Optical Interfacing.


In optoelectronics the term interface pertains to the point in the system where electrical signals are turned into light variations or vice-versa. In our studies of transmitters and receivers we have touched upon these subjects and  now it makes sense to look at the electronics behind these.  We shall start with the drivers, which in the transmitter took the electrical signal and converted to light, and in the receiver did the opposite. Let us look at the two figures below:




Next >

fig 1 Analogue driver                                     fig 2 Digital Driver

Here we see the drivers for both analogue and digital systems, you will notice that there appears to be an LED at the collector load in both cases but in fact both could be used to drive an injection LASER. In fig1 you will notice that the driver is in fact a standard class A amplifier and has a DC bias This is needed because the signal alone cannot overcome the breakover voltage (about 0.7 V in silicon and 0.2V in germanium of the transistor) It is a potentiometer because even the same make and model of transistor will have a variation and the pot allows us to tune for best current at the base, so we drive the transistor neither into saturation nor cutoff.  The input  causes a variation of voltage at the LED thus varying the intensity of illumination which we can then send through a lens and into the cable. We have already looked at the receiver end when we used an LM386 to amplify directly the varying illumination fed through, but there is no reason why we could not use the configuration above in reverse except for space and possibly cost, as we'd have to match impedances.

You will notice that the digital transmitter in fig 2 has no bias. This is because the transmitter is in fact acting as a switch and not an amplifier. We can now pass preset and unvarying levels of voltage through it. We could pass 5 volts and get a pulse of light, then eight binary digits. For a zero, we may pass 2.5 volts, for a 1 the full 5 again. Or we may pass 0 volts for a zero, but, unless we time perfectly, we run the risk of misinterpreting the data. It makes a more complex circuit to actively rather than passively mark zero, but it is a more accurate one.