How to Choose a Capacitive Sensor for Liquid Level
For a high or low level signal through a plastic tank, start with an external capacitive switch. For a metal vessel or a continuous level reading, choose a different arrangement. The deciding test is whether the actual liquid and the drained container remain distinguishable—not simply whether the sensor detects water.
Do you need a level switch or a continuous reading?
Choose the required result first. A switch tells the controller whether liquid has reached one sensing zone. A continuous system reports level across a defined measuring span. Both may use capacitance, but they require different hardware and setup.
| Your application | First arrangement to evaluate | Key selection boundary |
|---|---|---|
| One level point, with no liquid contact | External capacitive point-level switch. | A compatible non-metal wall and repeatable empty-to-liquid contrast. |
| One level point in a metal or unsuitable vessel | A process-rated point-level probe installed through a suitable connection. | Wetted materials, seals, process conditions and the probe's reference arrangement. |
| Level trend, inventory or a proportional signal | A dedicated continuous capacitance system or another continuous level technology. | Measuring span, calibration, vessel geometry and required accuracy. |
Two switches can control a band, but do not measure the whole level
A lower and an upper switch can support refill control between two limits. They do not reveal the level between those limits. Likewise, an IO-Link connection does not by itself mean the device provides a calibrated liquid-height value; check which process data the exact model supplies.
A wetted probe needs its own compatibility review
Do not treat a probe as an external proximity switch with a longer tip. In continuous capacitance measurement, conductive-liquid and non-conductive-liquid arrangements can use different probe insulation, reference geometry and calibration. Endress+Hauser's principle guide explains this distinction. Follow the selected instrument's installation instructions rather than a generic grounding rule.
Can the sensor detect through your tank wall?
An external capacitive switch is a candidate through compatible plastic or glass. An ordinary external proximity sensor should not be selected to detect liquid through a continuous metal wall. Check the actual sensing location, including coatings and reinforcement—not only the tank's material name.
The sensor responds to a changing electric-field environment. The empty wall already contributes to its baseline; the liquid must create a distinguishable additional change. Wall thickness, curvature, sensor-to-wall gap and nearby hardware all form part of that installation.
Check the wall and mounting together
- Wall construction: record material and thickness variation, ribs, double walls, labels, conductive coatings and metal near the sensing area.
- Face position: follow the model's contact or clearance instructions. A curved tank may leave an air gap under a flat sensor; a fixture must hold that gap consistently.
- Production movement: check tank bulging, removable-container placement, vibration and changes after cleaning or maintenance.
- Surroundings: test with the final bracket, guards, cables and nearby equipment in place.
Do not calculate usable reach by subtracting wall thickness from catalog sensing distance. A rated distance obtained with a specified test target is not a guaranteed liquid-detection distance through a layered wall and air gap.
How to read a wall-thickness claim. Balluff describes selected Smart Level products operating through glass or plastic walls up to 10 mm, in the context of conductive media and that technology. This is a family-specific example—not a 10 mm allowance for any capacitive switch. Ask which exact model and liquid conditions support a supplier's claim. Balluff application reference.
For a metal tank, consider a suitable process connection or an engineered non-metal sight section, if permitted by the process design. The sight section must represent the tank's real level; trapped liquid, blockage or delayed equalization can defeat the measurement.
Will it detect your actual liquid—not just test water?
A successful water demonstration does not prove reliable oil, detergent or solvent detection. Many oils give less dielectric contrast than water-based liquids. Solvents vary widely; the category name alone is not a useful sensitivity specification.
Test the formulation and operating conditions
Identify the actual liquid, concentration range, temperature range and expected contamination. Then test the least favorable production conditions. A dielectric-constant table can help explain why signals differ, but it cannot predict the whole wall–gap–liquid arrangement.
Conductivity and dielectric permittivity are different properties. Their practical influence depends on the measurement design. Do not apply a minimum dielectric value from one probe family to an unrelated external sensor.
Decide what foam and residue should mean
Should foam count as “level reached,” or should the switch respond only to bulk liquid? After draining, must the controller report “below level” while a wet film remains? Define these states before teaching the sensor.
Compensation is model-specific. For example, ifm describes its LMC buildup-discrimination approach as frequency sweeping. It is not evidence that an ordinary adjustable capacitive proximity switch has the same capability. Ask for the sensing principle and the documented behavior, not simply an “anti-foam” label.
How do you find a reliable empty-to-full setting?
Use the manufacturer's setup procedure, then challenge the setting with both rising and falling level. A usable setting must detect liquid at the required point and release after the liquid falls away—including the normal wet or coated condition after draining.
Clean and empty
No bulk liquid at the sensing point.
Liquid present
Bulk liquid reaches the sensing zone.
Drained, with residue
No bulk liquid here—but the wall is not clean.
- Fix the geometry first. Use the final wall, bracket and sensor position. Perform setup with unintended machine movement or filling prevented; do not make uncontrolled adjustments during production.
- Establish the intended states. Follow the model's empty/full teach sequence or manual adjustment instructions. Use the real liquid and include the drained, wet-wall condition.
- Check both transitions. Fill slowly through the sensing zone, then drain through it. Record the liquid heights at which the output changes and releases. Hysteresis means those heights need not be identical.
- Repeat without retuning. Challenge wall variation, temperature, residue, remounting and the permitted operating conditions. A setting that works only immediately after cleaning is not enough if the normal process leaves coating.
If the sensor exposes a signal or margin value, log it alongside the output. If it only has an LED and an adjustment control, record the setting, actual level and switching results. Do not invent a capacitance value or assume dial divisions are linear.
A documented test explains why the fixture matters. Texas Instruments' TIDU736A reference design used external electrodes on a plastic container with a 2 mm wall and tested water. Its setup instructions emphasize minimizing the air gap; they also warn that residual film affects the capacitance reading. This is an electronics reference design, not a qualified industrial level switch. The useful lesson is to preserve the test conditions—not to reuse its wall thickness or resolution as a sensor rating. TI reference design, Section 6.
If liquid-present and drained-wall behavior cannot be separated reliably, stop increasing sensitivity. Revisit the sensing position, mounting, compensation capability or measurement method. Adjustment cannot create a distinction that the installation does not provide.
Which specifications should you confirm before ordering?
Request the exact model and suffix, installation drawing, datasheet and setup instructions. Each important specification should answer a requirement of the tank or controller.
| Check | What the evidence should establish |
|---|---|
| Sensing arrangement | External switch, wetted point probe or continuous system; permitted wall, gap and mounting conditions. |
| Adjustment and compensation | How to teach or adjust, retain or lock settings, and handle the specified coating or foam—not an assumed feature. |
| Electrical interface | Supply, wiring, PNP/NPN or other output, load limits, residual voltage and leakage where relevant, connector pinout and PLC input compatibility. |
| Timing and state logic | Response and release delays, startup behavior, output polarity, and what the controller sees during loss of power or connection. |
| Environment and process | Ambient and medium temperature limits, ingress protection and cleaning compatibility. For a wetted probe, also seals, pressure rating and chemical compatibility. |
| Replacement consistency | The delivered suffix, configuration and firmware where applicable match the evaluated version; changes affecting the application are identified. |
For a rapidly rising level, sensor response is only part of the stopping delay. The PLC, valve or pump response and continuing inflow also affect the final liquid level. Set the physical detection point using the complete process behavior, not the sensor's switching speed alone.
What must a production trial prove?
The trial should demonstrate reliable detection, reliable release and the intended control action under agreed operating conditions. Define acceptable trip-height variation, response time and false-state behavior before testing; there is no universal number of cycles that proves every liquid application.
Illustrative example: the tank drains, but the switch still says “liquid present”
Assume an external sensor works with water in a clean plastic reservoir. In the proposed detergent application, the filled tank is detected, but a film remains on the inner wall after draining and the output does not release. This is a hypothetical troubleshooting scenario, not a reported customer test.
First separate sensing from wiring. Check the sensor's documented indication and output against the PLC input. If the sensor output has released but the PLC still reports liquid, investigate polarity, wiring, input configuration or program latching. Do not retune a sensor to compensate for controller logic.
If the sensor itself remains in the liquid-present state: compare the drained coated wall with the cleaned empty wall at the same setting. Then repeat with detergent present. Lowering sensitivity is useful only if detection of the real liquid still works across the required conditions.
Decision: do not accept the original water demonstration as sufficient. If no common setting detects detergent and rejects the expected drained film, evaluate a documented coating-tolerant model or another sensing arrangement, then repeat the trial.
Record the conditions that a replacement must reproduce
- Installation: exact sensor, wall sample, mounting dimensions, cable route and saved settings.
- Process: liquid formulation, temperature, fill/drain behavior, foam, residue and cleaning state.
- Results: trip and release heights, incorrect transitions, recovery after startup, and the actual PLC/control response.
- Variations: container replacements, sensor remounting and the agreed operating extremes—without quietly retuning between passing tests.
An ordinary proximity switch and a successful level trial do not, by themselves, establish a safety-rated overfill protection function. Where failure can endanger people or cause a hazardous release, the protection architecture and validation need an appropriate application-specific safety assessment.
When should you choose another level-sensing method?
Change approach when the required states remain indistinguishable, the vessel prevents a suitable installation, or the required result exceeds a point switch's role. Do not keep adding sensitivity to an unsuitable arrangement.
- Metal vessel, one limit point: evaluate a compatible process probe, float or other point-level device with a suitable process connection.
- Heavy residue or foam: investigate a device with documented discrimination for the actual medium, or another principle. Check the proposed alternative's own coating and foam limits.
- Continuous level: compare a complete capacitance system with radar, ultrasonic or hydrostatic measurement as appropriate. These require their own access, geometry and process checks; they are not drop-in sensors that see through any tank wall.
For a focused supplier enquiry, send the required level point or span, liquid and temperature range, tank-wall details, a dimensioned mounting photo, expected residue/foam, controller interface and acceptance criteria. Ask the supplier to identify the proposed exact model, its application limits and the test needed to close any uncertainty.
Choose the sensor that distinguishes the required process states in the real installation—not the one with the largest headline sensing distance.
Sources and method references
- Balluff: external capacitive level sensing — non-metal walls and the product-specific Smart Level example.
- Endress+Hauser: capacitance measuring principle — conductive and non-conductive media in probe-based measurement.
- Texas Instruments: TIDU736A reference design — Section 6, documented water-test fixture, air gap and residual-film limitations.
- ifm: LMC applications — a frequency-sweeping product family's buildup-discrimination approach, not a generic capacitive-switch capability.
The selection workflow and illustrative scenario are engineering synthesis, not manufacturer test results. Product-family examples do not establish capabilities for an xsz sensor model. Hero: AI-generated concept illustration; state diagrams: conceptual illustrations.