The effect of clock and carrier frequency offsets on the performance of a direct-sequence spread-spectrum multiple-access system

It is widely known that the probability of symbol error in a direct-sequence spread-spectrum multiple-access network is dependent on the carrier and spreading code chip phases of the interfering terminals with respect to the signal of interest. As a result, for a given total interference power, the performance is worse for the case of one strong interferer when compared to the case of many weak interferers where the Gaussian assumption on the multiple-access interference holds. The author assumes a random-spreading code model and shows that if there is an offset on the carrier and spreading code frequencies of the interfering signals relative to the signal of interest, the above difference in performance between the cases of a set of balanced and a set of unbalanced interferers is drastically reduced. The result is that subject to such frequency offsets one may make the Gaussian assumption on the interference power regardless of whether it is composed of one strong interferer or many weak interferers of the same total power. In a network it may be desirable to actually force the offsets by putting a slight offset on the clock oscillators of the various transmitters. It is shown that to achieve a small variation of the noise variance, the spreading code and carrier offset frequencies should be greater than the symbol rate