Fullscreen HTML5Med Res (??MB)HTML4

Vintage Radio Loewe’s Loewe’s 1927 1927 OE333: OE333: simplicity simplicity itself itself By Ian Batty Talk about a minimalist radio set! Did they go a bit too far in stripping it back to essentials? Known as the Ortsempfänger (local receiver) OE333, this set must be the ultimate in electrical simplicity for anything short of a crystal set. At the time this set was released in the late twenties, radios were taxed based on the number of valves. So there was an incentive to keep the valve count down, as long as it didn’t hurt performance too much. So they thought: why not combine several electrode assemblies into one “valve” for a compact, low-cost radio? That’s just what young Baron Manfred von Ardenne did. Obtaining a patent at the age of 15 and dropping out of high school, he sold the patent to Dr Sigmund Loewe and took up work with him. Loewe and his brother David had established Radiofrequenz GmbH in 1923 in Berlin, and Loewe’s company began releasing 3NF-based sets in 1927. Valves of the day were expensive; the 3NF more so. But Loewe offered a repair service that would replace blown filaments and restore “as new” performance. In practice, both filaments were always replaced. While the 3NF is probably the best known of Loewe’s “multi-valves”, they also released the 2HF dual tetrode and the metal-shielded WG36, containing an RF pentode, triode local oscillator and pentode IF amplifier. It would have been possible to put just the valve assemblies into the envelope, but von Ardenne and Loewe decided to put all minor components in as well: two anode load resistors, two grid resistors and two audio coupling capacitors. While this was a practical construction, the use of only six connecting 100 Silicon Chip Front view of the OE333. “Lautsprecher” is German for “loudspeaker”. Top-to-bottom, left-to-right: the Kleinempfänger DKE38 (July 2017), Loewe OE333 and Grebe Synchrophase (July 2016 & February 2018). Australia’s electronics magazine siliconchip.com.au pins did mean that all signal coupling would have to be at audio frequencies, as no tuned circuits could be placed anywhere between the input connection and the output. A close look reveals the six minor components individually encased in glass sleeves, and presumably evacuated via sealed-off pips. This would be necessary to prevent any escaping gaseous material from these components compromising the near-perfect vacuum that the valve sections needed to continue operating. How it works Two of the triode valves (stages 1 and 2) are designed for high-gain operation, while the third is designed for driving the loudspeaker. The filaments for stages 1 and 2 are wired in series across the 4V filament supply; stage 3 has the full 4V applied. All three use “dull emitter” thorium-coated wire. The valve characteristics are plotted in Fig.1 (for stages 1 & 2) and Fig.2 (stage 3). Notice that the anode current (Anodenstrom) is shown in micro- amps (10-6A) for stages 1 & 2, so this is a very low-current valve. But the amount of gain you can get from a triode is based on its amplification factor (MU), modified by the anode load value, and this is principally the valve’s anode resistance and the load resistor, both in parallel with the following stage’s grid resistor. Despite its anode current of only tens of microamps, the amplification factor exceeds 50. There is a somewhat similar RETMA valve, the octal IH5G/GT. The 3NF was mainly used in radios; however, one ingenious company used it in a clockwork-motor radio station identification machine! The radio circuit is simplicity itself – see Fig.3. Stage 1 works as a grid-leak demodulator. It is biased via the G1 pin, set at 1.5V by a battery tapping. Varying this, I was able to reduce the set’s gain, but no circuit shows such a feature, so volume control was achieved by swinging the antenna circuit coils apart or together, to vary coupling. Fig.1: plot of the 3NF valve characteristics for stages one and two. siliconchip.com.au The antenna circuit, as with all welldesigned receivers, contributes to performance: at 600kHz, there was a voltage step-up of almost 13 times between primary and secondary. The tuning capacitor is straightline capacitance, so stations crowd together at the top of the band. Like the DKE38 Kleinempfänger I reviewed in the July 2017 issue (siliconchip. com.au/Article/10728), it uses a solid plastic dielectric, giving a compact design. There are no bypass capacitors anywhere on the set (relying on the low impedance of batteries instead). I did experience feedback with untidilyplaced test leads at one point. Stage 1 capacitively-couples demodulated audio to stage 2. Stage 2 is internally biased by its grid resistor returning to ground. As the filament is at the top of the series string, this puts some -2V of bias on the valve. Stage 2 capacitively couples to output stage 3, which picks up -7.5V external bias via its grid resistor from pin G3. Fig.2: 3NF valve characteristics for stage three. Australia’s electronics magazine July 2020  101 I used an Amplion cone speaker for testing, but any high-impedance horn or cone speaker will work. You could also use a conventional movingcoil speaker with a transformer with around 4kW primary impedance. Clean-up Fig.3: circuit diagram for the Loewe OE333. As reinforced by the photo of the underside below, this is an extremely simple radio. The resistance-capacitance coupling between stages sets a low-frequency limit under 50Hz. I am indebted to a comprehensive analysis of this set on the Radiomuseum website (see references), but am unable to thank the anonymous author. Construction The close-up photo of the 3NF valve (Fig.4) shows a central coppercoloured capacitor, flanked by two smaller resistors to left and right. The metal cylinder starting at the centre of the picture and extending upwards 102 Silicon Chip past the getter silvering is the output valve (stage 3), while the horizontal cylinder near the top and partly obscured by the getter is one of the stage 1/2 valves. Connecting leads can be seen entering the mounting press at the bottom. The timber casing, and exposed antenna coils, make this little set slightly susceptible to hand capacitance. Be aware that the primary coil is the smaller of the two – this set initially tuned no lower than 850kHz, but tuned correctly once I swapped the two coils over. Australia’s electronics magazine This set was in excellent cosmetic condition, having been bought at auction from the Historical Radio Society at RadioFest 2019 in Canberra. Electrically, it is some 90 years old, and I was a little apprehensive about the 3NF’s condition. Applying power (very carefully) didn’t seem to get much response. Cleaning up the contacts helped a bit; however, performance still seemed lacking. But the set came good after maybe half an hour of operation. This can happen with old valves, especially with thoriated filaments. The thorium coating is only a few atoms thick and can degrade over time. Typically, it recovers in operation. While the recovery can be sped up by applying over-voltage to the filament, thankfully I did not need to do this. Frankly, I would probably not have risked such a rare and valuable device as replacements run many hundreds of dollars online. Once the set was working, I had little else to do other than check its performance. You’ll note that I haven’t shown many measurements in the circuit diagram (Fig.3), as there are few points that I can probe due to it mostly being a sealed set. For testing, I wanted to discover its best performance, so I added a variable capacitor between the signal generator and the antenna terminal. This allowed me to achieve optimal matching at any frequency, and let the antenna tuning work to its optimum. Removing this capacitor and inserting a standard broadcast-band dummy antenna reduced the gain by about 2~3 times, so this set does demand a properly-designed antenna for best performance. How good is it? The OE333 showed significant harmonic distortion at levels above 5mW output, so sensitivity testing was done at 1mW output. That may not sound like much, but my Amplion speaker gave a comfortable volume level in the workshop during testing. siliconchip.com.au The tuning range ran from 546kHz to 2350kHz, evidence of a large capacitance ratio (about 16:1) in tuning capacitor C1. Unlike a superhet, the antenna circuit does not need to track with any other circuit (such as a local oscillator), so it made no sense to add the expense of a trimmer that would have only reduced the maximum tuneable frequency. Sensitivity varied with tuning, the best being 5mV at 1150kHz and the worst being 10mV at 95kHz: see the table in the circuit diagram (Fig.3) for more details. Selectivity also varied: ±6kHz at 600kHz, ±14kHz at 900kHz, ±18kHz at 1150kHz and ±25kHz at 1650kHz, This variation in bandwidth is not unexpected for a single tuned circuit, but does permit more than one station at a time to be heard towards the upper end of the band. Audio performance was only fair. For a 400Hz signal at 1mW output, total harmonic distortion (THD) was 6%; at 5mW, it rose to 10%, and clipping occurred at 10mW output, with 20% THD. As noted above, broad selectivity allowed a few stations near the top end of the band to overlap, confirming the limits of any radio which only has antenna tuning. Unlike the DKE38, the OE333 could not take advantage of regeneration to improve its selectivity. There were regenerative, mainspowered, dual-band versions of this set: Loewe’s EB100W and R645W, among others. These took up an unused connection to the V1 anode, brought out through the envelope and tucked up inside the hollow of the press (stem) that supports the internal elements. It would be fascinating to compare these for both sensitivity and selectivity, given the DKE38’s impressive performance. The OE333 is also nowhere near as good as the Grebe Synchrophase (July 2016: siliconchip.com. au/Article/10016 and February 2018: siliconchip.com.au/Article/10977), but then, almost no radios of the era can match it in performance. Remember that the five-valve Grebe used stateof-the-art neutralised RF amplifiers and audio coupling transformers, that together resulted in the voltage gain of a six-valve set. The Synchrophase needs just 35µV siliconchip.com.au of signal for a 1mW output. The Loewe, with no RF amplification, needed almost 230 times the same RF level to achieve the same 1mW output level. But add in the Grebe’s extra two valves (type 01A, with a maximum amplification factor of eight each), and two audio transformers (gain of three each), and my back-of-the-envelope calculations put a hotted-up Loewe on par with the Grebe. Consider that I only had to throw about 10m of wire out of the workshop door and onto the carport to bring in 774 ABC Melbourne (about 60km away) at an acceptable listening level down here on the Peninsula. Radio National and five other metropolitan stations also came in at usable levels. A proper long-wire antenna would bring up all local stations well. No-one with any sense of history (or of preserving value) is going to “improve” a classic set like this one, so let’s take it for what it is. It’s a milestone in radio history. Not only does it perform creditably for such a simple design, but its compact form with exposed components would also have made it a ‘pride of place’ addition to the modern household of 1927. Forget those boring sets with their large, imposing, closed cabinets and dial after dial after dial to twiddle and misadjust. This set is an example of the ‘steam engine effect’. Major parts of the mechanism are exposed to view. Even relative novices could point to the antenna coils, to the three-in-one valve, and not only recognise them, but maybe even say a few words to onlookers to show that they were au fait with the wonder technology of the age. Put this marvel of 1920s engineering next to any old timber-cased radio of the day, and I’m pretty sure I know which one would attract the most interest. Fig.4: interior of the 3NF valve, showing the stage 1 & 2 triodes at top (horizontal cylinders) and vertically orientated output triode in between. Would I buy one? I might. Expect to pay at least $800 locally, more online/overseas. I’m thinking of saving up a bit of money and seeing what turns up. It would be so cool. Loewe OE333 versions There are many similar radios from several manufacturers, including mains-powered regenerative versions, and versions adding the dual-tetrode 2HF in the RF circuit. Search for 3NF Australia’s electronics magazine Source: lampes-et-tubes.info/rt/rt175.php July 2020  103 at Radiomuseum (see link below) for more information. Side view of the OE333 showing the power cable and both antennas. Special handling The rear antenna coil could be rotated. This acted as a volume control by varying the coupling between the two coils. The battery leads on this one were flexible enough for my testing, but you should examine them and treat them with care. They can become brittle with age. Be alert for the visually-identical 3NFB. Although it also uses a 4V filament supply, all three filaments are wired in series. Otherwise, it appears to work identically to the 3NF. I was super-cautious about the filament voltage, but so long as you only apply the recommended 4V, you should have no dramas. But be aware that thoriated filaments can take a little while to regenerate. There’s more information on thoriated cathodes on page 93 of the February 2018 issue of Silicon Chip in my article on the Grebe Synchrophase (siliconchip.com.au/Article/10977). If you acquire one of these but get no useful output after maybe an hour of operation, or the HT current is a lot less than 3mA, first check that you have the biasing correct. If all supply voltages are OK, you may consider revitalising the filament. I recommend that you be really sure of the need to do this, and that you research the process thoroughly before proceeding. Conclusion Special thanks to Jim Easson of the Historical Radio Society of Australia (HRSA) for the loan of this rare and remarkable radio. Thanks also to Giorgio Basile of http://lampes-et-tubes.info for his superb website and the supply of the close-up photo of the 3NF. Not an HRSA member? Hop on to http:// hrsa1.com and have a look around. Further reading • Tyne, Gerald E. J., Saga of the Vacuum Tube, 1977, Howard W. Sams, Indianapolis (pp446-450). • Ernst Erb’s Radiomuseum (http:// radiomusuem.org) has heaps of circuits, photos and articles on the OE333 and other implementations of the 3NF and its cousins. There is also a very thorough two-part analysis (in German) of the 3NF: siliconchip.com.au/link/ab23 siliconchip.com.au/link/ab24 • A stunning photo essay: http:// lampes-et-tubes.info/rt/rt175.php SC 104 Silicon Chip Australia’s electronics magazine siliconchip.com.au