Fullscreen HTML5Med Res (??MB)HTML4

Vintage Radio The Emerson 888 mini-mantel set (UK Version) By Ian Batty Emerson’s 888 radio was dubbed Vanguard in its US release, with a stylised rocket as part of the logo, overlaid by the word VANGUARD. From left-to-right, Regency’s TR-7, Zenith’s Royal 500, the Emerson 888, Toshiba’s 9TM-40 “robot” and Admiral’s 7M1. V anguard was the name of the US rocket that placed their second satellite into Earth orbit. It was intended to be the first, but when the Soviet Union successfully launched Sputnik I on the 4th of October 1957, they scrambled to respond. After the failure of the Vanguard TV-3 launch, they decided to quickly get the Explorer 1 satellite into orbit using a Juno I rocket. That was followed by Vanguard 1, making it the second successful US orbital launch of a satellite. The satellite launched on that rocket was retrospectively named Vanguard 1. Vanguard 1 continued to make useful contributions to space science until 1964. It, and its third launch stage, are the oldest artificial objects in orbit around the Earth, with an expected lifetime of some 185 years to run. The British release of this radio lacked the VANGUARD label, perhaps because “Vanguard” failed to resonate in the same way in the UK. A history of Emerson Victor Hugo Emerson (an early recording engineer and executive) started Emerson Radio Corporation in 1915 as Emerson Phonograph Co., based in New York City. Although Emerson introduced the first radio-phonograph combination sold in the USA, the company remained in obscurity until 1932, when, during the Great Depression, it introduced the “Peewee” radio. It sold like hotcakes, becoming ‘the’ radio to have. Emerson Radio & Phonograph converted to military 104 Silicon Chip production for World War II in 1942, when it held one-sixth of the US radio market. In 1947, among its first post-war products, Emerson offered a television set with a 10-inch (25cm) tube. Between fiscal years 1948 and 1950, the high demand for television allowed Emerson to more than double its sales. In 1953, Emerson Radio and Phonograph purchased Quiet Heet Corporation, which entered the company into the air conditioning market. Although radio represented only 15% of Emerson’s revenue by 1954, the company credited itself as creating the firsts of the clock radio, the solar-powered radio, and the hybrid pocket radio – the 838, reviewed in the October 2018 issue (siliconchip.au/Article/11276). They started producing tape recorders in 1955. Emerson Radio and Phonograph purchased the consumer products division of Allen B. DuMont Laboratories Inc in 1958. With this acquisition, a higher-priced line of television sets, phonographs and high-fidelity and stereo instruments, along with the DuMont trademark, were added to Emerson’s products. Unfortunately, by this time, almost every US household that wanted a TV set already had one, and many customers who were in need of another set were waiting for colour television instead of buying a replacement monochrome set. Emerson would be acquired by National Electric Corporation (NEC), ending some fifty years as an independent manufacturer. Emerson-branded products were finally discontinued in 1972 (see https://w.wiki/D6fJ for more details). Australia's electronics magazine siliconchip.com.au The Emerson 888 Regency’s TR-1 wasn’t a pocket set unless you had a large coat pocket. But it looked a bit lost on a shelf, so the ‘trannie’ would need to either become smaller, such as Sony’s TR-63, or larger. You could offer a full-sized mantel, as many manufacturers did, but mantels lose the cachet of portability. What about a ‘mini-mantel’ set? Released in 1958, Emerson’s 888 model is a convenient size, with a fold-back handle that allows it to sit safely at an angle. Similar sets include Regency’s TR-7, Zenith’s Royal 500, Toshiba’s 9TM-40 ‘robot’ and Admiral’s 7M1 (see the lead photo). In the hand, Emerson’s 888 is a simple brick with a thumb-wheel dial at the top. The volume control, fitted with a decorative key tab, demands that you hold the set in one hand and adjust the volume with the other – reminiscent of Regency’s TR-1, and less ergonomic than Sony’s TR-63. The dial is calibrated in metres rather than kilocycles (as would have been used back then). The tuning range is 550~200m (545~1500kHz) for medium-wave, with an original fixed long-wave frequency of 200kHz. There’s no separate band-change switch; long-wave is selected by tuning past the top end of the broadcast band to actuate an internal switch. Circuit description This radio follows the design that had stabilised by the mid-1960s. This UK release is the familiar six-transistor superhet, a scaled-down version of the eight-transistor US releases (Fig.1). The US releases featured an unusual direct-coupled two-stage second intermediate frequency (IF) amplifier and an audio preamplifier. Converter transistor TR1 is the familiar OC44. Both it and the similar OC45 use alloyed-junction construction, with the main difference being their cutoff frequency; over 7.5MHz for the OC44 or greater than 3MHz for the OC45. The circuit uses collector-emitter feedback, typical of European/US/Australian designs. While this gives similar performance to the collector-base feedback used in many Japanese designs, it has the advantage that you can inject a signal directly to the converter base without stopping the local oscillator (LO). Historically, collector-emitter feedback was used in the first transistor set, Regency’s TR-1. That ensured its grown-junction converter, with its limited high-frequency specification, would operate reliably over the broadcast band. This set’s LO tuning capacitor section has a cut-plate design. As this naturally forces the LO to track at 470kHz above the incoming signal frequency, no padder capacitor is needed on the broadcast band. It’s unusual to see cut-plate tuning capacitors in multiband sets, as the cut-plate construction can only give correct tracking over one band. But the 888’s long-wave band uses fixed tuning, so the cut plate’s LO offset has no effect on it. For the broadcast band, the ferrite antenna rod’s L1 primary is tuned over the range of 545~1500kHz by tuning siliconchip.com.au Fig.1: this cut-down set uses six transistors: TR1 (mixer/oscillator), TR2 (first IF amplifier), TR3 (second IF amplifier), TR4 (audio preamplifier) and TR5/TR6 (Class-B push-pull audio output). The demodulator is a single OA70 diode. There are also three IF transformers, one oscillator transformer and two audio transformers (phase splitter and speaker matching). In the UK, Cockburn & Gunn Ltd, operating from 1958, imported Emerson products from the USA. They became Emerson Electronics Ltd in 1962. capacitor VC1, with top-end trimming by TC2. Tuning the 888 to the very top of the broadcast band activates bandchange switch S1a/S1b. The antenna section, S1a, connects long-wave trimmer TC1 and 1100pF band-change capacitor C1 to antenna coil L1, thus pulling its resonant frequency down to 200kHz. C1’s high value of 1100pF ensures that broadcast trimmer TC2’s setting has virtually no effect on long-wave antenna tuning. Note that the C1 and C6 band-change capacitors are both ±2% tight-tolerance types. Broadcast LO tuning is by cut plate section VC2, trimmed by TC3. For long-wave, trimmer TC4 and 100pF bandchange capacitor C6 bring the LO frequency down to the required 670kHz. As C6 has a much smaller value than the antenna circuit’s C1, LO trimmer TC4 has a much wider adjustment range than antenna circuit trimmer TC1. In practice, in long-wave mode, it is designed to tune only to 200kHz, or close to that frequency. In common with other transistor converters, whether autodyne or separately excited, TR1 appears to work with almost zero bias. This implies that it’s working close to Class-B, as we’d expect with a self-oscillating converter stage. TR1 feeds the tuned, tapped primary of T2. This first IF transformer is permeability tuned by an adjustable ferrite slug. T2’s secondary feeds the base of the first IF amplifier transistor, TR2, an OC45. As this has an automatic gain control (AGC) voltage applied, its base resistor (R4) has a high value of 68kW. This allows the AGC control voltage to significantly reduce TR2’s bias on strong signals, thus reducing the stage gain and helping keep the audio output constant with stronger or weaker stations. The ‘cold’ side of T2’s secondary is bypassed to ground by an 8μF electrolytic capacitor, C7. This is not regarded as good practice, as electrolytics do not perform well above audio frequencies. That said, it worked just fine, even without a better-performing capacitor in parallel. TR2, like all alloyed-junction types, has considerable collector-base feedback capacitance. It uses R6 and C10 to cancel the feedback capacitance. As this circuit uses resistance and capacitance, it’s unilateralisation rather than simple neutralisation. TR2 feeds the tuned, tapped primary of second IF transformer T3. T3’s untuned, untapped secondary feeds the base of the second IF amplifier, TR3. TR3 works with fixed bias, having its own bias divider (R8/R9), and working at fixed gain. It’s also unilateralised, by R10/C14. Both networks (R6/C10 and R10/C14) use tight-tolerance type capacitors (±2%) and resistors (±5%). TR3 feeds the tuned, tapped primary of third IF transformer T4. T4’s secondary feeds demodulator diode D1. This, in turn, feeds 5kW volume control potentiometer VR1 as its load, with 10nF capacitor C15 filtering out all but the audio signal. The DC voltage developed across VR1 is fed, as the AGC voltage, back to the bias circuit of the first IF amplifier transistor, TR2, via 8.2kW resistor R5. TR2’s biasing from 68kW resistor R4 puts D1 weakly into forward conduction, improving the radio’s sensitivity. The audio developed across VR1 goes to the base circuit of audio driver TR4, an OC71, via 8μF capacitor C16. Using combination bias, TR4 feeds the primary of phase-splitter transformer T5. The output pair of transistors, TR5/TR6 (both OC72s), operate in Class-B mode. Their bias is derived from divider R17/R18. This circuit lacks temperature compensation, and this appears to be more common in English-designed sets. Australian designs, starting with our first transistor set (AWA’s 879P), incorporated thermistor compensation from the beginning. I’ve seen European equipment – which probably worked just fine in Europe – either go out of alignment, or just die, when exposed to our wider range of environmental temperatures. Top-cut is applied by 40nF capacitor C20. Local feedback is provided by 10W common emitter resistor R20, and there A top view of the Emerson 888 radio’s PCB with some of the important components labelled. You can see the battery holder attached to the volume control at the bottom. Converter Oscillator Coil 1st IF Transformer 1st IF 2nd IF Transformer Driver Transformer Outputs Output Transformer 2nd IF 3rd IF Transformer 1st Audio Demodulator Diode Volume Control 106 Silicon Chip Australia's electronics magazine siliconchip.com.au is overall feedback from T6’s secondary, via R19 (1.5kW), to T4’s unbypassed emitter resistor, R16 (10W). TR5/TR6 drive output transformer T6, and its secondary drives the internal speaker, or an earphone plugged in to the earphone socket. The set runs from a 6V supply made up of four AA-sized cells in a carrier. Restoration The review set was in good cosmetic condition, so a light clean had it looking just fine. Turning it on produced nothing. Usually, this points to a dead set, but I was able to inject a few millivolts of audio into the volume control and get an output. Further testing showed the RF/IF section was as dead as the dodo. Injecting a 470kHz signal into the demodulator produced nothing, and the cause was an open-circuit demodulator diode, D1. This was a reminder that, really, you need to be alert to any possible fault, no matter how unlikely. D1, the famous OA70 we probably used in crystal sets, is in a low-stress part of the circuit, never getting more than a few hundred millivolts compared to its maximum reverse rating of 22.5V. But there it was – as open a circuit as just leaving the multimeter leads lying on the test bench. Replacing D1 (with a near-equivalent OA81) brought the set to life, and it was just a matter of checking voltages, aligning it and putting it through its paces. Be aware that, in common with many British designs, this uses a 470kHz IF, with their other common frequency being 465kHz. If you’re unsure, get the manufacturer’s data or service sheets. Performance results It’s on a par with other six-­transistor sets of the day. I was puzzled at first, as it didn’t emit the usual front-end noise when turned up to full volume, but its specifications appear to be about right. In detail, for 50mW output, it needed just on 1000μV/m at 198kHz, 275μV/m at 600kHz and 225μV/m at 1400kHz, with signal+noise to noise (S+N/N) figures exceeding 20dB. The relatively poor long-wave sensitivity may have been due to my radiating test ferrite rod, as it was only ever specified for the 535~1605kHz broadcast band. My on-air weak station reference, Warrnambool’s 594kHz 3WV, rocked in at full volume. Regrettably, there are no local long-wave transmissions in the Geneva Frequency Plan of 1975, specifying band coverage of 153kHz to 279kHz. Non-directional beacons (NDBs), used in air navigation, are located at higher frequencies, just at the lower end of the 300kHz~3MHz medium-wave band. The closest NDB to me here on the Mornington Peninsula is the Moorabbin NDB at 398kHz. Using the European/US converter design of emitter feedback allowed me to inject a test signal at the converter base, and the levels there are consistent with the pickup effectiveness of the 888’s short ferrite rod. Its IF bandwidth is 1.25kHz for -3dB and 22.5kHz for -60dB. The AGC allows about a 6dB rise for a 28dB signal increase. That’s about as good as you’ll get with the single-stage AGC siliconchip.com.au The band-change switch is circled in yellow; the LW trimmers are also visible in this photo. The underside of the Emerson 888 PCB. Note that the volume control pot is secured with three screws. Australia's electronics magazine May 2025  107 The Emerson 888 has a distinctive volume control knob, resembling a door knob. The tuning dial is made from plastic with the “LW” setting just past the “200” mark. used in the 888. The audio response from antenna to speaker was 180Hz to 2000Hz for -3dB. From volume control to speaker, it’s around 180Hz to 7.8kHz. At 50mW, total harmonic distortion (THD) was 4.2%, with clipping at 70mW, giving a THD of 10%. That seems like a low maximum output power, but the clipping was symmetrical, which it would not have been with one faulty output transistor. At 10mW output, the THD was 4.6%. Low-battery performance was good: with a 4.7V supply, it managed a useful 35mW at clipping, albeit with visible crossover distortion due to the voltage-divider bias circuit. Is it worth buying? I think it’s worth having as an example of a major American manufacturer customising their design to suit an export market. It’s unusual in having the fixed-tuned long-wave provision. Any long-wave provision – even a fixed-tuned design – appears an oddity, given that long-wave was in decline when the 888 was released. The BBC, however, maintains its 500kW 198kHz Droitwich service, as its transmissions cover most of England and Wales, plus much of the Republic of Ireland. Its rubidium frequency synthesiser-controlled broadcasts are readily 108 Silicon Chip available as a frequency standard reference (see https://w. wiki/D8fu for more details). Special handling The tuning dial is secured by a central screw with a knurled head that is easily removed. The volume control knob is a press-fit onto a chamfered shaft – be careful when withdrawing the knob, as it is plastic and is easily damaged by injudicious levering-off. The board is secured to the case by one large nut and two small ones. Emerson states that you must include fibre insulating washers between the nuts and the circuit board. At least one nut would otherwise short out a circuit board track. Be aware that the medium-wave band is specified for a maximum frequency of 200 metres (1500kHz). I did try tuning up to the standard 1605kHz. While the LO would tune correctly, the antenna trimmer, even when wide open, would not bring the antenna circuit into tune. It did work perfectly well for a maximum of 1500kHz. The long-wave tuning is intended for 200kHz (198kHz for the major remaining UK station). While the LO will tune more broadly, the 1100pF antenna circuit padder (TC1) severely limits the authority of the long-wave antenna trimmer, TC3. For more info on this set, see siliconchip.au/link/ac4q SC Australia's electronics magazine siliconchip.com.au