Delay variations in some GPS timing receivers

We report a number of studies of variations in the delay through GPS receivers used for timing at NIST. As a first study, the data from the receivers built at NIST in the early 1980´s (called the NBS-type receiver) are compared over periods of many years to show variations in delay. These receivers vary of the order of 5 ns over a year, with the peak-to-peak variation of 10 ns over 6 years. A second study shows the temperature dependence of the antenna system and of the receiver unit of the NBS-type receiver. The antenna system of one receiver was held constant at 42°C for the entire experiment. The antenna of another system was held at various fixed temperatures, and the relative delay through the two antennas was determined at each temperature. Similarly we varied the receiver temperature of one system, holding the antenna temperature constant. Though the delay varied 2 ns over the course of the experiment, this variation was not correlated with temperature. We next report the delay variation with temperature of a commercial multi-channel GPS receiver, which we are investigating for use in an alternative common-view receiver. We studied the temperature dependance of two receivers separate from the antenna. In one unit, we found a 2.3 ns a peak-to-peak variation of delay with temperature and a 4.3 ns peak-to-peak variation in the other, both over a range from 10°C to 55°C. The delay variation in both appears to be approximately quadratic over some range of temperature