Making an Easy Fix in the Field
The GE Lighting Euclid Lamp Plant manufactures a variety of industrial,
household, and automotive lamps. In the production of automotive lamps,
the bulb, exhaust tube, and mount structure—consisting of lead wires and
coil—are sealed together by heating the bulb and mechanically pinching the
glass to form a seal around the lead wires and exhaust tube. This process
results in great residual stress on the bulb. To minimize the effect, the
bulb is heated above a critical temperature and then cooled in a controlled
fashion to relieve the mechanical stress, improve product strength, and
meet quality parameters.
GE had to monitor each lamp in a temperature range between 650°C
and 950°C. High levels of signal attenuation caused by environmental
factors and mechanical obstructions led GE to choose an IR ratio-thermometer
with peak-hold signal conditioning. After testing the device in the lab,
GE purchased a Raytek Marathon MR1S integrated ratio-thermometer. Once the
unit was installed on the production line, though, things worked differently.
Although the temperature values being reported were mostly correct, an occassional
value would be reported that could not be explained based on the process
parameters.
Guido Bergomi, quality engineer at the plant, immediately contacted Raytek
in Santa Cruz, California, to notify the company of the problem. Bergomi
was able to connect the sensor to a computer and record process data while
the unit was online. Thomas Heinke, design engineer at Raytek, analyzed
the data Bergomi provided.
"The correct temperature readings were in the raw data," says
Heinke. "However, the speed of the process, the momentary blockage
of the reading, and the highly attenuated signal could all combine to cause
the peak-hold algorithm to trigger on an invalid signal." Based on
the data from GE, Heinke was able to develop a custom peak-hold algorithm
for the sensor and send a firmware upgrade. The upgrade was uploaded via
two-way communications to the firmware in the Marathon MR1S's electronics.
"I thought the units would have to go back to the factory,"
says Bergomi. "I was surprised and pleased that the modifications to
the firmware could be made in the field. I just hooked the unit to my computer
and ran the program.
|
Return to Main Article
