloth-covered mains cord and a replacement speaker transformer, the project was complete. Some background Photo 6: the completed chassis, with the mains transformer painted green. You can also see the replacement speaker transformer. 106 Silicon Chip Australia's electronics magazine Stromberg-Carlson Australia was an autonomous operation and ran its business largely independently of its American parent. The Australian company began by importing receivers from the USA in 1927, and a year later, started the local manufacture of receivers and most of their components. In 1936, their production volume justified the construction of a large factory at Bourke Road, Alexandria, NSW. Stromberg-Carlson made receivers and components for themselves, as well as for brands including Audiola and Crosley. The 1930s were boom years for Stromberg-Carlson radios. In the war years, between 1939 and 1945, Stromberg-­ C arlson produced telephones and telephone switchboards for the Australian Army. Adverts from the 1940s proclaimed, “... there is nothing finer than a Stromberg Carlson”. Throughout their history, they primarily aimed for the high end of the market, with exquisite woodwork on many products. The radios continued to sell well in moulded plastic cases through the 1950s. The end of the radios was a line of distinctive portable transistor radios clad in patterned leather. Stromberg-Carlson tried to participate in the Australian television market, but they were not competitive and ceased all manufacturing in 1961. SC siliconchip.com.au