Theory and Performance of Narrow Correlator Spacing in a GPS Receiver

Historically, conventional GPS receivers have used 1.0 chip early-late correlator spacing in the implementation of delay lock loops (DLLs), However, there are distinct advantages to narrowing this spacing, especially in C/A-code tracking applications. These advantages are the reduction of tracking errors in the presence of both noise and multipath. The primary disadvantage is that a wider precorrelation bandwidth is required, coupled with higher sample rates and higher digital signal processing rates. However, with current CMOS technology, this is easily achievable and well worth the price. Noise reduction is achieved with narrower spacing because the noise components of the early and late signals are correlated and tend to cancel, provided that early and late processing are simultaneous (not dithered). Multipath effects are reduced because the DLL discriminator is less distorted by the delayed multipath signal. This paper presents the derivation of these narrow correlator spacing improvements, verified by simulated and tested performance. INTRODUCTION AND BACKGROUND The use of 1.0 (or even 2.01 chip correlator spacing in delay lock loops (DLLs) dates back to the early GPS receivers, starting with the Phase I GPS X-Set, Y-Set, Z-Set, and Manpack. In early theoretical papers, 1.0 chip spacing was implied (1,2]. Narrower spacing was mentioned in the classic multipath effects report [2], but not in a complimentary way. However, this dishonorable mention was with respect to coherent DLLs in the presence of strong multipath. For multipath signals of more reasonable levels, these adverse effects are not true, nor are they true for noncoherent DLLs. This has been verified [3], and is explained herein. The 1.0 chip spacing concept carried over to the GPS Phase II and Phase III equipment, and more recently to modern digital GPS receivers [4,5]. Although these cited receivers are of the P-code variety, this 1.0 chip spacing concept has also been used in most commercial C/A-code receivers. In fact, it has been such a standard approach that authors of papers describing their designs do not even mention the spacing, except that it is implied by their performance equations and test results.