5    Semiconductor Diodes

   What are semiconductor diodes? When the old timers of radio were playing with cat’s whiskers as ‘detectors’ in crystal sets, they were actually making a diode using a pointed phosphor-bronze spring and a lump of galena. One had to find a ‘good spot’ in order to ‘detect’ radio signals. What they had actually made was a rectifying diode. This should have been called a demodulating mixer, since the carrier wave and the two sidebands of the Amplitude Modulated signal are re-mixed in it and out comes the difference signal…the audio, ready for amplification.

       Many metals are conductors and insulators are mica, glass, plastics, ceramics, wood, air, vacuum etc. However, semiconductors are man-made. PURE Silicon and Germanium are the commonest parent materials and, surprisingly, are both insulators. There are two types, P and N. Semiconductor type P has been doped (infiltrated) with atoms (phosphorus, antimony or arsenic) which have one less electron than the parent atoms. P type therefore has a lattice structure that has holes in it where electrons are missing.  N type has been doped with atoms that have one electron more than the parent (indium, boron or aluminium). Therefore this lattice has an excess of electrons with no atoms needing them, so they are ‘spare’ as it were.

       When manufactured, both P and N type semiconductor materials are electrically neutral, i.e.each has no intrinsic charge. If we now take some P and N and join them together, the spare electrons in the N type see holes in the P type which need filling and off they go across the junction. But, as soon as electrons leave the N lattice it becomes positively charged. The P type gains electrons so it becomes negatively charged. Eventually this potential difference is large enough to stop any more electrons coming over from the N type and the action stops. This ‘barrier’ voltage is about 0.6v for silicon diodes, but only 0.4v for germanium, which is why the latter are used in small signal circuits. There is now a non-conducting area between the P and N material with no holes and no spare electrons and is effectively an insulator. (depletion layer). This means that there must be a capacitance between the two ends. We can use this effect to make variable capacity diodes. If we connect the P type anode to a negative variable bias voltage, increasing this reverse bias makes the barrier layer becomes wider still, so the capacitance falls.  Reducing the negative bias increases the capacitance again to a maximum at about .6volts. All semiconductor diodes display this effect, but specially designed ones are manufactured for the purpose. In circuit diagrams these diodes have a capacitor symbol next to them. Using vari-cap diodes to
       tune resonant tuned circuits means that the problem of temperature changes, affecting the previously used variable metal plate capacitors, can be almost eliminated. A variable resistor now adjusts a voltage which varies the capacitance instead. The supply voltage must be properly regulated and well smoothed of course.

       Using a diode we can obtain DC from an AC in a power supply. When the P type anode goes negative there is only a little capacitive coupling across the reverse biased diode, so virtually no current flows at all. When the anode goes positive however, the electrons from the N region  see a demand for them from the positive supply and off they go, called conduction of course.

       So large pulses of electron current can only flow when the diode anode goes positive. These half cycle flows of current is called rectification of  AC into DC. By using two diodes and a centre-tapped transformer secondary winding, each diode only conducts during alternate half cycles providing less lumpy DC! This means that the smoothing capacitor has charging bursts twice as often as before, so the ripple frequency is twice the input frequency and the value of the subsequent smoothing capacitor can therefore be made smaller.

       Centre tapped transformers are pricey, so the problem is got around by using four diodes in a ‘bridge’. Two diodes conduct during one half cycle the other two conducting on the other. There are always two diodes in series with the load, so the barrier potential of 0.6v each means that you lose 1.2v in total. You must allow for this when designing a PSU. Big currents need big diodes and possibly heat sinking too.

                                           That’s it once again. Vy 73 Mike G3JKX