A new digital-to-analog conversion algorithm

An algorithm is described which starts digital-to-analog conversion with the most significant bit. This algorithm can be realized by using only sample/hold circuits, and thus permits high-speed conversion. Error analysis shows that a conversion accuracy of 10 bits or higher is easily obtainable with a monolithic implementation. The architecture of data converters using this algorithm and some experimental results are also given to confirm the principles of operation. it is concluded that the proposed algorithm is universal in that it is applicable to cyclic, successive-approximation, sequential-flash, pipeline, and flash conversion.<
>