# How Accurate Is My Water Level Sensor?

Accuracy on a level sensor is a funny thing. What the specifications are really telling you is the error band as a percentage of a specific scale.

Level Measurement

Accuracy on a level sensor is a funny thing. What the specifications are really telling you is the error band as a percentage of a specific scale. To decode accuracy specs, we’re going to dive into some different accuracy ratings and explain them in detail.

We’ll use a water level sensor in our examples, just to keep it simple. The same principles apply to all level sensors, no matter what you’re measuring. We just want to remove any potential environmental factors to better explain how accuracy works; all things being equal (like temperature, for example).

There are a few ways to spec accuracy on a level sensor. Most depend on a measurement scale. Some of the more common methods include:

• Percentage of the distance of the current reading
• Plus or minus an actual unit, such as ± X mm
• Percentage of the full-scale as calibrated

Finally, once you understand how linear accuracy on your liquid level sensor works, you’ll want to add context to it. How much volume is at stake within the specified error band? This is tricky with non-linear tanks and wells. More on this later.

Ultrasonic technology is a very common water level sensor whose accuracy depends on the current reading. Ultrasonic sensors use sound waves to detect the liquid level.

These liquid level sensors can have a resolution as high as 0.1 in. (2.5mm) and an accuracy of 0.25 percent of the detected range (current reading).

What this means is that an ultrasonic will have higher detection accuracies at close targets than at targets farther away.

Here is our example using water level:

An ultrasonic water level sensor reading a full-scale range of 144 inches will have an accuracy of ±0.36 inches (at ambient temperature). The same sensor reading a distance of 75 inches will have an accuracy of ±0.18 inches.

The .25 percent detected accuracy is applicable whether the sensor is reading in distance (i.e.; in., ft. etc…) or volume (gallons).

Plus/Minus Actual Units
In some cases, accuracy is best understood as an error band of a solid unit of measure, such as ±1 inch.

Magnetostrictive technology is very accurate, and fits this scenario. Here, floats move up and down a stem. Changes in the magnetic field around the stem (caused by magnets in the float) are used to detect the location of the float, or in other words, the liquid level.

Magnetostrictive sensors have accuracy better than ±0.059 percent of the measurable length of the stem. Because the length of the stem depends on what the customer needs, the spec sheet will give a worst-case scenario–something like ±4 millimeters.

In reality, a shorter stem has better linear accuracy than a longer one.

Here is our water level sensor example:

A magnetostrictive water level sensor reading a measureable range of 6000 millimeters (its measureable stem length), will have an accuracy of ±3.5 mm.

A similar magnetostrictive water level sensor reading a measureable range of 3000 millimeters (its measureable stem length), will have an accuracy of ±1.77 mm.

Percentage of Full-Scale
It is very common to use pressure sensors to determine a liquid level. These sensors typically spec accuracy as a fixed percentage of the full calibrated scale.

The analog output of the pressure transducer will be calibrated to the full-scale of the measurement. If the output is 4-20mA, then 5mA equals 1/16th of the full range.

Many pressure sensors are fairly accurate, with an error band of ±0.25% full-scale or better.

Again, accuracy improves as the sensor’s full-scale pressure is lowered.

Now for our water level sensor example:
Let’s say you have flat-bottomed tank that is 1000 inches tall. A pressure sensor calibrated for 1,000 inches of water full-scale will have an accuracy of ±2.5 inches.

If you were to use this same water level sensor to measure a tank only 30 inches tall, the accuracy would still be ±2.5 inches.

If you used a water level sensor calibrated for a tank 30 inches tall, the accuracy would be ±0.075 inches.

When using pressure sensors to determine water level, it is highly recommended that you use a sensor with a full-scale range equal to (or close to) the height of your tank.

Tank and Well Dimensions
There are a variety of tank and well types and shapes out there. The dimensions of your tank or well are very important to put the accuracy of your level sensor into perspective.

Flat bottom tanks or wells with straight sides are the easiest to calculate accuracy and capacity.

Irregular shaped tanks and wells are more difficult, but there are tools to help you.

Some higher quality ultrasonic level sensors have different tank types programmed into the software. You will need to add the requested dimensions and the software will do all the calculations.

If your sensor does not have that capability, the use of a strapping chart for the particular tank in question is highly recommended. Strapping charts simply tell you what the volume is at any given linear height. Again, higher quality level sensors exist with strapping chart functions programmed into the software.

Having tank dimension calculators and strapping charts help you put a volume to your ± accuracy specification. If your tank is wide enough, even 4mm can represent a rather large volume.

You need to ask yourself how much volume you’re willing to allow inside the error band of your level sensor. The higher the accuracy, the higher the cost of your level sensor. Keep this in mind as you search for an acceptable solution.

And while you’re at it, let us know if you run into any questions about level sensors! We consider ourselves to be fairly educated in the matter, and our responsive techies and engineers are always ready to provide the right answer.

Kelvin Findlay is an Industry Analyst at Automation Products Group, Inc. He has been involved in the modern Internet-based monitoring market since its inception. He has played a key role for APG in market development and the analysis of user adoption and competitor activity for the last three years.

This post originally appeared on Automation Products Group's blog. To visit the APG blog, click here.

More in Flow Measurement