In this paper, the encoding/decoding circuitry for various t-unidirectional and burst uni-directional error detecting codes (UEDCs) is studied. Two new realizations of an encoder for all these codes are given. One approach assumes the use of carry-save adders to build a modular counter of 1’s (or 0’s) which is a basic element of an encoder. The new counter is proved to be faster than a commonly used counter built of ripple-carry adders. The second approach is based on the use of a multi-output threshold circuit Tn. With Tn-based encoders the design is easy for any known UEDC. Either encoder proposed here enjoys many advantages which are important for a VLSI implementation, such as: a highly regular structure, easy testability, the least amount of hardware used, and the smallest delay as compared to other similar circuits. Finally, it is shown that an efficient self-testing checker for any UEDC can be built by using the new encoders given here.