As Data conversion systems continue to improve in speed and resolution, increasing demands are placed on the performance of high-speed Analog to Digital Conversion systems. This work makes a survey about all these and proposes a suitable architecture in order to achieve the desired specifications of 100-200MS/s with 16-18 bit of resolution. The main architecture is based on paralleled structures in order to achieve high sampling rate and at the same time high resolution. In order to solve problems related to Time-interleaved architectures, an advanced randomization method was introduced. It combines randomization and spectral shaping of mismatches. With a simple low-pass filter the method can, compared to conventional randomization algorithms, improve the SFDR as well as the SINAD. The main advantage of this technique over previous ones is that, because the algorithm only need that ADCs are ordered basing on their time mismatches, the absolute accuracy of the mismatch identification method does not matter and, therefore, the requirements on the timing mismatch identification are very low. In addition to that, this correction system uses very simple algorithms able to correct not only for time but also for gain and offset mismatches.