Using non linearity in bipolar transistors for optical demodulator designs

Frequency conversion by the transistor used for optical sensing has used either the non linearity of the base emitter diode or the base collector diode. The optimum bias conditions for minimum insertion loss for both examples are analysed using the Ebers Moll transistor model. A simple Spice procedure is used to demonstrate the non linearity. If the b-c diode is used the non linearity is at least /spl beta/ (the current gain) times that when using the b-e diode; it is more when a finite emitter resistance is included, which for an HBT could mean a factor of 1000 times improvement. The frequency performance is reduced by a factor of about 25. The optimum bias for the b-e circuit defines the collector current; for the b-c circuit it is is nearly independent of collector current. A circuit with an added diode between collector and base, which can avoid forward biased b-c problems has identical non linearity characteristics to the b-c circuit and offers some improvement in high frequency performance.