A safe, efficient method for the external modulation of any light source, laser or non-coherent. The materials are inexpensive, and the equipment can be easily fabricated in an average home workshop.

a web page assembled by Chris Long, Melbourne, Australia.


This web page concerns a technology that has been largely forgotten since the 1930s. It is cheap and readily constructed with scrap parts. It does not use poisonous chemicals, and does not require high driving power. With simple design it can modulate a light beam to 1 MHz, and with special designs the cutoff can be extended to 15 MHz or more.

So why has this technology been forgotten for almost 70 years?

Prior to World War Two, one of the few countries to have a public broadcast television service was Britain. The BBC broadcast John Logie Baird's 30-line TV pictures on medium wave between 1929 and 1935. In November 1936 they upgraded to transmissions on 45 MHz, with two TV systems used on alternate days - the Baird Company's 240 line/25 frame pictures and Marconi-EMI's 405 line/50 field interlaced system. Within a few months the BBC standardised on Marconi-EMI equipment, and this 405 line TV standard survived in Britain until the 1970s. 

TV picture tube fabrication technology was in its infancy in the 1930s. Most CRT's were eggshell-thin by modern standards, and larger tubes were dangerous devices to put in a living room. Many receivers used narrow-angle electrostatic deflection of the electron beam (like modern CRO's) so that the maximum picture size was limited to a diagonal of around 15 to 25 cm, even with CRT's over a metre in length. Pictures ranged from the size of a postcard to about that of an A4 sheet. Any larger, and the CRT would be so expensive or so likely to implode that its domestic usage would be a legal liability to its manufacturer. The early electronic picture tubes also produced dim pictures in peculiar phosphor shades, ranging from green to sepia. Before Alan Blumlein's 'flyback EHT' development was generally adopted, they often derived their EHT from lethal mains supplies, which produced feeble acceleration voltages by modern standards - 4 kV was a typical value - and the phosphors were not aluminised to make best possible usage of the available light emitted.

Then along came the Scophony mechanically scanned TV receiver of 1937, with its rotating mirror polygon scanners, rear-projecting bright pictures of relatively high definition onto a ground-glass screen more than a metre across.

By far the largest commercial British TV set of the pre-war era, it was intended for usage in public places such as the lounges of expensive hotels. The Scophony projection TV was designed on completely novel optical principles, many of which have never been applied since, and these features were - not surprisingly - protected by patents (see reference/footnote list below).

The heart of this extraordinary receiver was an ingenious ultrasound system for modulating the light scanned onto the screen. This type of optical diffraction cell seems to have been devised simultaneously in 1932 by the American team of Debye and Sears; and by the French scientists Lucas and Biquard. However its full practical application to light modulation in television is due to J H Jeffree, the chief engineer of the British firm, Scophony Limited. Formed in 1930 by Solomon Sagall, Scophony's main emphasis lay in developing novel methods of mechanical TV scanning to produce large screen images. J H Jeffree championed the usage of ultrasonic light modulators for this purpose, as a result of which this modulator became known, at least in the late 1930s, as 'The Jeffree Cell'. 

In this TV application, the cell was lengthened to act as a sort of ultrasonic/optical delay line. As the ultrasound modulated by the incoming video signal travelled along the liquid in the cell, the device optically 'stored' the information from about 50 picture elements along its length. The image of the length of the cell was projected onto the screen by a mirror polygon rotating in the opposite direction to the movement of the ultrasound waves in the glass tank, effectively halting the projected image of those fifty picture elements. The velocity of the ultrasound in the liquid had to match the contra-speed of the scan rate - and the two could desynchronise at very high or low temperatures causing blurring. However, by using this 'storage' method, the optical efficiency of the mechanical scanner was multiplied fifty or more times - a particularly ingenious system.

It is especially unfortunate that no complete example of the pre-war Scophony TV receiver appears to survive, in spite of thirty years of enquiry by the author. Within three years of its introduction, the Second World War brought a complete cessation to all television work in Britain. Many TV engineers went to work on radar during the war, so that CRT production technology went ahead in leaps and bounds. This rendered the Scophony mechanical scanner uncompetitive after a remarkably short commercial life, extending only from one prototype at the end of 1936 to the cessation of British TV transmissions for the duration of the war in September 1939.

The ultrasonic diffraction cell in optical communications.

To those of us principally interested in the atmospheric optical communications, the storage aspect of the ultrasonic cell is more a liability than an asset. The cutoff frequency of the device as a modulator is set by the transit time of the acoustic wave across the light beam. The acoustic wave must traverse the light beam in one half the period of the highest modulation frequency required. For maximum modulation speed, the beam must therefore be of small cross-sectional area  - immediately suggesting that a laser would make the ideal source. In fact, gas lasers, which are notoriously difficult to modulate internally, would be ideally suited to modulation by this ultrasonic cell. In practical cases, the maximum modulation frequency is limited to about 20 percent of the acoustic wave frequency. Thus for excitation at 10 MHz - the frequency recommended by Jeffree - the maximum modulated frequency would be around 2 MHz.

The optical ultrasound modulator works on the principles of schlieren optics, where an image is formed by the diffraction of light past a blocking device, a knife edge, or an aperture. Dark field microscopy is one example, the Jeffree cell is another.

A common misconception is that ultrasonic optical modulators are costly and/or complex. In fact, such assumptions were all too recently expressed on the amateur radio 'laser reflector' web discussion page.

J H Jeffree was quick to dispel such assumptions in the late 1930s, and to that end he published a number of simple constructional articles, reproduced on this web page. These demonstrate the principles involved in very practical ways, and the October 1938 article on constructing an atmospheric optical communication system is particularly pertinent application. I can only add that if these effects could be practically applied in a home workshop in 1938, they should be much easier to achieve today with solid-state technology. The quartz crystal wafer required for the ultrasonic driver should be recoverable from any of the older frequency reference crystals manufactured in the 1940s, 50s and 60s - particularly those in insulating holders with socket-mount pins in their base. These are commonly available at hamfests and in electronic disposals stores. 

J H Jeffree's articles are worthy of Internet republication today, as they are only available in extremely obscure sources that are now well beyond the statutory period of their copyright. I also append a very brief extract of Fang-Shang Chen's October 1970 paper on 'Modulators for Optical Communications' in the Proceedings of the IEEE (pps 325 - 326) which, to my knowledge, was the last time that the subject of light modulation by ultrasound was covered in print!

I should emphasise that I have never built this equipment, but the principle of operation seems so promising that I would be fascinated to hear from anybody who gives it a try. In that case, please contact me at:


...or via PO Box 400, Mont Albert 3127. Victoria, Australia. Please note that I only download my e-mails once or twice a week - my reply may be a little slow.


Debye, P and Sears, F W: 'On the scattering of light by supersonic waves' in Proc. Nat. Acad. Sci., Washington, 1932, Vol 18, p.409.

Lucas, R and Biquard, P: 'Proprietes optiques des milieux solides et liquides soumis aux vibrations elastiques ultra-sonores' in J. Phys. Radium, 1932, Vol 3, p. 464.

PATENT: Scophony and John Henry Jeffree: 'Electro-optical light-valves' - British Patent 439236 (March 3, 1934); US Patent 2155659 and 2155660 (27 Feb. 1935, 25 Apr. 1939, both divided 2 March 1937). The patents were for the Scophony ultrasonic/optical storage principle.

PATENT: Scophony and Jeffree, J H: 'Light valve' British Patent 466212 (21 Nov. 1935). Piezo crystal assembly for generating ultrasonic waves.

PATENT: Scophony and G Wikkenhauser: 'Television light valve' - British Patent 471066 (26 Feb. 1936) - an ultrasonic light valve with variable frequency control to produce linear movement of the emergent light beam.

PATENT: Scophony and Jeffree, J H: 'Light valve' - British Patent 469018 (14 June 1936).

PATENT: Scophony, Sieger, J and Jeffree, J H: 'Electro-optic light modulation device' - British Patent 482665 (2 October 1936).

PATENT: Traub, E: 'Supersonic light valve' - British Patent 478840 (17 Oct. 1936).

PATENTS: Jeffree, J H: 'Light modulating device' - U S Patent 2155661 (2 Mar 1937, 25 April 1939. In Britain 3 March 1934) - assigned to Scophony Company.

PATENT: Scophony and Jeffree, J H: 'Light Modulator' - British Patent 497069 (10 June 1937).

Rao, K N: 'Diffraction of light by supersonic waves' in Proc. Indian Acad. Sci. A, 1939, Vol. 9, p 422.

Smith, Alva W and Ewing, Lewis M: 'Diffraction of light by supersonic waves in liquids', in Amer. Jour. Physics, 1940, Vol. 8, p. 57.

Giacomini, A: 'Cella ultrasonora di grande area per la modulazione della luce' in Alta Frequenza, 1943, Vol 12, p 409.

Giacomini, A: 'Cella ultrasonora di grande area per la modulazione della luce' in Acta Pont. Acad. Sci., 1944, Vol. 8, p. 49.

Giacomini, A: 'Cella ultrasonora di grande area per la modulazione della luce' in Ric. Sci., 1945, Vol 15, p 52.

Bhagavantam, S and Rao, B R: 'Diffraction of light by very high frequency ultrasonic waves' in Nature, London, 1946, Vol. 158, page 484.

Humphreys, R F; Watson, W W and Woernley, D L: 'Ultrasonic modulation of a light beam' in J. Appl. Phys., 1947, Vol. 18, p. 845L.

Bhagavantam, S and Rao, B R: 'Diffraction of light by very high frequency ultrasonic waves'  in Nature, London, 1948, vol 161, page 926.

Vigoureux, P: 'Ultrasonics', Pub: Chapman and Hall, London, 1952. A particularly comprehensive coverage of the propagation of ultrasonics in liquids, and coverage of the diffraction cell is given pps. 58 - 68. Valuable information on the mounting of quartz ultrasonic driving crystals in or on liquid optical tanks are given pps 10 - 17. The book has a very comprehensive bibliography. Highly recommended.

Hance, H V and Parks, J K: 'Wide-band modulation of a laser beam, using Bragg-angle diffraction by amplitude modulated ultrasonic wave' in J. Acoust. Soc. Amer, vol. 38, pps 14 - 23, July 1965.

Gordon, E L and Cohen, M G: 'Electrooptic diffraction grating for light beam modulation and diffraction' in IEEE Journal of Quantum Electronics, vol. QE-1, pps 191 - 198, August 1965.

Maydan, D: 'Acoustooptical pulse modulators' in IEEE Journal of Quantum Electronics, vol. QE-6, pps 15 - 24, January 1970.

Click here for a scanned copy of a 1938 Scophony BrochureThis document (about 4 Meg) contains pictures and details of the Jeffree Cell based televisions and other  early television equipment.  For other information on historical television receivers and equipment of this and other eras, visit the Television History web site.


From TELEVISION AND SHORT WAVE WORLD, London, August 1938 pps 461 - 464:



"In 1936 we published the first details of the then newly invented supersonic light control which was developed by the Scophony Company and is employed in their [mechanically scanned 405-line British television] receivers. The issue in which that information was given is now out of print and in response to many requests we publish here a resume of the original article with certain additional particulars of this type of cell which is playing such an important part in modern high-definition television".

Beams of light, produced by apertures, lenses and so on, follow paths that can be calculated by assuming light to be waves of very short wavelengths of about one fifty-thousandth of an inch. When the lenses and apertures are of ordinary sizes, thousands of times larger than these wavelengths, light behaves, for most practical purposes, as rays, and travels in a straight line. When, however, it encounters small structures of the order of size of these wavelengths, it may behave quite differently.

For instance, if we focus an image of a bright filament on a screen with a lens, we have a beam which we can consider as a bundle of rays of light, stretching from each point of the lens aperture to each point of the filament image. We may reflect this beam at some point by interposing a flat mirror, without otherwise disturbing the phenomenon. Suppose, however, we divide the mirror into a large number of small elements, about half of which, taken at random, are raised up one two-thousandth of an inch ) above the level of the remainder (without tilting them at all), then the result will be quite different from what we would expect of a bundle of rays of light. The filament image on the screen will now almost disappear, while round it, at some distance depending on the size and arrangement of the small elements into which we have divided the mirror, will be a coloured halo, formed of the light previously concentrated in the image. If we raise the mirror elements gradually, from zero to one two-hundred-thousandths of an inch, we shall see the filament image gradually fade out and the coloured halo gradually appear. If we cut a hole in the screen, large enough to pass the filament image but not the halo, we can control the brightness of the light passing through it, from full brightness to fairly dark, by a movement of one two-hundred-thousandth of an inch in the mirror surfaces.

Similar small differences are produced in the path of a light beam if it is passed through a liquid in which supersonic waves are propagated. Supersonic waves of wavelengths down to a fraction of a millimetre are readily produced in liquids by the agency of piezo-electric crystal plates, for instance, of quartz or tourmaline, and are propagated from a flat crystal as plane wave-fronts of compression and rarefaction. A light beam, passing through the liquid parallel to these wave fronts, is retarded more by the compressed regions than by the rarefied ones, as the amplitude of the supersonic waves is increased. The principle can, therefore, be used as a practical light relay [(modulator)], either by passing the central beam through a screen, which stops the diffraction spectra, or by letting the spectra pass and stopping the central image with a wire.

Moreover, the supersonic waves move in the liquid with a definite velocity of propagation. This means that when, in accordance with a television [video] signal, we modulate the electric oscillation applied to the quartz crystal plate, we get groups of supersonic waves moving along the liquid, each group corresponding to one of a series of successive modulation values, which are thus strung out in space in their correct order in the liquid.

Simultaneous Light Spots.

If we had a series of light relays, controlling a series of scanning spots following one another on the receiver screen, and if we could delay the modulation of each relay after the first, by the amount of time by which its spot follows the first spot, then we would be able to scan the screen with several light spots instead of one, and get a proportionately brighter picture. This is what the supersonic relay automatically does. Any point in the liquid, at a certain distance from the quartz crystal, will modulate light with a delay equal to the time needed for the supersonic waves to travel from the crystal to that point.

If we arrange our optical system rightly, we do not even need to divide the liquid column into a series of separate light relays. We can pass a beam of light through the whole column, along which supersonic waves are travelling, and then focus it on a slit or wire to separate out the diffraction spectra. We then have a scanning system which forms on the screen a line of light, which is an image of the whole column of liquid.

The size and magnification of this image is such that the images of the supersonic waves, could they be seen in it, would be seen to pass in the opposite direction to that of scanning, and at the same speed. When this line of light (which may be fifty picture elements long, or more) is caused to scan the picture, the modulated groups of waves travelling along the liquid show up as stationary picture details, each produced by the co-operation of what is equivalent to a row of fifty light relays or more.

A simple arrangement is shown in Fig. 1, in which A, B, C, D represent successive picture details as transmitted to the receiver. (The scan is supposed here to be a mirror [polygon] drum. The most varied kinds of scanners can be used in practice.)

Practical Application

In working out the practical application one of the first things to decide is the order of frequency to use. The higher this can be, the shorter the wavelength of the waves in the liquid. This is advantageous, since the angle at which the diffraction spectra are spread out is inversely proportional to this wavelength, being in radians, the ratio of the effective wavelength of light to that of the supersonic waves (for the first order spectra). Unfortunately a limit is set by the fact that high-frequency waves in their initial amplitude in about 20 cm of water or 4 cm of paraffin oil, and the attenuation increases as the square of the frequency. This makes 10 mc [MHz] the right order of frequency to use. The above two liquids are better than many others in this respect.

At this frequency the wavelength in water is about 0.15 millimetre, in kerosene 0.13 mm. Visible light has an effective wavelength of about 0.00055 mm, so that the diffraction liquids travel only a limited distance. At 10 mc they are reduced to a half angle of about 1/280 radians for water, 1/230 for kerosene. Though this may seem small, we shall see that it is adequate in practice.

Length of Liquid

Temperature changes affect the velocity of sound propagation in liquids; usually about 1 per cent per 3° or 4° C. It is often advisable, therefore, to restrict the number of elements shown on the screen, so that normal changes of room temperature will not seriously impair the accuracy with which the scanning motion follows up the wave motion in the light relay. In many cases fifty elements is an adequate number to use.

At 240 lines, 25 pictures per second [the Baird Company's TV standard in 1936], fifty elements (assuming one element equals one line width) are scanned in about 0.000026 second, during which time supersonic waves travel about 3.4 cm. in kerosene. This shows the order of length of the liquid column useful at high definition in simple sets.

Resonance of Quartz Crystals

The next point is frequency response. Quartz crystals are sometimes supposed to be inherently unsuitable for light relays owing to their sharp resonance in air, but it is all a question of damping. Crystals generating supersonic waves in liquids are fairly heavily damped and show reasonably flat resonance curves.

Sharpness of resonance depends on what we may call the "acoustic refractivities" of crystal and liquid. This quantity is the mass of substance traversed by unit area of sound wave-front in unit time; the product of sound velocity and density, for each body. The ratio between the values for crystal and for liquid is the acoustic "refractive index" between them, and determines the re sonance curve. For quartz-water it is about 9, for quartz-kerosene about 14, for quartz-air about 30,000. This explains the much sharper resonance in air.

To a first approximation the percentage off resonance, at which response falls to 71 percent, is:


                                                                 (n - 1) π

Where "n" is the acoustic refractive index. This gives 4 percent for water, 2.4 percent for kerosene. (The crystal is assumed to vibrate at fundamental frequency, not at harmonics, and to have one face in contact with the liquid.)

At an operating frequency 10 mc this means modulation frequencies of 400 kc and 240 kc respectively, for 71 percent response. Though sufficient for medium definition television this is not good enough for high definition.

There are various ways of getting over this difficulty, as in ordinary electrical and sound transmission practice.

Characteristic Curve and Operating Voltages

Fig. 2 is the theoretical operating curve of the light relay. According as one uses either the central beam or the diffraction spectra, one may consider either top or bottom as full black, but the shape of the two ends is not very dissimilar. How far it is realised in practice will be described below. It is calculated by usual optical theory, for monochromatic light, assuming sinusoidal variations of wave path along the liquid, and holds fairly well in most practical cases.

The required operating voltages are most easily calculated from the Lorentz and Lorenz law connecting refractive index and density changes, and from the piezo-electric pressure constants of quartz, since a crystal, in contact with a liquid on one side only, transmits to it, when vibrating at resonance frequency under a given voltage, pressures equal to those producible statically in the crystal by the same voltage. This is a short-cut method of calculation, but correct under the above conditions.

Omitting the working out, we arrive at voltages (reduced to r.m.s. values) of  approximately:


                                                                      20 . gamma . T

for liquids of refractive index about 1.43, such as kerosene.

E = elastic coefficient of liquid,

T = thickness of column of waves,

Gamma = operating frequency.

This formula agrees pretty well with observation. The variation due to refractive index is not very important.

We may note in passing that the pressures needed for full control are not very great, being of the order of half to one atmosphere in practical cases. The amplitude of the liquid vibrations is also of interest: it is of the order of   10-6 centimetres or less, or hardly more than of atomic dimensions.

Thickness of liquid.

To control as much light as possible, the central beam should have an angular spread about equal to the angle of the first order spectra. This enables one just to separate the beam from the spectra when focussed on a slit or screen. This angular spread causes the rays of the beam to stray from a crest of the supersonic waves towards a hollow, or vice versa, in passing through the liquid: they are not all parallel to the supersonic wavefront. The shorter the supersonic waves, the greater the angle of diffraction, which is the desirable angle of spread for the central beam, and the less room there is between successive supersonic wave crests for the light rays to spread. Hence we have to reduce the liquid thickness in proportion to the square of the supersonic wavelength.

The permissible thickness of liquid in accordance with this criterion is about:

 T = 1.3 x 104 λ2

where λ = wavelength of supersonic waves.

This is a maximum value, but it is usually better to use a smaller thickness, especially when the light in the undiffracted beam is being used. For kerosene at 10 mc the maximum T is 2.2 cm, and the minimum voltage therefore comes out about 34 volts rms. In practice it may be two to six times this, according to the arrangements used.


From TELEVISION AND SHORT WAVE WORLD, London, September 1938 pps 537 - 541:


By J H Jeffree

A good deal of work has been done in recent years on supersonic waves, and a few simple experiments on their optical effects will interest "Television" readers in view of their importance as a method of light modulation.

The principle is as follows: A thin plate is cut from a quartz crystal in a certain direction, for instance, perpendicular to the electrical or "X " axis, as shown in Fig. 1. Such a plate has the property of changing its thickness when a voltage is applied to its surfaces.

If alternating voltages of the right frequency are applied, the changes set up a resonant vibration as shown in Fig. 2, both surfaces going inwards or outwards together, and the middle plane standing still. The movement is then much greater than that which a steady voltage of the same value would produce, and causes a reaction on the electrical circuit, which enables such a crystal to be used for control of transmitters. For supersonic experiments, however, we are chiefly concerned with the waves that can be generated in a liquid in contact with such a vibrating crystal.

For this purpose the crystal can either be totally immersed in an (insulating) liquid, or stuck over a. hole in the side of a cell or on the end of a tube, so that one side only is in contact with the liquid. Owing to the fact that the crystal is vibrating in resonance, it needs nothing on the other side to "push against" , and actually gives a stronger effect when one side is free, so that the latter arrangement is preferable: but the former can also be perfectly well used for experiments if one has a suitable trough with flat sides. In either case the effect is, as shown in Fig. 3, that pressure waves travel away from the crystal and cause momentary local changes of density and, therefore, also of optical refractive index. 

If a beam of light passes through the liquid these changes, though small in themselves, can cause marked diffraction effects in it, owing to their sharp variation from point to point in the liquid.

To demonstrate them effectively, we must:-

(a): Excite the crystal at the resonance frequency.

(b): Pass light through the liquid parallel to the wave fronts.

(c):  Focus the parallel beam on a screen.

Cell Construction

Any sort of flat-sided transparent cell can be used, but an open-ended type is better, such as was suggested in recent mechanical television articles. The dimensions of that cell are reproduced in Fig. 4. and it can be made of glass plates with spacers, of glued or painted wood, or ebonite or the like. A container made of a transparent synthetic resin can be bought.

The best crystal frequency is about 10 megacycles, but anything from 7 to 14 mc. will do quite well, so that a standard 7 mc. crystal can be used. It should be mentioned, however, that the high accuracy of grinding, needed for a good transmitter crystal is wasted on supersonic work, and a plate cut in the right direction (from an un-twinned crystal) and ground to about 1/100th inch thick, is all that is really needed.

Electrodes are needed, and may be aluminium foil stuck on with seccotine. It is advisable to lay the crystal on a piece of plate glass, and smooth out excess cement by stroking with a card or ruler. Strips of foil longer than the crystal are convenient, the extra lengths being for connections; they can go off in opposite directions on the two sides. On one side, the width can be that of the hole in the cell which it has to cover, and the other strip can be stuck down only over an area equal to that of the hole. In this way power is not wasted in making those parts of the crystal vibrate which are in contact with the cell walls; effective vibration only occurs where there is an electrode on either side, well stuck down.

As an adhesive, seccotine is suitable, because it remains a trifle damp and does not crack off. Sufficient should be used to make round the crystal a small squeezed out ridge, which should be left. The tinfoil ends are made fast to firm connectors. (Fig. 5).

Two simple lenses are needed, one to take light from a small filament and make it parallel, for passage through the cell, the other to focus it on a screen on the other side. About 8 in. to 10 in. focus is very convenient for these, and those of 8 in. focus stuck on the mechanical television type of cell are quite suitable. Otherwise, however, they need not be stuck on, but can merely be mounted near the cell on either side.

General Arrangement

The general arrangement is shown in Fig. 6. The lamp, with a small vertical filament (i.e., parallel to the crystal) can be a motor headlamp bulb, or anything with a single fairly straight filament not wider than about a fortieth of an inch. Lens A is at its own focal length away from this; then come the cell, lens B and, at its focal length beyond, a screen, such as a piece of card. Simple adjustable mountings, such as are shown in the picture, are convenient, and a foot or so of extra baseboard length beyond the screen would be an advantage for later experiments.

The liquid in the cell may be almost anything that does not attack the materials and cements used in it, but care should be taken in experimenting with various liquids in a cell of synthetic material. Paraffin lamp oil, however, is one of the best liquids to use, and suitable for these experiments. The cell should be filled be fore any voltages are applied to the crystal, since it is not impossible, otherwise, to overload and crack it owing to lack of damping. For the same reason bubbles should not be permitted to lie up against the crystal face.

To avoid creeping of the paraffin, one may wipe the top of the filling hole as dry as possible, and stick down a piece of thin paper on it with seccotine, or use some kind of stopper. Leave a little air in the cell to allow for expansion.


For all supersonic experiments we need an oscillator; it is no use trying to work from D.C. or ordinary sound or picture signals! The simple circuit of Fig. 7 is satisfactory; the thick lines are a simple oscillator, the thinner ones show how to add modulation. 

Any tuning condenser between .0001 mfd. and .0003 mfd. is suitable, and the coil is 10 turns, centre tapped, for a 10 mc. crystal and made by winding D.C.C. wire of something like 24 S.W.G., round the four fingers of the left hand; 5 turns, then a twist for the centre tap, and 5 turns more. The first and last half turns can be bound round the rest to secure it together, and the whole made into a close circular coil about 1.5 in. diameter. There is no need for a more elaborate coil than this.

For a 7 mc. crystal use about 14 turns, and for 14 mc. crystal about 6 are suitable.

A small power valve is best for the oscillator, but almost anything with reasonable emission, and capable of acting as a triode will do. For modulation experiments a mains type with 200 to 300 volts H.T. is desirable, but a battery type with 100 volts a will produce the simpler phenomena. Grid leak and condenser can be almost anything for the present experiments, but low values such as 25,000 ohms and 20 pf. will be needed for television work.

A similar valve will do for the modulator. The coil centre tap is then to be earthed for H.F. by a small condenser - about .0001 for television work - and there is no harm in including this in the simple oscillator from the start. For safety in applying all sorts of modulation, condensers have been included in both input leads. Fig.8 shows the simple oscillator, Fig. 9, with modulation added.

While there are better circuits for television work, this one is perhaps the simplest and gives a positive picture with the sort of arrangement suggested in recent articles. This is the justification for giving component values more suited to television than to the simple experiments of this article.

Inclusion of a milliammeter in the H .T. lead is advantageous for checking behaviour of the oscillator.

Now connect up the filled cell across the oscillator tuned circuit, remove the modulator valve, if any, and switch on the light and the oscillator. First, check whether it is oscillating. The milliammeter will show changes on tuning if it is, and a screwdriver across the tuned circuit will draw sparks. If not, switch off and disconnect the cell; if oscillation then occurs, one can try connecting the cell across the half tuned circuit only (centre tap - one end) on the idea that the damping may be a bit too great for the valve: but unless the latter is a poor one it is probable that the cell is at fault. Look for a short circuit, or an excess of damp cement between the crystal electrodes. If the television value of grid leak (25,000 ohms) has been used, one can try a higher value.

Oscillation having been secured with the cell connected, tune rapidly through the range, and look for any movement in the liquid, spreading of the filament image or sharp maximum in H.T. milliamps. Having obtained such, leave the tuning there and try, by slightly rotating the cell right and left, to get a symmetrical fringe formation of maximum amount on either side of the filament image. This done, the tuning can also be adjusted for maximum effect.

If, in spite of the presence of oscillation nothing call be seen, the tuning may be quite wrong. In that case a faint effect may be found at one end of the tuning range, showing, which way to alter the coil. The most sensitive way to find such effects is often to remove the screen, and look through the cell at various bright and dark objects, keeping the line of sight as nearly parallel to the crystal face as possible and watching for the faintest doubling of outlines as tuning, and line of sight, are varied.

Having got the effect, it can be noted that the milliammeter shows maximum current at correct tuning, indicating that at resonance the crystal draws most power from the circuit. Now the phenomena of fringe formation can be studied.

Fringe phenomena

Each fringe, like the centre image, is an image of the filament, but the fringes are slightly coloured at the edges (apart from colour aberrations, due to the use of uncorrected lenses). The outer edges are redder, the inner blue. Well off tune, there is a single faint fringe each side, nearer resonance a second pair appears, then a third. Note that they are equally spaced from each other, and that the phenomena is not one of actual movement of the light (though it can give that impression; but of the successive brightening up of fringes further from the centre. There is, however, a slight change of spacing during tuning, the separation of fringes growing greater with increasing frequency. The spacing depends on frequency, the brightness distribution on nearness to resonance, that is, on the effective power.

As the fringes increase, the centre image darkens, becoming brownish. This means that blue light is being thrown out of it, into the fringes, more readily than red. With enough power, however, after darkening to a certain point at which it nearly disappears, it begins to brighten up again, and is then bluish. This means that there is a point, for any particular colour, at which maximum darkening occurs, and it occurs sooner for blue than red. Theoretically there then follows a whole series of faint and decreasing maxima arid minima for each colour, but in practice one hardly sees more than the first. Also, however, one can notice that the first fringe on each side now pass through a dark minimum, and (with enough voltage) the second fringes after them. The minimum of the second fringes, however, will be markedly coloured.

To see these effects at their best one should focus the images carefully, and also stop out all except a small part of the cell aperture, about a quarter inch square or so, to get good definition and a comfortable brightness. Fig. 10 shows some of them.

By turning the cell slightly, leaving the tuning set for maximum effect, one can see the effect of skewing the waves relative to the light. The maximum fringe brightness moves to one side, and the central image brightens; the fringes then fade out, to nearly zero, and on further skewing show a slight reappearance. Again, theoretically, one could get an unlimited series of ever fainter maxima and minima, with sufficient power.


From TELEVISION AND SHORT WAVE WORLD, London, October 1938 pps 595 - 597:


by J H Jeffree

This is the second part of the article describing simple experiments with the supersonic light cell. The experiments are of an elementary type which will enable an insight to be obtained in the principle of this class of light modulator.

Last month the operating principles of the supersonic cell were described together with its associated phenomena.

To use these phenomena for light modulation we can either make a hole in the screen, to let the central beam through, stopping the fringes, or else replace the screen by a wire to stop the central beam. Both methods are useful, and either enables us to project a direct image of the supersonic wave track on a further screen. For this we place, a little beyond the first screen or wire, a lens (4 in. to 8 in. focus is convenient) which can make an image of the cell on the second screen (Fig. 11).

Stopping out the filament image by the wire will now obliterate this cell image, until oscillations are applied to the crystal: then, however, the fringe light forms (on the second screen) a bright image of the wave-track.

Using a slit instead, to pass the central beam, we get a bright image of the cell (and it is most effective if all stops are now removed from the latter); and now the wave track darkens when power is applied (Fig. 12).

In each case, an effect of rays proceeding from the crystal is usually seen. These might be due to crystal irregularities, but are more usually caused by irregularities of the electrodes stuck on to it.

When everything has been accurately lined up to start with, the wave track will be seen to be stronger near the crystal end (it is inverted in the image, of course) since at the far end it is weakened by attenuation of the waves. Turning the cell should weaken it, but, if there are errors in focusing and lining up may actually strengthen some parts, where the light was previously passing not quite parallel, to the wavefronts. By moving the cell, plus its lenses, bodily to or from the lamp a little, one gets varying degrees of divergence or convergence of the light through it, with out much affecting the focus on the first screen; and one may experiment in trying to find settings, of combined skewness and divergence of the light, at which the wave track is as uniform as possible right across, thus correcting the attenuation effect. Such corrections are useful in producing good picture quality in television.

The attenuation will be quite obvious with a 14 mc. crystal, but with 7 mc. it is not very marked.

The subtle colour effects obtain able along the track (apart from o those due to lens aberrations !} can also be studied. A ground glass second screen shows these to best advantage.


It will now be of interest to demonstrate modulation The additions to the oscillator have already been described (Figs. 7 and 9). Using a slit to begin with, in preference to a wire for sharpness of effect, we can pass through it, at will, either the central beam or a selected fringe, and get positive or negative modulation. The lens of Fig. 11 can now be moved further away, to a position where it focuses an image of the slit on any suitable second screen, via some sort of rotating mirror or mirror drum. Fig. 13 shows it arranged with a piece of mirror mounted on the spindle of a gramophone turntable as a "scanner."

On applying modulation to the terminals marked, from, e.g., a radio set, and rotating the mirror, a sound track will be traced, in light, on the second screen. Alternatively, one can dispense with this, and view the slit direct via the rotating mirror and lens, focusing the latter to suit. Fig. 14 shows some sound tracks thus obtained substituting the camera for the eye.

The use of a 30-line mirror drum or the like, here, makes possible, of course, the production of a patterned area instead of a mere track, and or very beautiful effects are obtainable with music and speech inputs, when actual picture signals are not avail able.

Light Beam Signalling

We may close this section by mentioning that the set up of Fig. 13 with the rotating mirror removed, is suitable also for light beam signal ling and transmission. The last lens then simply focuses the slit on the distant station; the bigger and, in proportion, longer in focal length, one can have this lens, the more light one can put across to the receiving end.

It is not essential, when using the light relay for this purpose, to have a slit after it at all, since without it the filament image is still focused on the distant station, and the fringes go astray to either side and miss it. A slit would, of course, somewhat enhance the secrecy of distant communication by this method.

The general arrangement of such apparatus is shown by Fig. 15, and Figs. 16, 17 and 18 show simple apparatus suitable for loudspeaker reception over a few yards or so. For this purpose a two-stage R.C. amplifier suffices, with a couple of triodes, e.g., an MH4 or similar type in the first stage, and an output type in the second.

The cathode or sensitive surface of the photo-cell is connected direct to the first grid, and these are connected to "earth" through the photo-ce1l resistor R1 which is also the grid leak, and may be half a megohm. The photo-cell anode goes to the safety resistor R2, through which it receives the positive H.T. voltage. This resistor may be 5 megohms at least; it is there to en sure that even in bright illumination, or if the H.T. is too high for the cell, no serious damage will result.

Gas-filled cells have a maximum permissible voltage, above which a discharge passes spontaneously, and with such a cell the voltage applied  should be brought below this, for example, by adding a shunt resistor as shown dotted in Fig. 16. Otherwise damage can be done, and in any case the cell will not function when such a discharge is passing.

With vacuum cells the applied voltage does not matter, within. reason, so long as it is above 30 volts.

The higher the value of R1 the greater the sensitivity, but also the greater the positive potential applied to the grid of V1 for a given illumination. If this is enough to make the latter positive it will provide an alternative conducting path for the cell current, greatly lowering the sensitivity, so that the higher the photo-cell resistor, the less the permissible general illumination of the cell.

For this reason it is well to exclude extraneous light as far as possible, for instance, by a couple of apertures, in line with the transmitter, each being the size of the cell window; or by a "tunnel"; as in Figs. 18 and 19. In the photograph the cover of the cell and the screen and screen barrier has been removed.

An alternative way of applying the signals to the first grid is shown in Fig. 20, and has the advantage of avoiding variation of effective grid bias with general illumination. Is is useful when higher sensitivity is desired, with a sensitive cell, in the presence of extraneous light. Naturally, the effective photo-cell resistor value is that of two resistances in the grid circuit in parallel.

In case it should be desired to use such apparatus for the foregoing for long range communications by light beam, the methods of increasing range, from a few yards to many miles, may be mentioned briefly. In the first place, amplification may be increased a great deal in the receiver; with a little care in decoupling a thousand-fold increase could be got by adding two more stages. The image increases as the square root of this gain, so that it would thus be magnified about 30 times.

Secondly, a collecting lens at the receiver, as large as possible irrespective of its focal length, increases light pickup in the ratio of its own area to that of the cell window (nearly); and the range is thus increased in the ratio of their respective diameters. A 6-in. lantern condenser lens, for instance, would increase the range four or five times. Fig. 21 shows the arrangement. The aperture in the screen before the cell should then be no longer than is needed to pass the focused spot of light from the transmitting station.

Thirdly, a long-focus lens of sufficient diameter can be used as output lens at the transmitter, in place of the small one of 8-in. focus or so, as hitherto assumed. Whatever its focus, that is the distance at which it must be set in front of the filament image to project the latter on to the distant station. If, with the existing angles of divergence of light from this image, the area of the lens illuminated by it is then increased, the response at the receiver goes up in the same ratio, or the range goes up as the square root of it. A large telescope objective is most suitable for fulfilling this condition. One of 2 inch diameter and about 24 in. focus will nearly double the range, and larger sizes will increase it still further in proportion to their diameter. Fig. 22 shows the principle of this.

Then, of course, an arc lamp may be substituted for the filament light source, with a slit, about one-fiftieth of an inch wide, in front of the positive crater, to give the required shape of source. It must run on smooth D.C., however, if speech or music are to be transmitted. A further three to six-fold gain in range is thus practicable. These four measures together suffice to give a range of several miles.

Since the first instalment of this article, quartz crystals as there suggested, cut approximately, and operating at about 10 megacycles, have been made available by the firm supplying the synthetic resin containers also previously mentioned. These should be very convenient for supersonic experiments.


From TELEVISION AND SHORT WAVE WORLD, London, April 1939 pps 217 - 218:


by J H Jeffree

In preceding issues, details of a simple mechanical-optical [television receiver scanning] system for home construction have been given. This article describes the radio receiver for use in connection with this system. Full constructional details will be given next month.

The fact that the steel ball scanning system recently described works at somewhat reduced definition makes it permissible to use with it a vision receiver with lower vision frequency response than is customary. Since frequency response and magnification are inversely related to each other, this reduction in the former makes possible a proportionate gain either ill the latter or in some equivalent respect such as reduced cost or addition of refinements.

After a good many experiments, starting with attempts to cut the cost, above all things, to a minimum, it seemed in the end more profitable to include the sound, in a simple receiver of relatively low cost and high gain, rather than to provide vision only in the cheapest possible one. Sound is necessary, and its cost, if a separate receiver is built, has to he included in the total. We therefore present here a simple and somewhat unorthodox receiver with modulated 10 mc. vision output suitable for operating a super sonic light cell and giving also a small output of the television sound in tended to be fed into the pick-up terminals of an ordinary broadcast receiver, or into any sound amplifier, but sufficient also to operate headphones, which is a convenience when experimenting .

The design considerations start with the requirements of the super sonic relay. This, in the simple form used in the mechanical set, needs something like 100 volts of modulated 10 mc oscillation, the required modulation being negative, so that the sync. signals are peaks of oscillation. This suggests at once the use of some form of grid modulation, so as to restore the D.C. picture component by rectification at the modulator grid, without a separate valve. One could, for instance, use a small 10 mc oscillator driving a suppressor grid pentode with the vision signals applied to the control grid.

To obtain good definition with the simple light relay, however, some sort of correction of its frequency response curve is desirable, and the best arrangement is probably a band-pass input circuit to it, the peaks of the band pass being arranged on either side of the crystal peak. (Fig. 1.) This involves two 10 mc tuned circuits, and if a third, for the driver valve, is added, the adjustment is somewhat tricky. If it is wrong, the advantage of the band pass is lost. In this receiver, therefore we are grid modulating a self oscillating pentode to save the extra tuned circuit adjustment.

An oscillator in a band-pass circuit such as Fig. 2 automatically oscillates, if the two circuits are about in tune at the centre of their band-pass response; it will therefore only be necessary to adjust them, alternately, till the maximum crystal response is secured, to have the whole thing correct. (This would only not be the case if the crystal coupling drew so much energy, at resonance, as to damp the oscillation markedly.)

Now grid modulation of an oscillator is not always possible. In such a circuit as Fig. 3 it is possible only so long as the oscillation is gentle enough not to swing the grid into grid current; if it is too strong, and this occurs, the grid current, and not the modulation, regulates the oscillation. The desired condition only obtains w hen the effective feedback amplification ratio (grid back to itself) lies between 1 and 2, and as it approaches 2 the sensitivity to modulation goes up rapidly. Now since the effective value of this magnification factor depends on the momentary mean grid voltage which is following the modulation, distortion is probable and the arrangement is somewhat critical anyhow. It is, however, very sensitive to modulation.

If a resonant circuit be connected as in Fig. 4, Hartley fashion, between plate and screen of a pentode, and the valve oscillates, it will be possible to modulate it to some extent, as the conditions are different from the former case. The oscillation is likely, however, to be weak and easily stopped by negative modulation. If, however, the control grid be blocked to the oscillation by a choke (dotted) it will receive enough excitation from the screen, by inter-electrode capacity, to assist markedly in maintaining oscillation, and modulation can still be applied.

Moreover, the increasing effective feedback with positive modulation can be partly offset by the increasing screen damping with stronger oscillation, so that the adjustment of the circuit is not critical. It can, however, still be stopped oscillating by strong negative modulation, an occurrence which shows as blank areas in the picture following the highlights; so a small unmodulated I triode can be added in parallel to maintain a weak oscillation at such times. Fig. 5 shows schematically the resultant circuit. In practice, a 20 mmfd. condenser from screen to grid to supplement the inter-electrode capacity, improves the performance, without making it too difficult to preserve the higher modulation frequencies.


From TELEVISION AND SHORT WAVE WORLD, London, June 1939 pps 337 - 339:


BUILDING A MODULATED OSCILLATOR with some practical instructions for its use.

by J H Jeffree

Some months ago we gave an account of the elementary phenomena to be obtained with the supersonic relay, and suggested the use of conventional type of oscillator to drive it. Now that mechanical scanning arrangements are available rendering possible the reception of the television transmissions with this relay, it is thought that a more detailed oscillator design, suitable for this purpose, may be welcome. It is intended either for use on picture reception with an existing vision receiver, or for any other experiments on supersonic light modulation that may be contemplated, replacing, for such purposes, the cruder arrangement suggested last September.

This oscillator will give a positive picture with the simple scanning arrangements so far suggested, when driven with a positive picture signal from an existing vision receiver suitable for C.R. tube reception; thus being the converse of the oscillator included in the mechanical television receiver described in the preceding issue. It bas been designed to give, as simply as possible, a reasonably good frequency response to the picture signals and a reserve of power to drive the relay. Such a reserve is highly desirable in picture reception, because it permits operation of the crystal somewhat off tune, with consequent improvement of the over-all frequency response; it also allows the full range of light modulation to be obtained even if everything is not tuned up to the highest possible pitch of efficiency.

As can be seen from the accompanying circuit diagram, the arrangement comprises two pentodes, the modulator plate and oscillator screen being directly connected. This allows a high output from the oscillator with normal H. T. voltage (250 volts) while retaining the advantage, of plate modulation, that the "D.C." component of the modulation can readily be retained. To permit nearly 100 per cent modulation, which is desirable, the mean voltage at the plate-screen connection is kept low; this could be done by the use of a high load resistor direct to H.T. but would then result in loss of the high modulation frequencies by effects of shunt capacity; so a bridge connection, giving a sufficiently low effective load resistance, has been adopted instead.

Cathode bias is adopted for the oscillator to avoid the errors introduced by the appreciable time constant of the more usual grid resistor and condenser at high modulation frequencies. It is easier to make these unimportant with cathode bias, even if the cathode resistor is bypassed, as here, by a condenser. It is also convenient to have the oscillatory circuit at earth (D.C.) potential, and to have one side earthed to H.F. also, which the present circuit permits. Actually the values chosen for the cathode resistor and condenser are such as to produce a small, but desirable, gain in the frequency response at the highest modulation frequencies.

The modulator valve cathode bias is obtained with a fixed and variable resistor in series, to permit of setting the mean level of oscillation as desired. A mean value of oscillation is obtained with a total cathode resistance of about 1,200 ohms, and if desired the variable resistor can be dispensed with, at least for preliminary experiments. and a single fixed resistance of this value fitted. No bypass condenser has been included with this resistor; if greater sensitivity to modulation is desired, it could be bypassed with a 50 µF, 50 volts working electrolytic, but then a useful effect from the point of view of quality, namely the negative feed-back produced by the un-bypassed resistor, would be lost. The modulation curve without this negative feed-back, for a representative pair of valves, is shown in the figure, from which it will be seen that it is not unduly distorted, but the author prefers to omit the bypass at the cost of a little sensitivity and straighten the curve a little.

Oscillator plate and modulator screen go straight to HT + if the valves specified are used; with some similar types the screen voltage must not be the full 250 volts, and a screen resistor of about 5,000 ohms to the modulator valve should then be included.

The tuning inductance has ten turns in a space of 1.25 in. on the former specified, and is centre tapped. For a 10 mc. crystal, such as is now available for supersonic work, the correct frequency is given with the condenser specified about a third 'in'.

Five-pin valves have been used in this unit, instead of the possibly neater seven-pin form, for two reasons; first, because it was thought that many experimenters will have one or two, probably of the five-pin form, available, and secondly so that the unit can be used, in various ways, with triodes also. If a triode (eg., an AC/P) be used instead of the pentode modulator, the results on sound frequencies will still be good, but the input impedance to the modulator at high (television modulation) frequencies will be very low. Alternatively, for simple experiments with the light relay, any type of out put (mains) triode can be inserted in the oscillator holder and will give an unmodulated oscillation.

Some hints on the manner of use of the relay may not. be out of place; they were given in the previous articles referred to, but are repeated for the benefit of those who missed those. While the supersonic relay is very easy to use when its mode of operation is understood, it is possible to waste much time without result, if unguided experiments are made without such understanding. It is further necessary, on account of the patent position, that the relay be constructed at home. Instructions are necessary, otherwise the experimenter may break the somewhat fragile crystal. It may be mentioned that only experimental use of these relays, for television, is permissible without infringing the patents.

The essence of the construction is merely to fix a suitable crystal, with electrodes, over an opening in the side of a suitable transparent container. The type shown in the figure is suggested as suitable for use with the mechanical scanning arrangements that have been described in preceding issues; such containers are obtainable, as are crystals cut for supersonic work, from H. E. Sanders & Co. The crystal is of about 10 mc. frequency, but no great accuracy of cutting, as is needed for frequency control, is necessary on supersonic work.

If a crystal without electrodes is obtained, these must be stuck on. They should be of thin metal foil such as is used for wrapping the less expensive brands of cigarettes. Over one side should be stuck a piece as wide and three times as long as the crystal, to cover the whole surface, with the extra length projecting on one edge for connection. This side is to go in contact with the liquid.

On the other side a narrower strip, as wide as the thickness of the liquid in the cell, is stuck down over an area as big as that of the liquid column only, so as to avoid wasting power on parts of the crystal that will touch only the walls of the container. This strip, also, projects, unstuck, to the side opposite from the first, for the other connection. These electrodes should be fixed with Seccotine, the excess cement being smoothed out carefully with a strip of stiff card or the like, from the middle to two edges, till the foil is in close contact. with the quartz. These operations should be done on a sheet of plate glass, to avoid risk of breaking the crystal by flexure.

The crystal should be carefully fastened over the end of the cell, using a small amount of cement only, between it and the latter, to avoid forming squeezed out lumps that. would obstruct the waves passing into the liquid. After it is fixed, the edge can be gone round again out side to make it quite liquid tight, but the area of the back electrode, corresponding to the liquid column inside, should not be covered, as anything in contact with this area on the back is liable to affect the generation of waves. Suitable cements are Premo-fix, Seccotine and, for a final coat over the edges, one of the shellac preparations made for jointing purposes. These cements are specified because it is found that certain others are liable to loosen from the cell under the influence of the contained paraffin, and allow leaks. If a glue is used, care should be taken not to short- circuit the cell electrodes with it.

Common lamp paraffin [kerosene] is one of the very best liquids to use in the cell, and is therefore recommended, but others can, of course, be tried, so long as care is taken (with a cell of plastic material) not to use anything that will spoil the cell, or dissolve the cement.

The manner of operation of the light relay is, that an electrical oscillation of about the right frequency for the crystal (10 mc.) is applied to the electrodes, and causes vibrations of the crystal in the direction of its thickness, which are transmitted to the liquid in the cell.. The crystal actually becomes alternately thicker and thinner, ten million times a second; it does not (essentially) move from side to side as a whole. The vibrations in the liquid travel, as waves, along it away from the crystal; at this frequency the wave length is only about an eighth of a millimetre, so that some hundreds of these waves are present in the cell at once, like a series of parallel plates laid across the liquid column from one end of it to the other.

Now if we had such such a set of real plates, no light would pass through between them in any direction except parallel to their surfaces, and it would be obvious that if we attempted to pass it in any other direction we should get no results (as, for instance, in a. multi-plate Kerr cell). In the supersonic relay, how ever, the "plates" thus formed consist of the liquid, and are transparent; therefore they are normally invisible, and light can pass freely through them. In fact they have no appreciable effect on it, at all, unless it hap pens to pass exactly parallel to their surfaces, and this is one of the reasons why it is fairly easy to play about with such a light cell and get absolutely no result at all. The light is passing freely through it, but it is skew to the invisible "plates" of the cell, and nothing happens that can be visibly observed.

One necessary adjustment is, therefore, to turn the cell till the position is found where the light passes parallel to the wave fronts as above. The position is not very critical, but has to be right within a degree or so. At this setting, if the oscillation is also about right, a sudden spreading of the focused beam occurs, forming a diffraction pattern on the screen where it is focused.

This, however, indicates a second point. The effect desired is a slight spreading out of each elementary ray of the light passing through the liquid into a number of slightly divergent rays. This effect would be clearly visible, without more ado, if we used a single very narrow ray, but then the amount of light handled would be very small. To use the whole length of the liquid column a parallel beam of light is needed (to pass between the 'plates' at every point) and this is best got from a concentrated filament such as that of a motor headlight bulb, with a lens at the right distance from it to focus its rays into a parallel beam going out to infinity; i.e., at its own focal length away. A bulb of the 36-watt, 6 or 12-volt type, with a single straight coil of filament, is suitable; the filament should be vertical, and a lens of 8 in. focus at 8 in. away will provide a sufficiently parallel beam. A smaller filament is permissible, for experiments, but if a wider one is used the lens must be of proportionately longer focus.

Components for the MODULATED OSCILLATOR.

  RI 1,000 ohms 1 watt (Bulgin).
  R2 15,000 ohms 3 watts (or two 10,000 ohm 1-watt in parallel, as in photograph) (Bulgin).
  R3  30,000 ohms 1 watt (Bulgin).
  R4  250 ohms 1 watt (Bulgin).
  R5  250,000 ohm. 1 watt (Bulgin).
  R6  5,000 ohms W.W. volume control (Bulgin VC44).
  C1 .0003 uF. tubular (Bulgin).
  C2  trolitule variable (Premier).
  C3  0.1 uF. tubular (T.C.C.).
  3x   5-pin ceramic (Clix).
COIL FORMER (Raymart).
CHASSIS (Eddystone).
TERMINALS (Belling-Lee type B).
   2    Pairs.
   6-BA bolts and nuts.
  18--Gauge tinned copper wire
   Systoflex sleeving.
  2-   A.C. (Pen 5-pin Mazda, or Premier).


Brief extract (less than 10%) from Fang-Shang Chen's October 1970 paper


(from 'Proceedings of the IEEE', October 1970, pps 1452 - 1453)

On Acoustooptic Modulators

The light beam to be modulated traverses across an acoustic beam that is amplitude-modulated by the signal to be impressed on the light with the Bragg angle θ as shown in Fig. 8. Part of the incident light will be diffracted by the periodic [refractive] index variation produced by the acoustic wave via the photo elastic effect, and both the diffracted and the undiffracted light intensities follow the envelope of the acoustic wave, their variations being in-phase and out-of-phase, respectively, with the modulation of the acoustic signal. The device is essentially a baseband intensity modulator with its cutoff frequency determined by the transit time of the acoustic wave across the light beam.

A few other factors must be considered in order to achieve successful operation of this device. The acoustic beam must be allowed to spread, either from diffraction or by focusing, so that the Bragg scattering interaction takes place over the bandwidth of the modulator. The light beam must also have a diffraction spread of the same order of magnitude as that of the acoustic beam in order that the light diffracted by the acoustic carrier and its sidebands will overlap to produce an efficient intensity modulation. To diffract nearly 100 percent of the incident light, the number of lines of the grating intercepted by the light beam N (see Fig. 8) must be greater than or equal to 1. This will be satisfied when the diffraction spread of the light beam is made slightly larger than that of the acoustic beam.

The bandwidth of the modulator due to the finite transit time of the acoustic wave across the light beam can be determined as follows. The rise time of the diffracted light intensity due to Bragg scattering by a step-function acoustic wave is shown as:

                                                      tr ≈ 1.3 0 a)

where va is the velocity of the acoustic wave and 2wo is the diameter of the light beam (assumed to be Gaussian) at its waist. Using the familiar relationship between the rise time and bandwidth of a low-pass filter:

                                                      Δf = 0.35/tr

the modulation bandwidth becomes:

Thus the acoustic wave must traverse the light beam in approximately one half of the period of the highest modulation frequency in the band.

The diffracted light beam has the same diffraction angle β as the incident light when the diffraction spread of light and acoustic waves are about equal. In order to insure a good separation of the diffracted and undiffracted light beam, let (Eqn.1):

                                                      θ = β

The Bragg angle is:

where n is the refractive index and fa is the acoustic frequency, and

                                                     β ≈ 2λ/nπω0

for a Gaussian light beam. Upon substitution into (Eqn.1), one obtains (Eqn.2):

From (Eqn.1) and (Eqn.2), the relation between the bandwidth Δ.f and the acoustic carrier frequency fa can be found as (Eqn.3):

                                                          fa  ≈ f

Thus the bandwidth of the acoustic modulators is limited to about 20 percent of the acoustic frequency, which in turn is limited to the order of  1 GHz at present from practical considerations in making transducers.

Maydan [see reference near head of web page] has built an acoustooptic pulse modulator using As2S3 glass. The pulse rise time was 6 ns,  fa=350 MHz, and over 70 percent of the incident light at 0.63 micron wavelength was diffracted with an electrical power of 0.6 watt applied to the ZnO transducer. The bandwidth of this modulator is Δ.f = 59 MHz from (Eqn.3), and the calculated pulse rise time tr is   tr = 0.35 Δ.f  = 6 ns,   in agreement with the measurement. Using rutile and fa = 750 MHz, the rise time has been decreased to 3 ns but with less efficiency (5 percent of incident light diffracted with 1 watt of electrical power) at 0.63 micron [HeNe laser] wavelength.

Although the bandwidth of acoustooptic modulators is not large compared to electrooptic modulators, good acoustooptic materials with high optical quality can be found, which makes this type of modulator suitable for use inside laser cavities.


Back to Modulated Light DX

Last update to this page: Friday 17 June, 2005



Page count since 201005: