Dept. of Science & Engineering
Oregon Health & Science University
In this research we have developed and analyzed a set of techniques for converting analog voltages into pulse streams based upon the Inter-Pulse-Interval (IPI) as well as conversion back to analog voltage. A MOSIS 0.35 Âµm chip was designed and fabricated that contains both circuits. The basic chip design and test results are also presented. The IPI presented here is the asynchronous version, without a clock to drive the sampling process. We have studied low power versions of this circuit. We show that the IPI conversion requires significantly less power than more traditional signaling techniques such as pulse-width modulation (PWM), sigma-delta ADC, etc. This low power is obtained without slowing the sampling rate of the input signal. We show that the IPI conversion process generates intrinsic distortion between the voltage input and voltage output, and analytically prove that the way to compensate for this distortion is to increase the sampling rate. In asynchronous IPI, there is non-uniformity in sampling rate from one part of the input signal to another part of the input signal. For reducing distortion, the main thing that is relevant is the ratio of the worst-case or slowest sampling rate over the input signal frequency, based on Nyquist's theorem. The IPI output is suitable for transmission using broad-spectrum, pulse based ultra-wide band (UWB) techniques, with the potential application in the area of ultra-low power wireless sensors, especially for biomedical applications. We also show that IPI representation has more immunity to high-frequency attenuation over an interconnect wire compared to more conventional signaling techniques, such as pulse-amplitude modulation (PAM), pulse-width modulation (PWM), etc. This concept can be useful to address high-speed signal integrity issues.
Div. of Biomedical Computer Science
School of Medicine
Ravi, Sanjay, "Inter-pulse interval based mixed signal representations" (2008). Scholar Archive. 336.