No abstract

associated electrical and electro-optical hardware, the cost of redesigning the system architecture, and, in some instances, the difficulty in digitizing certain modulation waveforms. This report is written primarily for the fiber-optic design engineer that uses analog modulation, for whatever justification, but many aspects will also be of use for those who design with the traditional digital approach. Based on the LEDs we have tested, third-order linear dynamic ranges up to 75 dB and second-order linear dynamic ranges up to 59 dB have been observed in 3 kHz noise bandwidth at frequencies up to 20 MHz. For links with substantial loss or for frequencies much greater than 20 MHz the performance will be degraded. For the lasers tested, third-order linear dynamic ranges up to 68 dB and secondorder linear dynamic ranges in the vicinity of 50 to 55 dB have been observed in 3 kHz bandwidth at a modulation frequency of 70 MHz. Measurements at 400 MHz indicated a third-order linear dynamic range of 59 dB. Link lengths were typically 500 to 600 meters, but some measurements with a short link indicate that laser performance degrades for link lengths of several hundred meters because of reflections into the laser cavity from distributed regions of the fiber. Factors that affect fiber-optic system performance are discussed. These include electro-optical transfer characteristics, average power level, modulation sensitivity, device risetime, fiber dispersion, excess microbending loss, excess noise and distortion mechanisms for lasers, temperature effects, and lifetime, particularly with respect to an BMP environment. Techniques to make the best use of available linear dynamic range are discussed. These include using an intermediate FM carrier, using optical rather than electrical summing of FDM carriers, and hard-limiting of angle modulated carriers (followed by bandpass filtering) prior to introduction to the optical link.

No abstract

A 12-bit A/D converter having an aperture uncertainty of 25 psec and a maximum sampling rate of 5 MHz was used to experimentally determine the magnitude of the spurious-free dynamic range that could be obtained with present off-the-shelf technology. A larger number of quantization bits would have resulted m unacceptably slow speed and an unacceptably low maximum frequency limitation. A smaller number of bits would have yielded a smaller spurious-free dynamic range. Since we did not have a digital processing facility to analyze the A/D directly, a D/A converter, deglitcher (sample-and-hold), and a high quality spectrum analyzer were used to evaluate the A/D performance. Consequently, the results reported are characteristic of the composite A/D-D/A-deglitcher system rather than the A/D alone. These results, however, are directly applicable to a first generation towed communication buoy system which would probably use the' composite A/D-D/A-deglitcher architecture.

No abstract