Crystal radios

Author: Peter Hadgraft

I intend to write a series of articles on crystal sets used on the broadcast band, I will assume the reader has a basic knowledge of resonant circuits, along with frequency and wavelength. If you don"t follow please ask or check in a suitable book.

The P2 model crystal radio from USSR Broadcast frequencies are say 535khz to 1600khz with corresponding wavelengths of 560 metres to 187 metres. BC stations use a vertical antenna which creates vertically polarised waves and are best received (under ideal conditions) by a vertical antenna. The aim is to get your antenna as high as possible to get as strong a signal as possible. Because of the wavelengths involved your antenna will be shorter than one quarter of a wavelength, and will therefore look like a capacitor plus some resistance. The earliest crystal sets resonated the aerial with an inductor, e.g. a coil of wire with fixed taps or sliding contacts. This is called “loading the antenna”. By varying the number of coil turns the antenna resonates at different frequencies.

The energy in the resonate circuit is then tapped by a second circuit which includes a detector and earphones.

The tap from the detector circuit is adjusted so that it is possible to hear the required station while rejecting others close to it in frequency.

Remember a crystal set is a compromise between sensitivity (strength of station) and selectivity (ability to separate stations). The aerial tap affects the sensitivity mostly and the detector tap the selectivity. They will interact, you may need to move both taps when you want to select different stations. For a practical circuit, wind 135 to 150 turns of 26 gauge wire (0.4mm, approx. 25gns in weight) on a 200mm length of 50mm PVC plumbing pipe. The size of pipe, wire or number of turns is not critical, start by experimenting. Tap the coil every 10 turns from one end and every 5 from the other. This will allow coarse and fine adjustments, another idea is to wind the coil on a cone rather than a cylinder, also you can use sliding contacts to do the tapping. Use a germanium diode and high impedance headphones of 2000 ohms or higher.

The performance of the crystal set will depend on every part of the set. The most critical are the aerial and earth, together they determine the energy fed in to the rest of the circuit. The higher vertically the antenna, the stronger the stations, but there is a limit to how high you can put them. It is possible to make the antenna appear to be higher than it actually is.

  • (a) Horizontal wires in L or T configuration can be added to the top of the antenna, this is called a "capacitance top-hat"
  • (b) The aerial can be loaded as in below (right), this is called "bottom loading". A coil can instead be placed at the top of the antenna as below (left), this is called "top loading", however this is difficult to adjust or tune to different frequencies.
  • (c) The resistance of the aerial can be reduced, instead of a single strand or one stranded wire a number of parallel separated wires can be used. This is called a "cage" or "sausage" aerial. Typically six wires are used in a hexagon configuration using insulated spacers. The wires should be the same length over the total antenna.

Now lets improve on our crystal set design

The aerial is "bottom" loaded by means of the variable aerial tap, until the aerial circuit resonates at the station required To give high sensitivity and some selectivity the circuit formed by the coil, taped at about 80 turns and a single gang 415 pF capacitor is tuned to the required station. To improve selectivity the antenna circuit and capacitor controlled circuit are now two "close-coupled" circuits tuned to the same frequency. The detector circuit consisting of a germanium diode (OA85, OA91 etc.) and 2000 ohm earphones is tapped at about 25 turns. The detector is now matched to the tuned circuits for best selectivity.

The most common earth is a connection to your (steel) water pipe where it enters the ground. The dirt around your house has resistance and this can be improved

(a) Use three spikes (steel, copper or copper-clad steel) driven a metre in to the ground and a metre apart in a triangle configuration, connected together and to the set.

(b) Add chemicals to the ground, the commonly recommended one is ammonium chloride solution around the earth spike. However this is very corrosive and you may need to replace the earth spike every few years.

(c) Use a counterpoise, this is a mat of wire laid out under the antenna. This can be a series of radial wires joined at one earth point to the set. Alternatively the wires can be in a grid joined (or soldered) at each crossing point with the earth connection at a suitable place near the aerial downlead. The counterpoise can be buried if you prefer, use a smooth edge garden edger tool to make a thin slot in the ground and push the wire in to it.

Here is another circuit to try:

Use the coil already would for the previous circuits for the left-hand side (aerial input), for the detector side wind up to 100 turns of 0.4 mm wire on a tube which will just fit over the other coil. This could be a PVC or cardboard tube. Tap this coil every 10 turns at the top and every 5 turns at the bottom. Experiment with the coils side by side and inside each other as well as with different tapings.

We now have a loaded antenna with 2 coupled tuned circuits. The coupling on the LH side is tight the second coil has variable coupling; tight if the coils are inside one another and loose if side by side. This arrangement improves selectivity) at the expense of sensitivity). The looser the coupling the better the selectivity.

Another source of loss is the wire used in a crystal set. The wire in the 150 turn coil (26 gauge wire) in a crystal set has a DC resistance of about 2 ohms However Broadcast Band waves only travel on the surface of the wire, "the skin effect" BC waves penetrate to a depth of about 0.05 mm to 0.09 mm. The wire is 0.4mm diameter its resistance at BC frequencies is about 3 to 5 ohms which is not significant. If you built a Short Wave crystal set, this resistance could be significant as the resistance increases with frequency. Some purists like to use Litz, which consists of a number of insulated strands of very fine wire. eg. Antique Electronics of America sell one litz wire which consists of 54 strands of 38 gauge wire (0.1mm), and another one which is 25 strands of 41 gauge wire (0.07mm) The 38 gauge litz is ideal at BC frequencies , The 41 gauge litz will not change resistance noticeably until the frequency is above 3.5 mHz

The biggest loss obviously is in the detector circuit, the average resistance of a crystal used in 1920 was about 2000 ohms at BC frequencies, earphones of 2000 ohms were recommended because maximum power is available to the earphones if the resistance of diode and earphones are equal.

Fig 8 is a variation of the circuit in fig 7 using the same coils.

The two coils are kept well separated, preferably mutually at right angles

The top tuned circuit is configured as a wave trap. It is tuned to a station you wish to reject, it will look like a high resistance to that frequency. The rest of the circuit is the same as the improved version of fig 6.

The Crystal detector or diode rectifies the radio wave so that it feeds a DC signal, varying at an audio frequency rate to the earphones. This audio is what we hear.

When the diode is forward biased it does not initially conduct until the voltage reaches a threshold value at A. There is a slight curve between A and B , and then a relatively straight line from B to C and beyond. The initial threshold voltage varies from diode to diode eg.
  1. Lead Sulphide (Galena) ..... 0,3 volt
  2. Germanium ..... 0,3 volt
  3. Silicone ..... 0,6 volt
  4. Silicone Carbide (Carborundum) ..... 3 volt
  5. Silicone on metal (Schottky) ..... 0,4 volt
The curves AB and BC are much the same for all solid state detectors theoretically there is little to choose between them. To overcome the threshold voltage the diode can be permanently forward biased so that it starts operating no matter how small the radio signals applied.

The potentiometer should allow about 1 milliamp to flow from the battery. Measure the current flowing in one of the arms from the potentiometer, adjust until current just starts to flow. Alternatively listen on the earphones to a weak station and adjust for maximum loudness (hard to judge)
You may have seen circuits using multiple diodes, which supposedly increase the output of the crystal set, remember you are limited by what the aerial and earth system can deliver to the crystal set. Each diode added increases the resistance and voltage drop. The earphones should correspondingly have a higher resistance, an extra 2000 ohms for each diode added. Note that ordinary circuits use one diode, a half wave rectifier, using one half of the available radio wave. The circuit below shows an example of full wave rectification.

Wind another coil of say 60 turns on a former which will just fit over the existing coil. Tap every 5 turns with one lead to the earphones from the centre of the coil. The diodes have the same number of turns between them and centre tap. One diode works on one half of a cycle of radio signal, the other diode on the other half cycle, there is only one diode conducting at a time.

So you have built a crystal set but can"t get any 2000 ohm earphones, don"t worry there are ways around this.

In the basic crystal set circuit, the detector is tapped down on the tuned circuit coil. This matches the detector circuit to the tuned circuit at a point of reasonable balance between selectivity and sensitivity, the tuned circuit looks like a resistor of say 40.000 ohms. The diode and earphones together have a resistance of say 4,000 ohms.
  • The coil has 80 turns, tapped at 25 turns, the turns ratio is 80/25 = 3.2.
  • The resistance is transformed as "turns ratio squared" Turns ratio = 3.2
  • Resistance ratio = 3.2 squared = 10.24

    The tuned circuit resistance is transformed down at the 25 turn tap by a factor of 10.24, i.e. from 40,000 to around 4,000 (the resistance of the detector circuit)

    In the same way you can make a low impedance earphone look like high impedance earphones. Suppose you have 150 ohm earphones, you need to match these to 2,000 ohms.
  • The resistance ratio is 2,000/150 = 13.33,
  • corresponding turns ratio = square root of 13.33 = 3.65

    An audio transformer is required which has a turns ratio of 3.65 to 1, a small power transformer can be used such as 240 volts to 65 volts, 115 to 37, or a transistor radio transformer.

    Good quality earphones have their diaphragms placed close to the magnets. Some have adjustable spacing. The low impedance earphones are likely to have thick diaphragms, be insensitive and not sound good. Don"t expect too much.

    Moving coil earphones typically 8 or 16 ohms are very inefficient (about 1%) but sound good. They are probably not a good choice for crystal sets.

    Another variation on our crystal set

    Up to now we have varied the coupling between circuits by variable tapping and coupling between coils. Figure 14 shows how we can couple two circuits using a capacitor.

    The two coils are kept separated, between the top of the two coils add a 150 pf variable capacitor. This varies the amount of energy fed from the antenna side of the circuit to the detector side of the circuit. Note that both sides of this capacitor are above earth, it must be insulated from your hands and earth, if you put your hands directly on the capacitor you will change the coupling so put a knob on the shaft.

    Peter Hadgraft
 |  Next >

Add comment:

code protection:*