Liquid level tank measurement with the DrDAQ - PART 2

Glovisol · Post by **Glovisol** » Sat Oct 24, 2015 9:51 am

: Liquid column calibration setup

SIMPLE MATHEMATICS

As already anticipated in PART 1, the differential sensor is ratiometric and it exibits an offset in the output (with zero differential pressure) in the order of 200 mV with a supply voltage of 5,00 V. Picolog measures the output voltage with one of the Ext. channels (Channel 1, in our example) and then performs the following operations in the calculated parameters channels.

In EXT. 1 (Channel 1 input) : Sensor Output Voltage
Parameter formatting: mV
Digits width:3
Decimal positions:2
Min. Value:0
Max Value:600
Scaling equation is: X*1000

Calculated Parameter Channel 2: Adjusted Sensor Voltage
Note: here we subtract the offset output voltage from the Sensor Output. This voltage is shown in EXT. 1 above when the sensor is still not connected to any pipe and has zero pressure differential. With my sensor the offset voltage is 202 mV, hence:

A=Sensor Output Voltage
Equation: A - 202
Parameter formatting: mV
Digits width:3
Decimal positions:2
Min. Value:0
Max Value:300

To calibrate the system, we use a water column of 40 cm (we would use diesel fuel, if we had to measure a fuel tank). With the assembled system we pump into the 40 cm water column and Channel 2 shows a value of 184 mV, hence the sensor sensitivity (with water, it would be different with a liquid of different density) is 4.6 mV/cm. Therefore:

Calculated Parameter Channel 3: Column Height

B=Adjusted Sensor Voltage
Equation: B/4.6
Parameter formatting: mV
Digits width:3
Decimal positions:2
Min. Value:0
Max Value:99

Finally we have a tank with a height of 50 cm and a capacity of 100 litres, therefore:

Calculated Parameter Channel 4: Tank content

C=Column Height
Equation: C*100/50
Parameter formatting: litres
Digits width:3
Decimal positions:2
Min. Value:0
Max Value:100
NOTE: in the enclosed files & pictures, the Min. & Max values are respectively 70 & 90 litres, in order to show retracing capability and data loss with time due to air leaks.

With these values a column height of 40 cm corresponds to 80 litres.

Of course one could concentrate all calculations in EXT. 1 and do away with the calculated parameter channels, but I think it is useful to keep all parameters separately under control.

THE RATIOMETRIC PROBLEM

Because of the ratiometric nature of the sensor, variations in the supply voltage of the sensor introduce a reading error as follows.

Supply...........Offset...........Error............Corrected & Linearised Error
V..................mV..............Litres............Litres

5,5................220..............7,74..............7,7
5,4................216..............6,02..............6,2
5,3................213..............4,73..............4,6
5,2................209..............3,01..............3,1
5,1................206..............1,72..............1,6
5..................202...............0,00..............0,1
4,9................199..............-1,29............-1,4
4,8................195..............-3,01............-3,0
4,7................192..............-4,3.............-4,5
4,6................188..............-6,02............-6,0
4,5................184..............-7,74............-7,5

Considering that the USB + 5V can vary in the range 4.75 to 5.25 V, We could have a random reading error in excess of +/- 3 litres, which might not be acceptable in some instances. The solution is to use a precision + 5 V supply to feed the sensor, which needs approx. 6 mA. To do this the jumper between pins 11 & 12 of M1 (DrDAQ Buffer) must be removed and the stable + 5 V applied to Pin 12.