Industrial Valve Position Sensor Feasibility Study
In the summer of 1997, I was asked to bid on the design and
prototype of a small board that incorporated a Hall Effect sensor,
an amplifier, and a 4-20 mA current loop interface. I spoke with the
prospective customer, an electrical engineer at a company that
designs and manufactures industrial valves, concerning the
application. He explained he and his mechanical engineer colleague
were developing a sensor to measure the position of the valve shaft
relative to the closed and open position with a repeatability of 0.1
degree. They planned to embed two magnets in the bottom of the valve
shaft to create a magnetic field that would rotate as the valve
changed position. This would allow them to place the sensor board
outside of the active valve volume (viz., the space in which the
material flows) on the "dry" side of the valve cap. Because the
valves are constructed of steel, a good conductor of heat, the
sensor, by virtue of its proximity to the material, must withstand
the temperature of the material flowing through the valve. This
temperature range, -40ºC to +150ºC, includes all of the military
range and then some, to a point where readily available electronics
do not even operate.
It did not take long to determine that a Hall Effect device would
not do the job. The thermal stress induced by the package as the
temperature varies over such a large range would cause the sensor's
output to vary randomly causing large errors. Hall Effect devices
also require relatively large field strengths and operating power,
two significant disadvantages.
I spoke with a colleague and friend, an industrial physicist,
about the problem. He suggested magneto-resistive devices, which use
a material that changes its electrical resistance based on the
strength of the magnetic field around it. I found a company, NVE
Corporation, right here in the Twin Cities, that manufactures sensor
devices that incorporate the Giant-Magneto-Resistive (GMR) effect.
These devices were a perfect match for the valve application. They
behave predictably as temperature varies, they work at high
temperatures, they are small, and they require very little power to
operate. Finally, they are inexpensive. I obtained some samples
immediately.
Then the real innovation occurred. My colleague explained how we
could use two sensors in a way that would cancel the variation over
temperature. The method required both measurement and calculation,
not a problem today thanks to small and inexpensive
microcontrollers. I explained my preliminary findings to my client,
and was contracted to perform a feasibility study of the dual-sensor
idea.
I used a stepping motor and collected data manually for both a 1-
and 2- sensor system. I entered the data into an Excel spreadsheet,
and performed the calibration and sensor calculations over a
complete 90º angular range. The results exceeded my expectations,
and I wrote a final report describing my experiments, my findings,
and my recommendations. The customer liked the idea, and hired me to
develop the product. But that is another story you can read by
clicking
Sensor Development.
