Norcim rc electronics club page 19……
FURTHER RADIO CONTROL ELECTRONICS NOTES.
ANOTHER MYSTERY CIRCUIT FOUND BY DAVID CAUDREY. This is a 35 MHz transmitter section probably of the 1980 era. It is shown below in two parts to be more diagrammatically clear.
The first transistor stage controls the coder pulse rise and fall times. The output of this stage swings the varicap which is part of the 17.5MHz fundamental Crystal oscillator stage. (second transistor). The primary of the collector coil is tuned to 35MHz.
The secondary coil matches the input of the RF amplifier transistor. The collector tuned circuit of this amp rejects harmonics and its secondary matches the low ohms input of the RF output transistor. The output appears to be a double Pi-L filter to the telescopic aerial. An analogue meter was used to indicate transmitter output via the little circuit above the output filters. Transmission bandwidth with harmonic and spurious output would be expected to be better than -40dB at 15KHz. with this circuit.
Had a mail recently from Barry Lennox re the above circuit and he reckons….’’It shares a few details with some of the German sets of the 80's; so, could it be Multiplex or Graupner or ?’’ Thanks for that Barry, a good starting point. Must admit that the German manufacturers were onto the 35MHz band for a couple of years before us in the UK so could even be late seventies.
Now back to David’s expertise with another reverse engineered receiver from yesteryear…….
RCS Mk II GUIDANCE SYSTEM RECEIVER ∆
We are kind of wishing that early R/C electronics Guru Dave McQue ‘was still with us’ to add some meaningful text to go with the above RCS historic circuit. All we can say is that it is a typical super-regenerative single channelreceivers of the time. Such simplicity of those days around perhaps 1960 ? Almost certainly designed by
The transmitter circuit is shown next…….
The RCS MkII transmitter circuit is shown above. The crystal oscillator (NKT675) uses an overtone crystal for feedback. The collector coil with 47p capacitor oscillates at the 27MHz band. (third overtone) The secondary of the coil drives the output transistor via emitter coupling (helpful with HF gain with early RF transistors).
The natty output coil filter is almost a single Pi filter and will suppress harmonics by around 20db.
Meanwhile the AC22s form a multivibrator circuit offering around 800cps tone?. With the button open, transmission of the 27MHz carrier continues. Button closed produces the ‘tone’ (a square wave of on and off carrier at 800cps). Although this direct switching of the carrier was perfectly adequate for the type of non-selective frequency receiver, the sidebands produced would probably span the entire
27MHz band. This phenomena was sorted when selective receivers came along.
<![if !vml]><![endif]>Notes on the reverse engineering of the RCS Guidance System circuits on Page19 Part 1 Receiver
More than the customary effort has been given to reverse engineering of this complete system because it had been purchased by a fellow Hungerford club member ‘David Quinton’ for nostalgic use in a vintage model. It was purchased in a non- functional but complete state and the object of the exercise was to attempt to get the system to work.
The receiver was tackled first and found to be non-functional due to previous replacement of the original germanium input transistor by a modern silicon transistor with too much current gain. A change (not shown) to the bias arrangement for this transistor has returned the circuit to working condition. Ideally it would have been preferable to source and fit a period germanium transistor but the condition of printed circuit tracks was not conducive to this.
With the addition of some period connectors the receiver circuit was then hooked up with a battery and the included Elmic Corporal rubber powered escapement and tested with a signal generator square wave modulated 100%. at approximately 300Hz.
The escapement clicked away merrily as the modulation was switched off and on, the pin of the escapement wheel vertically centred with no modulation and rotating through 90 degrees with modulation present. The mechanical arrangement of the escapement wheel thus effecting ‘one press for left/ two presses for right’ control when operating a rudder via a slotted follower.
<![if !vml]><![endif]>Part 2 Transmitter
Attention then turned to the transmitter. The transmitter circuit was found to be non-functional due to the original output transistor being open circuited but functionality was restored with its replacement by a suitable period transistor.
However when the 27Mz. modulated output waveform was viewed on an oscilloscope the waveform below was observed:
Clearly not 100% or square wave modulation!
Attention then was centered on the modulation circuit. The modulation waveform is produced by a text book astable multivibrator circuit which oscillates only when the control button is pressed. When the modulation astable is not oscillating one of its collector outputs is arranged to turn on a germanium transistor in the emitter circuit of the rf. power output transistor permitting the latter to drive the aerial circuit with a 27 MHz continuous wave; other collector output being unused. However, when the astable is oscillating (button pressed) it might be expected to turn the modulation transistor off and on periodically but, from the waveform above, it clearly does not do so.
To investigate the reason for this, an oscilloscope probe was attached to the base terminal of the modulation transistor to obtain the next waveform:
Clearly the astable is not turning the npn modulation transistor completely off until late in the ‘off’ period; with conduction reducing exponentially from the start of the period. A similar waveform of greater amplitude appears at the collector of the astable transistor but with amplitude significantly less than <![if !vml]><![endif]>that of the waveform observed at the collector of the unused astable transistor:
(200mV/division vertical 1ms/ division horizontal Referenced –ve. Rail)
Reasoning that the loading resulted from the path via the IK resistor , the relevant cross coupling capacitor of the multivibrator circuit and transistor emitter-base junctions, it was concluded that 1k and 6k8 resistors were mutually in wrong positions. However exchanging the positions of these components resulted in loss of the rf. output and consequently the circuit was returned to its original condition.
(5V/division with 12Vsupply. Referenced –ve. Rail)……..
The receiver undoubtedly responds to the weak modulation but with greatly reduced sensitivity compared with its response to the 100% modulated signal generator but the reasons underlying the design are lost in the past.
D.C. Oct. 14
Initial discussion to cure the problem with the Tx, involved minor circuit modification. To date these have been resisted in the interest of keeping the Vintage RCS system ‘as it was designed’ and in its original form. Although the system is now working, Some caution may be necessary on the first flight in a model. A small slow model being best to show any range problem. A fine example of one of the first UK manufactured radio control systems.
Although modifications to the original transmitter circuit have been resisted, a couple of thoughts did come to mind…….
Recently a friend, Ian Kirkpatrick, lent me some copies of ‘Model Aircraft’ magazine which disappeared from the book stalls in the early 1960s. The April 1957 edition of ‘Model Aircraft’ carried an article reviewing single channel radio control receivers then current. The review article is pretty comprehensive; giving a description, detailed perspective drawing, circuit schematic and specification for each receiver and the drawings alone must have required considerable effort by the anonymous author. Unfortunately the article is confused by what must have been cut and paste errors, because the circuit schematics do not correspond with relevant drawings and descriptions. Also the schematics contain anomalies (unknown component values etc.) so it is likely they are the result of reverse engineering rather than produced from data supplied by the manufacturers. The article was probably the work of an old friend, the late Dave Hughes, who worked for ‘Model Aircraft’ before founding ‘Radio Modeller’ with Norman Butcher when MA. closed. Dave was a much respected modeller, graphic draughtsman and journalist.
The schematics are reproduced below, redrawn and with comment on anomalies etc.
The similarity of the schematics, together with the schematics on Page 17 of this site, demonstrate how little scope there was for producing lightweight sensitive receivers in the 1950s. All of the circuits operate by virtue of the phenomenon that oscillations in an oscillatory circuit build up more rapidly from a very small signal at the resonant frequency than they do when building up from circuit noise. It is interesting to note that the principle was also applied in early transistor receivers, despite the inherently greater noise of bipolar transistors compared with valves.
Frog alone managed to produce a hard valve receiver which embodied a sub miniature valve in conjunction with an ad hoc relay. The purpose of the twin armatures and contacts of this relay can only be guessed. Most manufacturers made their own relays because standard relays were too insensitive and heavy. The contrast between radio control in 1957 and today’s 2.4GHz sophisticated ‘buy and fly’ equipment is almost unimaginable but Terry and the author are thankful to have experienced both.
A RETRO R/C RECEIVER PROJECT from David Caudrey…..the aim of this project was to produce a radio control receiver based on early carrier wave valve circuitry. The design however includes an interesting twist that could well have been used in those days gone by, but nobody simply thought about it at the time. The Quench Oscillator could have been used for producing both the HT and the Heater for the valve. This would have resulted in a single (relatively) low cost single supply battery. The work so far is outlined in the circuit below. This however will be added to, and change with development of the project. Semiconductors of the day, where possible are being used.
For readers unfamiliar with super regenerative receivers, it might be advantageous if the following is read in conjunction with Page 17 of this website.
The notion of a project to build a ‘retro’ valve receiver was initiated by the discovery of a Hivac sub miniature thermionic valve type XFY10 among my odds and ends. I can only assume that I had kept this valve for reasons of nostalgia, mistakenly thinking that it to be an example of the Hivac XFG1 gas filled valve which was widely used in super regenerative radio control receivers up to the early 1960s. The alternative to the gas filled, or soft, valves in those days were the much larger hard valves DL91 and DL96 which were used in a circuit in which a single valve furnished both a 27MHz oscillator and the essential quench control oscillator, operating at a lower frequency. These valves, which were designed to provide the audio output power for portable radios, were capable of passing <![if !vml]><![endif]>sufficient anode current to energise a suitable electro-magnetic relay; which is most probably be well beyond the capability of a sub miniature (hearing aid) valve like the XFY10. However in 1957 Frog/Triang in the UK did market a super regenerative receiver which used a DL68 sub miniature valve, embodying what must have been a very special relay. Without such a sensitive relay it would not be practical to recreate a similar circuit but with additional amplification, provided by a transistor, a more robust relay might be used. By 1960 Germanium transistors, capable of operating at audio frequencies and low radio frequencies, were readily available. Therefore my projected, circa 1960, retro receiver circuit will include some form of transistor amplification.
Valve receiver installations needed several sources of power derived from zinc-carbon cells or batteries:- 1.5V for the filament supply, 22.5 or 45V for the anode supply and a supply for the control actuator, which was usually 4.5V. How much simpler things would be if these supplies could be provided from a single four cell battery by a DC to DC converter?
DC to DC converter circuits are usually based upon an oscillating transformer and how convenient it would be if this oscillator could also operate at the quench control frequency?
The foregoing schematic shows the circuit of a quasi-sinusoidal transistor oscillator operating at quench frequency (~200 kHz) and generating the operating supplies for the XFY10 valve. In addition to the oscillator transistor two additional transistors are used in the circuit shown. These provide a degree of stabilisation for the 22V anode supply against variation of the battery voltage but in its simplest form only the single oscillator transistor would be required.
Possibly the easiest way to produce a 27MHz super regenerative detector utilising the XFY10 valve would be to try it in a circuit typical of those used with the DL91 and DL96 valves and to see what happens. However the exercise is likely to be more successful if a circuit is designed to take account of the XFY10 characteristics. Little data is available on the XFY10 so it was decided to measure and plot its characteristics. The photograph shows the valve mounted on strip board for test purposes:-
<![if !vml]><![endif]>Results from the tests are shown in the chart below with Anode Current plotted against applied Anode Voltage with Grid Volts as a parameter.
The characteristics plotted indicate that, with a load of 5kΩ, the mean anode current change which might result from the difference between the build-up of oscillations from noise in a quench cycle and the build-up when an external signal is present is likely to be less than 500uA. Consequently the direct or low frequency voltage signal developed across a 5kΩ load resistor will be less than 2.5V but in any case it should be sufficient to turn on a transistor in a suitable circuit configuration.
The 27MHz test oscillator circuit schematic which is shown below. This circuit satisfies the need for 27MHz oscillations free from limiting and squegging such that an optimal amplitude, frequency and coupling can be determined for the quench control waveform.
Eventually the circuit will become the basis for a test receiver of external signals.
The Test Oscillator breadboard circuit is shown below :-
Below is shown the oscillator waveform under ‘no signal’ conditions. From this 27MHz oscillations can be seen to build up from noise under control of a 180kHz quench waveform. The double humping effect suggests that the quench frequency is too high <![if !vml]><![endif]>and that oscillations build up every other cycle of the quench waveform and consequently it might be worthwhile to reduce the quench frequency to 90kHz. To date it has not been possible to get a sensible photograph of the oscillator waveform when an external signal is present. All that can be observed is that a weak signal greatly increases the amplitude of the oscillations.
The next photograph shows the effect of roughly halving the quench frequency to approximately 90kHz. Here the build-up of oscillations is more ‘text book’. It should be noted that some external low level signal activity is present to obtain this wave form. <![if !vml]><![endif]>In ‘complete absence’ of signal activity the amplitude of oscillation achieved is very small and a small but significant external signal from a generator results in large amplitude continuous oscillation. It should be noted that, to obtain these waveforms, no direct connection is made to the circuit via the oscilloscope probe. The waveforms are obtained via a single turn pick up loop held close to the oscillator coil as direct probe connection gives anomalous results in such a sensitive circuit.
The next figure show the 27MHz Oscillator mounted on the Quench Oscillator/Power supply breadboard and operating with all supplies derived from a 5V bench supply. The transistors and components to the left of breadboard stabilise the anode supply for the valve and indirectly its filament supply. These are not essential for a practical receiver which could embody only a single transistor and components associated with the transformer. However stabilisation is desirable to reduce the number of variables during development testing. A one milliamp full scale meter movement displays the valve anode current for a weak signal; the resultant anode current reduction being about 100uA.
Quench and 27MHz waveforms for the breadboard circuit are show in the next photograph. For reference the quench frequency is 200kHz.
Previous waveforms were obtain using a signal generator to provide the quench waveform with the valve anode and filament supplies provided by batteries.
It is intended that the receiver will operate from an approximately 300Hz. 100% tone modulation, driving a rubber powered sequential escapement directly i.e. without an intermediate relay.
The next two photographs show the oscillator output waveform when responding to a 27MHz.signal modulated 100% at 277Hz and the corresponding wave form across the 100nF filter capacitor, which will be amplified and rectified to drive the escapement .
<![if !vml]><![endif]><![if !vml]><![endif]>
Having separated the tone signal it is necessary only to amplify and rectify it to produce a direct current drive for the escapement and the next figure shows the breadboard schematic for a complete tone receiver. The component count is fairly high but not significantly greater than was used in later transistor super-regenerative tone receivers, once transistors capable of working at 27Mhz became readily available. Also it must be remembered that the use of two batteries, necessary for valve operation, is avoided.
Whether it will be possible to convert the breadboard circuit to a practical retro receiver will depend upon the availability of contemporaneous components. The greatest difficulty will be sourcing of suitable circa 1960 npn. transistors (preferably Germanium) as tall narrow types ( e.g. OC139) are preferable for a compact layout. Also, as has been mentioned, the only Germanium transistors to hand have low Hfe values and it has been necessary to the resort to the use of a Silicon transistor (BFY52) for the voltage convertor/quench oscillator circuit of the breadboard. Similarly a BFY52 transistor has been used to drive the Elmic Commander escapement because of its high current consumption. Another problem will be the sourcing of a germanium junction diode for the valve heater supply. Germanium signal diodes (e.g.0A90) are available but these exhibit a three volt drop at 25mA and consequently it has been necessary to use a transistor junction or modern Schottky type in the breadboard circuit.
Thanks for reading !