This paper summarizes experimental work at Bell Laboratories in the field of multipath propagation as it affects digital radio. Field and laboratory measurements are presented on commercial-quality high-speed digital radio systems working in the 4, 6, and 11 GHz common carrier bands. Circuits and equipment used in these systems are all state of the art, but no attempt has been made to describe them in any detail. Modulation types are offset-4 PSK, conventional 8 PSK, and 16 QAM. Outages(BER > 10^{-3})vastly exceeding accepted objectives are measured on unprotected radio hops. A very effective method for reducing outages has been the use of a space diversity combiner followed by an amplitude slope equalizer. Results from single, as well as dual-polarized digital radio relay systems, are presented. The paper also describes the laboratory measurement of an equipment "signature" as a means to compare different digital radio designs in their sensitivity to multipath fading.

An experimental study of multipath fading effects on an 8-PSK 6 GHz digital radio system was conducted on a 26.4 mi test link near Atlanta, Georgia. Results were obtained for a non-diversity arrangement and a space-diversity arrangement employing a cophasing combiner operating at 70 MHz and fed by two antennas separated by 30 ft. Multipath fading was found to have a significant impact on bit error rate performance, i.e., the digital radio system was fragile by comparison to conventional analog FM radio. Modest in-band linear amplitude dispersion, 0.2 dB/MHz for example, was sufficient to cause aBER\geq10^{-3}. Such dispersion occurred at (average power) fade depths as shallow as 20 dB. However, average fade depth was a poor indicator of BER performance. The measured hop missed short haul outage objectives by an order of magnitude without space-diversity, and was close to acceptable when diversity was activated.