Fuel level sensors are popular tools for fuel consumption monitoring it the GPS tracking industry. Very important stages of its installation – are sensor and tank calibrations. In this article, we have collected the main tips that can help to avoid calibration errors and reduce time to install the units.

Let’s begin with the difference between sensor calibration and tank calibration. Essentially, fuel level sensors measure level, not volume. However, customer most likely wants to see the fuel volume in reports. To calculate the fuel volume we have to know the tank shape. In fact – fuel tank calibration is the calculation of shape and formula to convert fuel level into fuel volume. It could be done via filling the tank with small portions of fuel and recording the sensor output. Tank calibration by fillings is a time-consuming procedure and customers may try to look into sensors, which do not require tank calibration. Nevertheless, it is not possible, as the tank calibration procedure refers to the tank and not to the sensor.

Most popular fuel level sensors for GPS tracking have RS232/RS485 output interface. The sensor comes from the factory pre-calibrated to default length and the sensor output signal could change from “0” (empty) to “1023 (or 4095)” (full). After cutting this sensor – the output range shifts down and while decreased maximum output signal is not a problem, the empty tank output value cannot decrease below zero. As a result, the blind zone appears, proportional to the length of the sensor cut. Same problem with sensors with analog voltage output, as after cutting and output range shifting sensor should send negative voltage to the empty tank and are unable to do that.

Thus, after cutting one should calibrate the sensor to its new length by immersion in fuel by 0 and 100% (“empty” and “full” calibration).

Therefore, one has to calibrate the sensor after the cut and then perform a tank calibration. Now let’s think about simplifying the process.

Someone can notice that during the tank calibration we fill the tank from empty to full level. Sensor calibration “empty”-“full” is not required if sensor configuration software can calibrate the sensor itself based on tank calibration data.

For example, it is possible to calibrate the tank using Eurosens Dominator sensor configurator and the first and last points of tank calibration would be sensor calibration points at the same time (Fig. 1). Therefore, one can save about 5 minutes of the sensor calibration procedure.

Fig. 1. Tank calibration using Eurosens Dominator configurator 

Next step is tank calibration. How can one avoid this procedure?

If the customer will be satisfied with fuel height in mm or percentage, you can skip the tank calibration. However, everybody wants liters or gallons. The only question – how to perform the tank calibration.

How does the shape of the tank affect the calibration?

For tanks in which the cross-sectional width is constant in height, two points are sufficient – full and empty. Examples of such shapes: a perfectly rectangular tank (without roundings) or a vertical cylinder (not horizontal!).

Fig. 2. Ideal tank shapes, which don’t require the tank calibration

  The tank calibration results can be plotted on a graph. For the tanks shown in Fig. 2, the graph will represent a straight line.

Fig. 3. It makes no sense to calibrate the entire tank since all points fall on one straight line

However, such tanks cannot be found in real environments. Real tanks have rounded, dents (Fig. 4).

Fig. 4. Most of the “rectangular” tanks have roundings

What needs to be taken into account that the real calibration curve is not a straight line (see Fig. 5).

Fig. 5. Graphs of the calibration tables for rectangular tanks with roundings and irregularly shaped tanks

The graph shows that in the part of the tank where the walls are vertical and not curved the points of the calibration curve still on the straight line, and the graph is curved in the places where the tank is rounded. Thus, one could say that to calibrate rectangular tanks with roundings, you can use a different step of filling portions – fill in smaller portions in the area of the curvatures and larger ones in the middle of the tank, where the walls are vertical. This will save you time while maintaining the accuracy of the calibration.

Hence the answer to the question – what will happen if the tank is calibrated only at 2 points – full and empty? The error will increase depending on the size of the roundings, with the error being higher when the fuel is inside the rounding zone.

But for tanks of complex shape (Fig. 6) calibration with a constant step is required, since there are no straight sections of the graph, as can be seen from Fig. 5. The same applies to the shape of the “horizontal cylinder” tank.

Please note that plotting calibration tables allows you to verify the calibration. Calibration defects are clearly visible on the graph even before the customer complaints about the inaccurate operation of the fuel consumption control system. On the calibration graph, vertical or horizontal lines state about calibration errors (Fig. 6).

Fig. 6. Tank calibration error

For partners’ benefits, Mechatronics developed simple software that compares calibration files with each other by building plots in Excel.

There is absolutely no way to calibrate by filling in portions. What can be done then?

It happens, especially during installation on large storage tanks. Filling tens of cubic meters of fuel in portions is a very complicated procedure. In this case, there are several solutions.

Calculate the calibration table based on the tank dimensions. For horizontal cylinder tanks, you can use an online calculator, which will generate a calibration file for the Eurosens Dominator level sensor.

Fig. 7. Tank calibration calculator

This method does not take into account the existing shape defects and elements of the tank, which can affect the volume and calibration.

Build a calibration table based on a three-dimensional model. 

It is used when the accuracy is required higher than in the previous method, but the tank cannot be calibrated with fuel. Then one could build a three-dimensional model of the tank based on the dimensions one sent and adds the inclination of them if needed. The model is used to calculate the tank calibration.

Question : I have 10 identical cars. Is it possible to calibrate the tank only once, and then copy it everywhere?

That’s quite possible but it is necessary to check a few things. Individuality is possessed by both outwardly similar-looking fuel tanks and sensors, each of which is slightly different in its measuring electrodes (not to mention the fact that they can be cut in different ways). Often it is possible to selectively check the “uniformity” of the tanks by filling several tanks to full and thus measuring their total volume.

After that, one will have to carry out a full-empty calibration for all sensors so that the sensor output signal during calibration does not depend on the individuality of its electrode. After that, calibrate one tank and use it for the remaining tanks of equal volume and shape.

If the sensors are different, what should I do when replacing the sensor?

The individuality of the sensors mainly lies in its electrodes. If one replaces the Eurosens Dominator modular sensor, then the electrodes from the replaced sensor can be used. In this case, it is enough to check the internal value of the sensor’s detector after assembly with old electrodes – if it matches in the “empty” state with the value in the calibration table from the old sensor, then there are no obstacles to using the old calibration.


Artem Smirnov, [email protected]