Aqua-Data RV4X Calibration

July 2020 Procedure

 

Each flow sensor is individually calibrated at a hydrodynamic towing tank at QinetiQ, Gosport, UK.

We use a specific RC3 measuring unit that is dedicated to the measurement of the sensor, and this has previously been calibrated against a fixed precision unit that generates electrical signals in direct proportion to the current excitation, so as to simulate an ideal sensor with a fixed flow of 3.64m/s. This precision calibration unit uses ultra-stable 0.01% precision parts for the critical components, and we check this annually using a calibrated current source and a verified voltmeter (the voltmeter is verified internally against a calibrated voltage source).

Having calibrated the RC3 display unit so it displays the ‘nominally’ correct speed (as per design criteria) then a number of measurements are made at various speeds, logged every 50 milliseconds from a calibrated calibration rig onto a laptop, and assigned against the unique serial number previously assigned to the flow sensor. The flow sensor is rigidly bolted to a heavy trolley that is moved by a regulated electric motor and gearing at a constant speed, parallel to the side of a 270 metre tank of depth 5.4 metres. The sensor is held at a depth of between 1 and 2 metres. The depth is not a calibrated value; it is nominal, so as to reduce side effects to a negligible level.

 

The trolley is moved at an exact speed, as the trolley has its own timer system which displays the velocity, calculated from an encoder wheel mounted to the trolley. We verify the encoder wheel and thus velocity measurement by comparing it to a calibrated Dimetix laser distance sensor and a precision steel rule. The ambient temperature is fairly constant at a nominal 20 C, and even allowing for worst case thermal effects on the steel rule, a length error of more than 0.03% cannot occur over a 20C temperature variation. Hence we have a high level of confidence that the velocity is extremely accurate.

The time spent at each speed is a minimum of 10 seconds, and a nominal time of 20 seconds. Readings are taken at various exact speeds, as follows:

0.00 m/s

0.25 m/s

0.50 m/s

0.75 m/s

1.00 m/s

1.20 m/s

1.50 m/s

1.80 m/s

2.00 m/s

2.30 m/s

2.80 m/s

3.30 m/s

3.80 m/s

These readings are converted into a table of 13 data points. These points in turn are used to create a piecewise linear curve that corrects the sensor readings to be correct at the given sample points, and linearly interpolates in between. The graph being interpolated is very nearly linear in any case, so the interpolation is very precise. For speeds from 3.8 m/s to 4.0 m/s the above data is extrapolated upwards.

Since the sensor contains no moving parts, and the sensitivity is a function only of the magnetic field generated (in turn only a function of the excitation current) and the hydrodynamic effects of the geometry, then provided the sensor is clean and no physical damage has occurred there is no mechanism whereby drift can occur, except in the drift of internal gain setting resistors potted inside the sensor.

 

These are all precise and high stability parts, and are guaranteed free from significant drift by the manufacturer. The combination of all the possible errors mentioned above is such that the precision can be guaranteed within its working range to +/- 1.0% of range, and +/- 0.03m/s over a temperature range of -20 C to + 60 C.

 

Special calibration is possible for units that are intended to be used in more extreme temperature environments.

Aqua Data RV2 and RV4 Calibration

November 2016 Procedure

 

Each flow sensor is calibrated at an 80m hydrological tank in Southampton Solent University or another suitable facility. 

We use a specific RC2 measuring unit that is dedicated to the measurement of the sensor, and this has previously been calibrated against a fixed precision unit that generates electrical signals in direct proportion to the current excitation, so as to simulate an ideal sensor with a fixed flow of 3.64m/s. This precision calibration unit uses ultra-stable 0.01% precision parts for the critical components, and we check this annually using a calibrated current source and a verified voltmeter (the voltmeter is verified internally against a calibrated voltage source).

Having calibrated the RC2 display unit so it displays the nominally correct speed (as per design criteria) then a number of measurements are made at various speeds, and these are logged every 0.5 seconds from the RC2 onto a disc, and assigned against the unique serial number previously assigned to the flow sensor. The flow sensor is rigidly bolted to a heavy trolley that is moved by a regulated electric motor and gearing at a constant speed, parallel to the side of an 80 metre tank of depth 2 metres. The sensor is held at a depth of 1 metre. The depth is not a calibrated value; it is nominal, so as to reduce side effects to a negligible level.

The trolley is moved at the desired nominal speed, and the actual speed of movement is measured by timing the travel between various fixed points at the side of the tank. For low speeds the separation of the fixed points is either 3m or 4m depending on the selected speed. For higher speeds the separation is 50 feet The Imperial (feet) measurements are used for historical reasons, and these are converted to precise metres using the correct scaling factors. The trolley has its own timer system which displays the velocity between two fixed sensors built into the tank structure.

The distance between the marker posts is fixed and confirmed with a precision steel rule; the ambient temperature is fairly constant at a nominal 20 C, and even allowing for worst case thermal effects on the steel rule, a length error of more than 0.03% cannot occur over a 20C temperature variation. The timing interval is measured by a precision timer/counter Fluke model 1953A, which is independently verified against a TTi counter model TF930 and a TTi pulse generator model TG4001. These instruments are not currently NPL calibrated, but are checked to agree with each other to within 1 ppm. Hence we have a high level of confidence that the timing is extremely accurate.

Readings are taken at various nominal speeds as follows (the number of readings at the higher speeds is limited by the tank length):

0.00 m/s
0.02 m/s

0.05 m/s

0.10 m/s

0.18 m/s

0.25 m/s
0.50 m/s
1.00 m/s

1.50 m/s
1.80 m/s
2.30 m/s
2.80 m/s

3.30 m/s

3.80 m/s

100 readings
200 readings
120 readings
60 readings
40 readings
100 readings
40 readings
24 readings
16 readings
14 readings
10 readings
10 readings

7 readings

6 readings

These readings are converted into a table of 14 data pairs of actual speed (as measured by the distance/time method outlined above) versus instrument displayed speed. These pairs in turn are used to create a piecewise linear curve that corrects the sensor readings to be correct at the given sample points, and linearly interpolates in between. The graph being interpolated is very nearly linear in any case, so the interpolation is very precise. For speeds from 3.8 m/s to 4.0 m/s the above data is extrapolated upwards. At higher speeds in which turbulence may occur, the unit may not give correct readings. It has been empirically verified that the flow appears non turbulent and linear extrapolation is appropriate for speeds up to 6m/s, with errors of less than 5% and typically less than 2%, provided the prove is mounted c orrectly and parallel to the stream lines.


If data is obtained in which the linearity of the resultant transfer function differs from the expected transfer function by more than 1%, the probe is recalibrated or may be rejected.

Since the sensor contains no moving parts, and the sensitivity is a function only of the magnetic field generated (in turn only a function of the excitation current) and the hydrodynamic effects of the geometry, then provided the sensor is clean and no physical damage has occurred there is no mechanism whereby drift can occur except in the drift of internal gain setting resistors potted inside the sensor. These are all precise and high stability parts, and are guaranteed free from significant
drift by the manufacturer. The combination of all the possible errors mentioned above is such that the precision of the electronics can be guaranteed within its working range to +/ 0.5 % of range, and 1 5mm/ s over a temperature range of 10 C to + 60 C. Special calibration is possible for units that are intended to be used in more extreme temperature environments.