Fullscreen HTML5Med Res (??MB)HTML4

Vintage Radio By Dennis Jackson The 1924 RCA AR-812 superhet radio receiver This was the world's first commercially available superheterodyne radio and a “portable” set to boot – the RCA AG-814 external aerial loop antenna and model 100 loudspeaker puts its total weight close to 30kg! This set uses just six UV199 triodes, with most components tucked away in the “catacomb”, a metal container sealed with a wax-like substance (in this case rosin). The intention was to prevent competitors learning about how the radio was designed. 96 Silicon Chip Australia’s electronics magazine siliconchip.com.au W e are fortunate to be living in a time of rapid technological progress. Yet few of us are aware of the great minds whose work long ago underpins many important aspects of that modern technology. Much of their hard work is now taken for granted, as if the facts and techniques that they worked so hard to acquire have always been obvious. By around 1900, Guglielmo Marconi had put together the bits and pieces gained by the discoveries of the great researchers before him to become the “Father of Radio Spark Telephony”. But Major Edwin Armstrong, formerly of the US Signal Corps, could be referred to as the father of modern radio, having played the major part in the development of the superheterodyne radio receiver and being the inventor of the regenerative detector. He also developed the Armstrong oscillator which helped to make audio modulation possible. He came up with and eventually put into practice the concept of frequency modulation which has now become the norm for both radio and television transmissions. A brief history of the superhet History records that Armstrong was not the first to come up with the idea of the superhet radio. Canadian engineer R. A. Fessenden had made observations concerning beat notes in the transmission of radio signals using Morse code around 1900. Radio technology developed slowly until the first world war of 1914-18 resulted in an urgent need for radio receivers superior to the tuned radio frequency (TRF) sets then in use for communications. There was also a need to develop direction-finding equipment to detect enemy ships at sea. Lucien Levy of the French signal corps obtained a patent for a superheterodyne receiver in 1917. Americans entered the conflict in Europe in April 1917 and sent over an expeditionary force. Major Armstrong was attached to this and he quickly became aware of the poor performance of the sets of the time. He set about investigating the lack of sensitivity and selectivity. Armstrong believed the problems could be overcome by mixing the Recreation of the wiring diagram for ► the AR-812 radio. siliconchip.com.au Australia’s electronics magazine August 2019  97 incoming signal with a locally produced signal to produce a beat note of a fixed frequency, which could be further filtered by fixed tuned circuits, avoiding the necessity to use variable tuning for each stage of RF amplification as in the TRF receiver. Armstrong then built the first practical Superhet radio, an eight-valve set which performed better than any others of the day. This was around the time when the armistice was signed, and so the need for radio sets became less urgent. He applied for a patent covering the Superheterodyne radio on 30 December 1918 and was undoubtedly the first to take out a patent on the Superhet in the USA. The first consumer Superhet set, the RCA AR-812, came on to the US market in March 1924. These were sold by the Radio Corporation of America, better known as RCA. They were built by the Victor Talking Machine Company. RCA did not manufacture wireless sets until the beginning of the 1930s. Edwin Armstrong and Harry Houck, who is usually credited with the development of the second harmonic mixer, were the primary engineers for the AR-812. These sets sold for US $269.00 without batteries, speaker or antenna; a considerable sum of money by today’s standards. A Ford Model T motor car could have been purchased for a similar sum at that time. It is interesting to note some sources claim about 80,000 units were sold. It appears that few of these sets made it to Australia. Getting hold of an AR-812 Despite this, I noticed one of these sets for sale on eBay about four years ago. It was being offered by an antique shop in Queensland. All of the UV199 valves were missing and I noted that the chassis layout was very unusual, which I found both puzzling and interesting. I was aware of the RCA AR812 at the time but knew little of its history or development. Since it was the first true Superhet set, I became convinced that this set would be a very worthwhile addition to my collection of mainly 1920s radios. The AR-812 duly arrived at my home in Hobart and I must confess to sneaking the box into the workshop through the back gate to avoid any 98 Silicon Chip awkward questions about what I had purchased, how much I spent etc. No time was lost getting it all laid out upon the workbench. The long, narrow table-top cabinet with its carry handle was in good nick. A dose of paint stripper would remove the several layers of very dark varnish applied sometime in the distant past and a careful sanding followed by a couple of coats of spray lacquer would restore its appearance to its previous glory. A large central front panel hinged down on two locating pins to reveal the works. Central to the interior of the hinged panel and screwed to it was the mysterious metal box known as the “labyrinth”. Two variable tuning condensers were fitted, one on each side of the box. The one on the left was used to manually tune the inbuilt frame aerial and the other on the right to separately tune the local oscillator. Two pairs of oscillator coils were mounted immediately under the oscillator condenser. Beginning the restoration As mentioned earlier, all UV199 valves had been previously removed. I had made a bad mistake by being a bit overeager and had ordered, after a good deal of searching on the internet, a set of UX199 valves of the period via eBay from the USA. Upon finally receiving these, I discovered I had ordered the wrong type. I really needed the earlier type, the UV199. They are the same valve but the UX199 has a narrower base with shorter pins and a different pinout. After a further search through the internet, my luck changed and I was able to eventually purchase the six required (and rare) UV199 valves in two lots, four being boxed new old stock. The mysterious catacomb box proved to be not so mysterious, due to its seals being broken. Its encapsulation, which resembled (and strongly smelled of) pine rosin had partly melted and oozed out to expose its secret contents. My understanding is that the electronic contents of this catacomb were sealed and encapsulated to protect RCA’s then-innovative Superhet circuit from prying eyes and to ensure any servicing required within was done by their staff. A numbered terminal strip ran along Australia’s electronics magazine the back and a list of various test points between the strip and the valve sockets was published in the owners’ handbook, allowing ohmic readings to be taken along this strip to determine if all was well within. There are also wires connecting terminals on that strip to various valve pins. There was most likely an exchange system available to servicemen when required. My nerve failed me when I considered the consequences of melting away the rosin, so I began picking away at it with a sharp piece of wire. This worked as the rosin was brittle and came away in small pieces. Both step-up audio coupling transformers were missing from the set, but the set came with two transformers which appeared to be a good fit, despite appearing to be of much more recent construction. After this set became operable, I experimented with various other transformers, but the two which came with the set gave the best results. This surprised me because they have a very low DC resistance, the primary being just 300W and the secondary, 800W. The original transformers that would have come in the set had coils with 1000W and 6000W DC resistance respectively. Upon further investigation, it became apparent that the AR-812 frontend circuit was very unconventional (perhaps not surprising, given that superhet conventions hadn’t been established yet when it was designed!). Initially lacking a circuit diagram, I began drawing one out on a large piece of paper using coloured pencils. The mystery deepened and I made only modest headway until I purchased a CD from the USA with a scanned copy of the original operators’ manual and a well-drawn circuit diagram made by another enthusiast, who was apparently also an excellent draughtsman. RCA did not readily give out information regarding the secrets of their catacomb, but there isn’t much to go wrong inside it, except for open circuit conductors. My resistance check revealed one open-circuit coil which I removed from the laminated plate and rewound using litz wire salvaged from a disused IF transformer, taking care to count the turns accurately. There were a few loose wire ends visible, apparently caused by the volatile elements in the rosin drying out siliconchip.com.au The chassis is mounted to the front of the radio with a tuning gang on each side (C1 & C2). There are two compartments on the front of the case which each store half the required batteries. The original set was powered from six A batteries (1.5V each), two or four B batteries (45V/22.5V), and one C battery (4.5V) to provide negative grid bias. While the set is shown with an external loop antenna in the lead photograph, there is an internal antenna located on the rear of the case. It is possible to attach a handle to the top of the set for carrying, but the weight makes this somewhat prohibitive. The back of the chassis shows the connections made from the catacomb. The purple-labelled components above the tuning gangs are Karas Harmonik high-impedance audio transformers which were tested as replacements for T1 & T2. siliconchip.com.au Australia’s electronics magazine August 2019  99 over time and causing the rosin to crack as it shrank, breaking fine wires. I resoldered these into place using the newly-acquired circuit diagram as a guide, along with a certain amount of deduction. No retuning was necessary because the IF transformers had fixed tuning using mica capacitors, and the incoming RF and oscillator output were separately hand-tuned from the front panel. All of the RF inductors, except for the oscillator coils, are recessed into rectangular cut-outs in a laminated iron plate within the catacomb, and they also have laminated iron cores. The audio coupling transformers are also mounted within. The AR-812 also has an internal aerial wound on a thin timber frame encircling the rear of the chassis compartment. There is a simple switch mounted on the inside rear of the case to switch between the internal frame aerial and an external long wire aerial. There are also two drawer-like compartments at the front of the set to house the dry cell batteries, one on each side of the main chassis compartment. Originally, three large single cells producing 1.5V each were connected in series in each compartment, and these batteries were then connected in parallel to give a total voltage of 4.5V for the A supply to heat the filaments of the valves. Four 22.5V dry cell batteries, or two 45V batteries, were connected in series to provide 90V for the B+ supply to the plates of the valves. A small three-cell, 4.5V tapped battery mounted in a pocket at the rear of the chassis compartment made up the C or bias battery. With all of the internals reassembled, all UV199 valves in place and a final positive check for faults in the catacomb completed, I plugged a set of high-impedance (2000W) headphones into the speaker socket and set the speaker switch set to cut out the last two audio stages. I connected my trusty vintage radio power supply from Electronics Australia (March 1990) to supply the B voltage. I prefer to use three alkaline D cells for the A supply via the dropping rheostat as there is less chance of damaging those precious filaments. I also make up a C bias battery by connecting three alkaline cells in series. I connected a long aerial and a good Earth and switched it on but, you guessed it, I heard nothing on the headphones. It was all doom and gloom. But then, few restored radios work the first time. After fiddling with this and that for a couple of cold frosty winter nights, I distinctly remember hearing a very faint whisper in the headphones. Eureka! Things could only get better, and they did; there was a reasonable signal detectable immediately after the detector, indicating that all was well with the RF section. The audio from the audio stages was weak when using a horn speaker. These are usually sensitive and I tried several types, all of around 2000W DC resistance. The audio section is relatively simple and all voltages were around about what one would expect. I suspected that one or more of the valves might have had low emissions. I didn’t have any known-good UV199 valves to swap in for testing. An opportunity presented itself a year or two later when rebuilding a Browning Drake receiver from about the same era. It was recommended that a UX199 valve was used in the first RF stage of this set to make neutralising easier. I made up a socket to fit a UV199 and fitted it in parallel with the 201A or UX199 used in the Browning-Drake, so I could individually test my UV199s in the Drake. One of the UV199s from the RCA AR-812 Superhet proved to have reduced emissions, so I sought out a replacement valve, which improved the audio output considerably, but it was still quite weak. AR-812 performance Although the UV199 valves were The inside of the catacomb with most of the rosin melted away. The leftovers were picked at with a sharp piece of wire, but traces of it can still be seen. 100 Silicon Chip Australia’s electronics magazine siliconchip.com.au Here's an alternative version of the circuit diagram, taken from a service manual. This was included as a supplement to the following circuit diagram on page 102 as it more clearly shows all connections from filaments to HT etc. Source: www.rfcafe.com/references/radio-craft/radiolas-ar-810-812-radio-craft-june-1930.htm passable for RF amplification by 1924 standards, they were only used in the audio output stage because there were no better types available at the time. Audio amplification is not a role that these valves are well suited to. Attempts to improve this situation were later made by fitting a special socket, which had to lay on its side to give clearance to the audio output socket, so a more suitable valve could be used. An extra HT battery was also required to provide the higher HT re- quired for the audio valve. One advantage of the AR-812 is that it is very economical on battery power. The UV199 requires only 60mA for the heaters; only six valves are performing eight functions, achieved by reflexing two of the valves (ie, using them for both RF and AF amplification at the same time). It is fairly easy to tune in stations and to operate the set. Stations are tuned in by using the oscillator dial on the right, which has excellent selectivity. The dial on the left, used to tune the inbuilt aerial, has very broad tuning, making tuning in stations easy, especially once found and marked on the paper dial inserts provided with the set. It is a good idea to back off the filament rheostat before switching off and to slowly turn it clockwise to increase the heater current after switch on until a comfortable (but not excessive) sound level is reached. This avoids damage to the filaments The other side of the catacomb after replacing the two transformers at left. The new transformers had a much lower primary and secondary DC resistance that what the originals were rated at, but performed just as well. siliconchip.com.au Australia’s electronics magazine August 2019  101 102 Silicon Chip Australia’s electronics magazine siliconchip.com.au This circuit diagram was drawn by Alan Douglas using a program called TANGO, and is reproduced here from a scan. The RCA AR-812 is a reflex receiver and one of the first superheterodyne sets. Each valve in the AR-812 has quite a low current draw of ~60mA, which is why this circuit can be powered from dry cells. RF amplifier V1 is reflexed to function as an IF amplifier, while V2 performs as both the local oscillator and mixed (first detector), meaning the circuit effectively has eight stages. V3 is used for further IF amplfication, V4 is the second detector and V5/V6 are both used for AF amplification. The set uses a fairly low IF frequency of 45kHz, although some documents indicate it being as low as 40kHz and as high as 50kHz. Some versions of this circuit have the two transformers connected to the grid of V5 and V6 with a turns ratio of 1:3; this circuit has a ratio of 1:6 which matches the service manual. A wave trap may be needed for local stations as they can come in at multiple places due to the set’s design. A cleaner version of the circuit can be found at https://antiqueradios.com/gallery/main.php?g2_itemId=48147 on the second page, but it does have some slight differences to the circuit shown above. Connection diagram (left) and continuity test (right) for the catacomb. by overheating, which can cause the thorium coating, which improves cathode emissions, to boil off. While this set’s performance isn’t high by today’s standards, it would have been pretty good when it was released nearly 100 years ago! Putting it in historical context I have the AR-812 set up and working as I write this, and I can say that it is now performing well. Apart from the sound from the large horn speaker being sibilant and metallic, it’s at least as good as a smaller transistor set on local stations, considering the limitations of their small speakers. One can imagine a family crowding closely around the set, listening in a medium-sized room, but there is still a little to spare because although I have the volume control full on, I do like to back off the filament rheostat to lengthen the life of the valve filaments. There were once claims of coastto-coast reception in the USA, but I can’t substantiate that performance. It is possible to receive some of the stronger Melbourne stations at night here in Hobart, although they come in weakly. After all, this set uses very low gain valves (with a theoretical gain of about five times) and there is no RF amplification in the front end; this results in noticeably louder reception at night. I have another example of an early superhet, an Ultradyne L2 from 1925. Robert Emile Lacault’s Ultradyne L1 came on the market late in 1924, and siliconchip.com.au as far as I can tell, it was the second superhet available to the public. The set’s layout is very different, with the Ultradyne being more conventional and an excellent performer for its time. This Ultradyne set uses eight UX201A four-pin triodes. I also have a 1927 RCA Radiola 60 which was probably the first mainspowered superhet, using then-new five-pin screen-grid tetrode valves and single point tuning. This set gives much better performance than the first battery-powered, cutting-edge superhets. I also have several five- and sixvalve TRF sets from the same era (also using 201As) which perform well on local AM transmissions. It must be said that a set such as Major Edwin Howard Armstrong’s AR812 represented a great leap forward for radio in the early 1920s and superheterodyne receivers are still widely used today for AM reception. The few remaining sets such as these should be restored to working order for the benefit of all those of us in the future who can appreciate the genius of their designers and inventors. SC An advertisement for the AR-812 with its original horn speaker. The radio sold for US $269, without batteries; nearly the same price as a Ford Model T! Australia’s electronics magazine August 2019  103