Performance limitations for distributed amplifiers

Abstract

AbstractThe model of a distributed amplifier as two cascaded chains coupled at discrete points by active devices is considered from a network theoretic point of view. The active devices are represented by a simplified equivalent circuit, consisting of a current source and of input and output capacitances. The connecting structures are assumed to be lossless, but no initial assumption of image parameter lines is made. This mode of operation is deduced as an optimum from the network equations. The chains can be identified as gate and drain lines formed using a number of MESFET's, if these are assumed to be representable by the equivalent circuits described above. The objective is to sum the gains of the individual devices with maximum gain bandwidth product (GBW) as limited by the parasitic capacitances. If each chain operates with a single resistor, i.e. source in the gate line, load in the drain line, it is shown that with two resistor (input and output) terminations per line, maximum GBW is the sum of individual element GBW's, and requires constant‐k image parameter operation of the lines. Under these conditions the gain must peak sharply at the edge of the band. Any deviation from this mode of operation, e.g. to suppress the peak, will result in a reduction of GBW, and one of the illustrative examples consists of two simple non‐image LC ladder chains designed so the response peak is suppressed with nominal deviation from maximum GBW. The above results strictly hold only under the simplifying assumptions that the whole structure is lossless: although the assumptions of the paper represent ideal operating conditions, the results still provide an upper bound for the limits of reachable gain.