The construction can vary according to your needs and supplies. In general, you will need to connect to an air hose. A standard 1/4" NPT pneumatic air compressor hose male quick connect fitting connector plug can be used. An adaptor or two will expand this to a 1.5-inch pipe fitting. We first expanded to a 0.5-inch outer diameter pipe, which connected to a T connector so that an Omega reference gauge and our gauge were in parallel. Then we expanded the 0.5-inch pipe to the 1.5-inch adaptor. Our unit is about eight inches long, but it can be made much shorter by using a shorter pipe and not using the T connector.
Inexpensive pressure gauge.
The hardest part will be to weld the brass plate to the adaptor. Make sure that this is done professionally, because having a brass plate blow of the end of a pipe could cause serious injury. Then glue a strain gauge to the center of the brass plate. We used general-purpose Omega SG-7/350-LY41 gauges (a package of 10 costs $65.00 at the time of this writing).
The following illustration shows the pressure strain gauge in a quarter bridge circuit.
Quarter bridge circuit.
We tested our gauge by connecting it in parallel to an Omega PX236-150GV pressure gauge. The quarter bridge of our pressure gauge is connected to Channel 0 (pins 68 and 34) of our NI data acquisition system, while the Omega gauge is connected to Channel 1 (pins 33 and 66). Pin 67 of the data acquisition board is connected to the negative terminal of the power supply.
The Omega gauge sensitivity with a five-volt power supply is 5 PSI/mV. An NI LabVIEW program was developed to record data every 1 PSI (0.2 mV) of the Omega gauge reference -- first with the pressure increasing from 0 PSI to 40 PSI, then decreasing from 40 PSI down to 0 PSI (to obtain hysteresis information), and then increasing again to 40 PSI (to get repeatability information). We limited the pressure to 40 PSI to prevent serious injury of accidental blow outs caused by higher pressures. (Click here for a compressed ZIP file containing this LabVIEW project.)
The program computed the best third-order polynomial curve fit for each section of data to eliminate problems caused by not taking data at exactly the same pressure for each section. In order to give the students experience with other methods of analysis, these equations were copied to Excel for further comparison.
Using the third-order polynomial equations, as well as a linear trend line, we can develop a calibration graph and residual error curves as illustrated in the following images.
As we can see, this gauge, chosen at random, has a linearity of about 9%, which is fairly typical in the batch of gauges that we built. A small matter of calibration should bring the linearity to within acceptable limits. The hysteresis and the repeatability of this unit are about 1.6%, which is also typical for our units.
Les Hammer is an adjunct instructor at the Colorado School of Mines.