Prediction of wave solder machine parameters based on printed circuit board design characteristics /

This thesis shows the prediction models developed using board

Bibliographic Details
Main Author: Johnson, Ronald Charles, 1959-
Format: Thesis Book
Language:English
Published: [Place of publication not identified] : [publisher not identified] ; 1994.
Subjects:
Description
Summary:This thesis shows the prediction models developed using board
design characteristics to predict key wave soldering process
parameters. Circuit boards that are to be wave soldered go
through four major steps: fluxing, preheating, soldering,,
and cleaning. Establishing process settings for a new
circuit board requires using several production boards and
running the machine multiple times until there is a minimum
of solder defects. The use of these predictive models will
help reduce experimental testing on production circuit boards
for optimizing machine settings in the soldering process.
Models developed for the key wave soldering process
parameters included conveyor speed and preheat temperature.
The modeling research involved experimentation with several
methods: multivariate regression, multiple linear regression,
regression for matrix combination of both parameters, and
thermal radiative equations. The model building data set
used in the research has process data and design
characteristics for about 200 board designs. The process
data comes from developing optimized empirical machine
settings on the tlectrovert's Ultra Pak 337 wave soldering
machine at Westinghouse.
The final model results started from 26 initial design
variables identified by process engineers at Westinghouse.
Conveyor speed modells design variables included: material
(heatsink, laminate, B-Stage) thicknesses, laminate type,
component population (total components) and volume (multi-
hybrid packages, mechanical parts, axial components) per
bareboard area, and mass per bareboar d area of the circuit
board. The preheat temperature model variables are: conveyor
speed, bareboard thickness, lead population per bareboard
area, component (mechanical and connector) volume per
bareboard area,, weight per bareboard area and density of the
circuit card. The preheat model includes conveyor speed to
account for the rate of heat absorption as the board passes
through the preheat zone. The models were validated using
data from new board designs. The completed models, conveyor
speed and preheat temperature, were used to predict values
for the process parameters of the new board designs and then
compared to the actuals. Predictions for the process
parameters averaged within 2% or better of their actuals.
Statistics for model building and validation data sets were
compared and found to be close to each other.
Item Description:Vita.
"Major subject: Industrial Engineering".
Physical Description:xvii, 177 leaves : illustrations ; 28 cm.
Bibliography:Includes bibliographical references.