Return to List

Memoirs of a Radio Serviceman

     In fact I have never earnt my living as a radio seviceman, but I did, while still at school and college, run a nice little sideline in      mending radios, usefully supplementing my otherwise slender income.  I rapidly became so competent at it that I would tackle any make, even the dreaded pre-war Philips mains table radios, which were such a birdsnest inside that most repair shops would not touch them.  They were quite a challenge since many of the components were not marked with their value, the capacitors coated in pitch, making them look like lumps of coal.  My expertise was built up gradually, over a number of years, and always based on practical work, with a soldering iron, constructing circuits and then improving them, and taking them apart again to reuse the parts to build something else. 

      It all started when I was still a kid.  I had long been into Meccano and mending clocks, and with an obvious mechanical bent, might one day have finished up as a mechanical engineer.  But one day, not long after going to the grammar school, my grandparents came across their old crystal set which, years before in the 'twenties, had been their first wireless.  They had by that time long been using a battery set (made for them by my uncle, who did indeed become a mechanical engineer, and a very successful one) which worked well enough - apart from not being very loud and suffering from ghastly distortion.  It lived in a handsome wooden case, also made by my uncle, which housed a "battery eliminator" for the High Tension and for the filaments a two volt lead acid accumulator which seemed to need charging every other day or so. The balanced armature loudspeaker resided behind faded speaker cloth and a fretwork grille, in the popular form of a stork standing on one leg in a pond, in a circular aperture of about ten inches. 

     Rather than throw the crystal set away, my grandparents though I might like it to play with. It came with a set of high-impedance headphones and I tried using the metal spring mesh mattress support of my parents' double bed as an aerial, and a redundant gas pipe outlet by the fireplace in their bedroom (our house was all-electric) as the earth.  I put on the headphones, twiddled all the knobs and listened for any voices or music.  To my disappointment, all I could hear was a faint noise which I was to come to recognise all too well later on as "mains hum". 

Typical crystal sets

   My father came to my assistance; like any man in those days who was even moderately handy, he had made a wireless set (battery operated, of course) in the late twenties or early thirties, before mains sets came down to an affordable price.  He had saved up and bought the components as and when funds allowed, and followed the instructions in one of the several magazines aimed at people wanting to build a wireless receiver.  Practical Wireless, The Radio Constructor and all the others vied for the money of  the impecunious would-be wireless constructor, each bringing their latest monthly issue out earlier and earlier, hoping to capture the sixpence  cover price from their competitors, as no one could afford to buy two. This leap-frogging continued until the latest issue of each magazine reached the first day of the month before the month stated on the cover, a magazine tradition which has continued down to this day.  (Nowadays, there is an odd anomaly, due to postal subscription readers being so numerous, and important a part of a magazine's sales.  Subscription readers don't like to see their magazine on the newsagents' stalls before they get their copy thorough the post.  So they get their December issue in late October.)

   Then, by about nineteen thirty seven, my father had saved up for and bought a Cossor model 396 all-mains three waveband superhet receiver, serial number 5022848, I still have the Instructions for Use and Guarantee.  (The picture below is of the Cossor model 72, although the appearance is identical to that of the 1937 model 396)  The Radio Trader described the set at the time (I came across a copy of the issue with the review years later) as having "excellent bass radiation and good crisp treble". This was true - by the standards of the time, audio frequencies above 4.5 kiloherz were simply not transmitted until the advent of VHF FM after the war.  In the 1950s I had the job of repairing the 396 on one or more occasions; it used a PX4 directly heated triode output valve (under the Cossor number 4XP) capable of about five watts output with 10% distortion. I found out the hard way that while the top cap of the frequency changer was the grid, as usual, that of the IF valve was the anode!  My father used occasionally sit up late at night and, with luck, pick up an Australian station or two on short-wave (the set covered the 16 m to 49m bands inclusive), until the novelty wore off or my mother complained of being woken up in the middle of the night when he eventually turned in, I never found out which.  But I'm getting ahead of the story.


Cossor 396 all-mains three waveband superhet. Reproduced by courtesy of Classic Wireless.   The three gang tuning capacitor shows the set had a tuned RF stage. 


We rigged up a wire from my bedroom at the back of the house, to the cherry tree half way down the garden, and an earth wire running down to a length of old iron pipe buried in the flower border.  In dry weather I assiduously watered the ground around the pipe, as I had been assured that this was the thing to do to ensure the best reception.  And wonder of wonders, I could receive the Home Service, the Light Programme and even the new Third Programme.  But the strange thing was that, once one had suitably adjusted the tuning capacitor, and two umpteen-position stud-switches which connected the aerial to the tuning coil and the tuning coil to the cat's whisker, out came voices or music, with no moving parts. To my mechanically minded soul, this was a wonder indeed: gears, pinions, worms, bevel and contrate gears I could understand, you could see them working, but here was something much more intriguing.

   As it happened, about that time men were returning home after the end of the Second World War, looking forward to civvy street and resplendent in their "demob suits".  The first thing they did when they got home, well - you know what the first thing was, but after that, the first thing they did was to paint the outside of the house, as it hadn't been done since they went off to fight, years before.  Eventually, after doing up the inside and sorting out the garden, they all got around to turning out the loft.  Down came the home-made battery sets they had constructed in the early thirties, before buying a mains set.  Of course, they had carefully stored the battery set, up in the loft, just in case the mains set ever went wrong, and there it stayed when they went off to war.  And of course if the mains set ever went wrong, the wife took it off to one of the numerous radio repair shops of those days, and had it mended - no way was she going to scrabble around in the loft. 

   So the word went round like wildfire that there was this lad in the next road who was interested in wireless.  Old battery-operated wireless-sets, and the components from which they were constructed, rained in by the bag-full, almost by the sack-load. I went to the greengrocers and scrounged one of the wooden orange crates with a partition in the middle, and then another and another - - .  Eventually I had a huge stock of variable capacitors, some with solid dielectric, some with built-in swash-plate slow-motion drives, of basket-weave coils (Igranic - "What are the wild waves saying?"), Benjamin anti-microphonic valveholders, Blaupunkt and Lisssen moving iron loudspeakers, differential reaction capacitors, wirewound one megohm grid leaks, and of course, battery valves with 2 volt filaments and B4G bases, such as the popular HL2 triode amplifier, not to mention complete receivers.  One of the most handsome of these was a Pye "portable" TRF set, with an internal frame aerial and a line-up including three ganged tuned circuits with two rf stages, incredibly all neutralised triodes

The inventor of the triode, Dr. Lee de Forest, reproduced from The Manual of Modern Radio by J. Scott-Taggart. Amalgamated Press Ltd. 1933

This called for substantial diecast partitioning to ensure stability and explained, along with the substantial wooden case, accumulator and HT battery, the enormous weight of the set. The front grille was of the popular "rays of the setting sun over the waves" pattern.  There were two substantial carrying handle on the side and the whole set was mounted on a turntable, so that it could be rotated for best reception.  In fairness, it described itself as "transportable" rather than portable.

   Armed with all this kit, in the late forties I made a one valve set:-

   soon replaced by a two valver and so on up to about five valves, the only expenditure being for a 2V accumulator and, before acquiring an "HT battery eliminator", a 120V HT battery.  Much of my knowledge about wireless at this time was gleaned from The Newnes Wireless Constructor's Encyclopaedia, by F. J. Camm, famous as the editor of Practical Wireless, and brother of Sydney Camm, designer of the Hurricane fighter plane, from a similar work from Odhams and from other such tomes.  (Some years later, when they were obsolete as far as I was concerned, they got used as drilling blocks!  Years later still I bought a copy of the Newnes book and others of the period, from which some of the accompanying illustrations are taken.)

   But another invaluable source of knowledge was Bill, the proprietor of a local radio repair shop in Compton Road, who seemed happy to put up with my hanging about and would yarn about the ins and outs of the trade.  By this time, about 1947 or 1948, his repair trade was still nearly all in mains radio sets, and he would give me old volume controls where the built-in mains on/off switch worked but the track was shot (or vice versa) and other bits and pieces including old linecords.  I repaired a number of these, joined them all in series and sewed them to an old "Utility" blanket, providing myself with a potentially lethal electric blanket. 

Reproduced from "Stray Signals by Point Contact" - in the Maplin Magazine, August 1991. 


Bill was a bit of a wide boy and I remember seeing him carefully wrap a valve in tissue paper, followed by several turns of stout corrugated cardboard and then in brown paper, addressed to the Tungsram Valve company.  Inside was a note on his headed paper saying that the valve had failed under guarantee, and would they please send a replacement.  He explained that in fact he had no idea how old it was, certainly well out of guarantee, as he gently and carefully hit the the packet with a hammer until he heard a gentle pop as the glass shattered. "Good old Tungsram" he explained "damaged in post - they always send a free replacement".

   About this time, old discarded mains radio sets started to come my way, relieving me of the bother of charging the accumulator, but posing a distinct health hazard.  After a few painful shocks, I quickly acquired a healthy respect for the mains and for the HT supplies used by these sets.  By this time I had transferred my allegiance to Charlie's radio repair shop in Kingsley Road. Charlie (that's what we all called him, his wife called him Bert) was a bluff northcountryman with set views on life.  He allowed his wife a carpet sweeper but would not hear of having a vacuum cleaner in the house, maintaining that over 90% of what they picked up was pile off the carpet.  He had never taken a driving test - they didn't exist when he started to drive.  As a youngster he had worked for a north country engineering firm which started to manufacture "motor lorries".  One of his first jobs was to deliver a lorry to a firm in Lancashire.  When he objected that he couldn't drive he was told to take the thing for a couple of turns round the  yard to get the hang of it, and then get on with it.  Later, with a friend, he went into making cars; that's another interesting story, but too long for here.  After that he set up a cycle shop, and when a market for wireless components developed in the late twenties and early thirties, cycle shops were the obvious places to stock them, in view of the mechanical nature of variable capacitors, valveholders etc.  


Typical wireless components of the 1920s and 30s,  reproduced from The Manual of Modern Radio by J. Scott-Taggart. Amalgamated Press Ltd. 1933

   By the time I started visiting his shop, the cycle part of the business had long disappeared and radio repairs was the main branch of his family business.  Television was just beginning to appear and since few could afford to buy a "televisor", renting was popular.  Charlie became the local agent for DER - Domestic Electrical Rentals - one of the main TV rental firms of the time. One of his sons was an aerial rigger and installed the essential outdoor aerial, without which reception was deemed to be impossible.  The other son made car batteries in a shed in the back garden.  He bought in cases or reused good ones from batteries taken in part exchange. The lead from old batteries was melted out and reused for the bars between cells, and for casting plates.  These were then pasted, assembled in stacks with spacers between and fitted into the cells.  The finished items, filled with acid, sealed with pitch, precharged and ready for use, were available in three price ranges, bronze, silver and gold.  Charlie explained to me that the more expensive batteries had more plates and thinner, the larger surface area giving a greater reserve of current capability and so a longer life.

   A number of us, of whom I was always the youngest,  used to foregather in Charlie's shop of a Saturday morning, when business was slack, and exchange yarns.  This formed a useful school which expanded my knowledge, otherwise gained from books borrowed from the Library, The Wireless Constructor's Encyclopaedia and similar books I had acquired along with all the radio parts, and from Practical Wireless.  When I was younger, my father used to buy me regularly a copy of the Dandy or the Beano - they came out on alternate weeks - while my sister got the Girls' Crystal.  However, as soon as I started collecting all those wireless components, I persuaded him to ditch the weekly comic in favour of the monthly Practical Wireless. This contained advertisements for Henry's Radio, G. W. Smith of Lisle Street, Relda Radio, Proops and many other dealers in ex-government wireless and radar equipment, a vast selection of which was still available for many years after the war.  Thus it was that in due course I saved up my pocket money for weeks and purchased an Indicator Unit 182A, a display unit from an airborne radar set. It contained a 6 1/2 inch electrostatic deflection cathode ray tube type VCR138A, having a blue flash and yellow afterglow.  This made the Indicator Unit 182A much cheaper than other Indicator Units fitted with the green short persistence VCR97 cathode ray tube, which people were turning into television sets.  I turned my unit into an oscilloscope, the first of many I have owned.

   Even before this, however, I had started repairing radio sets, using a stock of spare parts I had salvaged from old mains sets and from other pieces of ex-government equipment.  Thus I was usually able to get a set going at little or no cost to myself, so that the five bob or so (five shillings equates to twenty five pence) that I charged a customer was all profit.  A frequent complaint from a prospective customer was that their set was working fine, it was just that the sound was horribly distorted.  This was invariably due to grid current in the output valve, due to the valve being a little "soft" - a technical term meaning that the vacuum inside wasn't quite as good as it should have been - often resulting in a little blue glow around the anode, a problem particularly prevalent with Mullard output valves.  This upset the bias conditions, but a dodge could get around the problem.  I would explain to the customer that what their set really needed was a new output valve, but the valve alone would cost seventeen shillings and sixpence (eighty seven and a half pence), and then of course there was the additional cost of fitting.  But alternatively I could offer them my special fix for five bob, which would cure the problem and keep the set going probably for a year or more but no guarantee.  Invariably they opted for the five bob fix, which entailed lowering the value of the grid leak from its original value of half a megohm or so, to thirty three kohm.  The grid current no longer upset the bias conditions and the set sounded fine, if a little short on bass.  For special customers, I would increase the size of the coupling capacitor to compensate.  It was remarkably effective, I never had a customer bring a set back. 

   I learnt an awful lot from Charlie and my time with others in his shop.  One of his nice little earners was a neat small four valve long and medium wave radio in a smart veneered wooden cabinet.  These he  made up from kits factored by a wholesaler. He would also sell me just the parts I needed for a modestly lucrative sideline - a cabinet and chassis fitted with a five inch loudspeaker, a two gang tuning capacitor and a dial.  I would turn this into a set looking just the same as his, using valves and other parts salvaged from an ex-government airborne four channel receiver, part of the TR1196, a 4.3MHz to 6.7MHz transmitter/receiver.  The picture below shows one that had been converted: the original would have had a set of switched tuning in place of the two gang capacitor and scale shown. 

The 1196 receiver, available from Charles Brittain Radio of Upper St. Martin's Lane and similar ex-government stores in the 1950s.


I sold these sets to friends of my sister, fellow students at university, at a good profit for myself (if you don't cost the time involved) but still at a much lower price that in the shops.  But of all the yarns I heard at Charlie's, some concerned the dodges employed by unscrupulous radio dealers - like Bill.  I'm sure Charlie was above these, but one has stuck in my memory and comes back to me whenever I think of my father listening in on short waves before the war. 

   Long after the war many men liked to stay up late at night and listen to foreign stations on the short wave bands, often on a set that had been bought before the war.  These sets really needed a complete new set of valves, which - unlike transistors - gradually wear out and become less and less effective.  A customer taking such a set to a radio repair shop, if he was unlucky enough to have chosen one that was not too scrupulous, might get his set back with assurances that it was now all mended and that will be two pounds ten shillings please.  On getting it home, its performance on medium wave was the same as ever; the long and medium wavebands were easy to receive.  On short wave, they would at first be delighted: instead of finding only just one or two stations across the whole of the short wave bands like before, the set would now be bursting with dozens of stations. 

   Eventually it would dawn on them that they were finding the same few stations over and over again, all the way along the dial.  If they raised the problem with the dealer who had "mended" the set, he would assure them that their set was fine; it's just because of the Russian jamming, stations nowadays have to broadcast their programmes on lots of different frequencies, to get through to listeners.  In fact, it was a case of a grid leak again.  The grid leak of the triode oscillator section of the frequency changer stage was usually thirty three kohm, a little on the low side for long wave but quite acceptable, and just what was needed on short wave.  If this were replaced by a 330kohm resistor the set would operate as before on LW and MW, but on short wave the oscillator would "squegg", oscillate in bursts, about two hundred thousand bursts per second.  The result is effectively not just one oscillator frequency, but lots of them, so that each stations can be received by each of them in turn, as one tunes along the band.  I can't quote from my own experience of any case of this deception being practised, but Charlie maintained that he had had to advise a customer or two that that was what had happened to their set and I can well believe that the practice was widespread.

   In 1953 or thereabouts I was taking a BSc. Elec.Eng. at Imperial College and was very proud of a valve voltmeter I had constructed.  I already had a homebuilt multimeter - after the style of an "AVO", it even had ac current ranges - but its sensitivity was only ten thousand ohms per volt, due to its ex-government surplus 100 microamps meter movement. My new valve voltmeter used a double triode in a balanced arrangement making it very nearly drift-free, despite the vagaries of the mains, and the input resistance was simply the ten megohms of the input stage's grid leak.  I told Charlie about it one day, proudly announcing that it had such a high input impedance I could even detect grid current in an output valve.  Charlie said he could too, but he didn't need a valve voltmeter.  I said his multimeter wouldn't do the job, as it only had the same input impedance as mine, but he soon put me down.  "I just measure the output valve's anode voltage and see if it changes when I short out the grid leak."  It had never occurred to me that you could use the output valve as its own valve voltmeter!  Yes, Charlie may have had no qualifications, but he certainly knew a thing or two.

Gracious living, 1930s style

This page last updated - see homepage.

Return to List