Abstract: A binary-to-bipolar converter converts two binary signals into positive and negative bipolar pulses alternately corresponding to a "1" logic state of two binary signals. The bipolar pulses are defined by the low output impedance states of two transfer gates. The main object of the invention is to reduce the transient trailing of the rear edges in the bipolar pulses. This can be obtained by delaying the drop from the low impedance state to the high output impedance state by means delaying control signals on the transfer gates, or by an integrator circuit applying calibrated pulses onto the output of the transfer gates, in response to the rear edges of the "1" logic states in the binary signals. The converter also calibrates the rise times of the bipolar pulses and allow the use of CMOS circuits upstream of the converter.