Masters Theses

Date of Award

12-1988

Degree Type

Thesis

Degree Name

Master of Science

Major

Mechanical Engineering

Major Professor

Edward G. Keshock

Committee Members

T. W. Kerlin, G. V. Smith, C. Moore

Abstract

A non-contact velocity instrument was developed for fibers and similar materials. The instrument operated on a basis in which a short section of the fiber's length was heated using a high-intensity light source. Infrared (IR) radiation sensors detected the presence of this fiber section as it traveled between two points. A microcomputer monitored the sensor response and calculated the velocity.

The heat transfer principles involved in the velocity measurement included radiation from the lamp, absorption and conduction in the fiber, and convection and radiation heat losses. Fiber materials with low thermal conductivity were found to be best for this measurement technique. A one-dimensional heat transfer model of the fiber was developed to aid in studying fiber properties.

Various IR sensors were reviewed. Initially, a conventional IR thermometer was used. Later, pyroelectric sensors were selected for their lower cost, faster response and higher sensitivity. Their disadvantage was their susceptibility to shock and vibration.

Signal conditioning consisted of current-to-voltage conversion (if needed) , a differential amplifier and low-pass filter, which were developed from integrated circuits. Simple interfacing circuits were built to interpret "start" and "stop" times from the sensor response to the heated fiber. In later experiments, the conditioned sensor output was connected directly to an analog-to-digital converter.

A desk-top computer with a BASIC operating system was used to process sensor data and display the fiber velocity. In some experiments, machine language routines were used to increase processing speed. A portable version of the instrument was built with another student, Dilip Nawathe, incorporating an MC68000 microprocessor and LED display.

Three experimental versions of the non-contact fiber velocity instrument were tested. The first version proved the measurement principle and provided a means to evaluate major sources of error. A 200W tungsten lamp heated the fiber, while a chopper between the lamp and the fiber was used to pulse the radiation. A test fixture consisting of a variable-speed motor and two large pulleys transported a 4.7 mm black cable past the lamp and a single IR sensor. Difficulties arose in correlating the lamp illumination with the sensor response. Also, the time resolution of the microcomputer was a large source of measurement error.

In the second apparatus, the IR thermometer was replaced with two inexpensive pyroelectric detectors, eliminating the correlation problem. A slope sensing circuit detected signal peaks for start and stop times. Averaging and statistical error detection minimized the effects of errant pulses. An external clock and machine language routines reduced the time resolution to 1 ms. Measurements of cable velocities up to 90 cm/s showed standard deviations of about 10 percent of reading. The chief source of error was found to be the cable drive.

An enhanced version of the dual-sensor instrument used different lenses to focus the IR sensors on a small image. The peak detection circuit was eliminated, and the sensors were connected to an analog-to-digital interface. Sensor levels were then interpreted in software. Using a phonograph turntable to drive a 300 μm diameter nylon filament, velocities were measured from 48 to 74 cm/s, with standard deviations of 3 percent of reading. Standard deviations in travel time were from 3 to 5 ms, independent of velocity.

Feasibility tests were conducted for measuring the velocities of aluminum, paper and suspended powders (talc). Aluminum plate, 3.2 mm thick, and aluminum foil generated a poor IR sensor response. Paper, in contrast, had a very strong, but irregular response. Fine powder suspended in an air stream also gave negligible sensor response.

Methods for improving the velocity instrument's performance included replacing the IR sensors and lenses with higher quality components, averaging several readings and lengthening the distance between sensors. Such an instrument was projected to measure velocities of 50 μm fibers up to 7 m/s, with an accuracy of +/- 0.5 percent.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS