Hi all,
I've seen some weird behaviour in my strain gauges recently, so I've setup a test where I axially load a specimen in tension with a known force. I want to measure the strain and estimate the corresponding load force to ensure that I have everything set up correctly. Should not be difficult, given sigma = E*strain, and sigma = Force / Area.
What I want:
Strain readings that are in the range of +/- 10% of theoretical strain (or preferably better). I do not accurately know the Young's modulus of the steel specimen, but I'm confident it is in the range 195-210 GPa.
What I get:
Strain readings that overestimate/overshoots theoretical strain by approx. 3 times in magnitude (say I load the specimen with a known force of 1000N, I measure a strain that correspond to 3000N). This varies somewhat, but around 3 seems to be rule of thumb.
Setup
My hardware setup is as follows:
- T rosette strain gauge connected to terminal block NI 9949
- Half-bridge type I setup (also tried quarter bridge type I, then with terminal block NI 9944). Pin out is checked and rechecked
- NI 9237 bridge module connected to NI 9949 through RJ50-cable.
- NI 9237 connected to cDAQ-9171 chassis (also tried cDAQ-9174).
- Tried another of everything (module, cable, strain gauge, specimen), but error still persist
-Loading: theoretically between 2-20 MPa of stress.
I have a simple VI utilizing DAQ Assistant to aquire the data. These are the parameters:
- Gauge factor specified as per calibration certificate of strain gauge manufacturer (k = 2.09)
- 2.5 Voltage excitation value (also tried 5V, 10V)
- Gauge resistance 120 ohm as per calibraiton certificate
- Poisson ratio 0.3
- Signal input range: -1mV to 1mV.
- Continous input with sampling rate 8192 Hz, reading 1kHz at a time.
- I'm using the strain calibration in DAQ Assistant with "enable offset nulling" and "enable shunt calibration" with a shunt resistance of 100 kOhm at resistor R3
- I've also used the example "Strain - Continous Input.vi". This reads the same strain values as DAQ Assistant.
Question
I cannot figure out what I'm doing wrong. My hypothesis is that there are some scaling errors in software that I'm missing. Have anyone else experienced something similar to this?
Side notes
- A voltage excitation value of 10V does not actually supply 10V, but more like 6-something V. 5V supplies 5 and 2.5V supplies 2.5, thus I've gone with the lowest to be sure that the chassis/module can supply enough voltage.
- When I do offset / shun calibration, I get an initial error of 42%. This seems very high to me. DAQ assistant / calibration module corrects the initial voltage by around 1.4 mV and the gain factor and after calibration the error is 0% - but still - 42%? The strain gauge is brand new.
- The null offset is in reality not at 0% error, the strain signal is howering around 0.6-1 MPa just after calibration when in an unloaded condition. Is this normal? It does not affect the measurements drastically and I can always offset that in post-processing, but still I thought it should be closer to 0.
