Reports

Contents

Title: Pressure transducer comparison 12-2018
Date:2018-12-10
Data File: CS451test_20181210_air.csv
CS451test_20181210_water.csv
Refers to:sn 70011234, sn70011229

I wanted to compare the water level pressure transducers in the lab because we had been noticing large diurnal changes of about 5-6cm in the Mayberry water levels, which do not seem to be real. This sensor (sn 70011229) was installed on 2016-06-24 and we saw the diurnal fluctations beginning summer 2017. We swapped the sensor out 2018-07-12 and brought it back to the lab to test with another water level sensor that had been removed from SW for frequent NANs. I wired the sensors to a lab CR6, measured every 10 seconds, and stored the 1-minute averages.

I used the following equation to convert the psi readings into cm.

psi * 70.307 = cm

15592556818829.png

 

Figure 1. Back in the lab, I first tested it in air just by laying it on the lab bench. The pressure varied around 0, but there wasn't much diurnal fluctuation, <0.1 cm.

Figure 2. I also tested another sensor at the same time, laying them next to each other on the lab bench. Similarly, there wasn't much diurnal fluctuation  (<0.1 cm), but the pressure was a little offset, varying around 0.96 cm rather than 0 cm as expected.

Regression Data

Residuals

Figure 4. The linear regression between the two shows the noise in the pressure reading.

Figure 3. The temperature of both sensors matched well with each other, one was ~0.1 degC offset from the other, especially near the daily max temperatures.

Figure 5. After a week on the bench, the sensors were placed in a bucket of water. There is much less noise in the pressure readings, and one sensor is still about 1 cm offset from the other. I forgot to measure the actual depth of water, but I'm assuming the sensor reading 0 in air (sn 70011229) is more accurate.

Regression Data

Residuals

Figure 6. The linear regression shows a much tighter relationship between the two sensors when they are measuring pressure underwater. There is some hysteresis at the upper end.

Figure 7. Under water, the temperatures between the two sensors match well when they are underwater, with a minuscule 0.1 degC offset.

Conclusion

sn 70011229: I didn't see large diurnal fluctuations in either air or water, so the fluctuations we saw at MB were probably not caused by the sensor. In addition, the replacement sensor now at MB (sn 70011236) has similar fluctuations up to 6 cm.

sn 70011234: Seems ok, didn't see NANs in data, but probably has an offset of about +0.975 cm.

Â