Dynamic single phase logic with self-timed stages for power reduction in pipeline circuit designs

True single phase clock logic techniques, e.g. with alternating arranged Nand P-logic cells, yield easy to design circuits with standard cells and high speed potential. The disadvantages are a difficult clock tree design and high power consumption. To realize every logic function, dual rail or differential styles are chosen which increase clock load. This paper presents a method to speed up dynamic single clock circuits. The advantage of asynchronous logic is that the critical path delay is the sum of only the evaluation times of the single logic blocks without wasting time for waiting, latches, or redundant logic. Therefore, this work assembles small asynchronous chains of dynamic logic blocks into one period of the global clock to minimize the unused time per clock cycle (AC-TSPC). However, the synchronous single phase clocking scheme is maintained. The advantages of this method are shorter latencies for calculations, power reduction by smaller clock trees and no need for latches, and a simpler clock distribution network due to increased clock skew tolerance. The results of the simulations of an 8x8 bit multiplier in TSPC and in AC-TSPC show an enhancement in power-reduction of 40% for the logic and of 89% for the clock tree with a latency reduction of 40% and more in comparison with TSPC.