By Wu Peng, Senior Process Instrumentation Engineer · Last reviewed September 7, 2026
A capacitive fuel level sensor measures fuel height with two fixed electrodes, usually an inner rod inside an outer tube. Air between them has a relative permittivity of about 1.0, while diesel and gasoline sit near 2.0 to 2.2. Capacitance rises as fuel replaces air along the probe. There are no floats or moving parts, the output is continuous, and resolution of about 1 mm is routine in clean fuel.
That simple principle hides the two mistakes buyers make most often. One is matching the wrong output convention to an existing gauge. The other is assuming the calibration survives a change of fuel. It does not; a probe calibrated on straight gasoline can read a quarter tank of E10 as nearly full.
This guide covers the output menu from 240-33 ohm senders to CAN bus and the three accuracy bases that appear on datasheets. It then puts numbers on ethanol and water errors and shows where capacitive gauging fits against other tank technologies.
Contents
- Working principle
- Output signal conventions
- Reading the accuracy spec
- Fuel changes the calibration
- Water in the tank
- Aircraft fuel gauging
- Selection and installation
- FAQ
Working principle
The probe is a capacitor whose dielectric is whatever fills the gap between the electrodes. An empty probe reads a baseline capacitance set by its geometry. Only the wetted portion raises the value above that baseline. The extra capacitance per millimeter is proportional to (εr − 1), the relative permittivity of the liquid minus that of air.
The transmitter electronics excite the electrodes with an AC signal, measure the resulting capacitance, and convert it to a level output.
| Medium | Relative permittivity at 20 °C |
|---|---|
| Air | 1.0 |
| Kerosene and jet fuel | 1.8 to 2.0 |
| Gasoline (no ethanol) | 1.9 to 2.1 |
| Diesel | 2.0 to 2.2 |
| Ethanol | about 24 |
| Water | about 80 |
Values shift slightly with temperature; the electronics compensate for small drift, not for a different liquid.
Two numbers from this table drive everything else on this page. Fuel near 2.1 adds about 1.1 units of signal per wetted millimeter; water at 80 adds about 79 units, roughly 72 times more.
Bare-electrode fuel senders work only on non-conductive liquids such as these hydrocarbon fuels. Conductive or corrosive media need a PTFE-insulated industrial probe instead; see our capacitance level transmitter for that class of instrument. Our guide to the dielectric constant in level measurement covers how permittivity behaves across other technologies.
Output signal conventions
Most gauge problems reported as sensor faults are convention mismatches. A capacitive sender emulates whichever gauge standard it is configured for, and there are more standards than most people expect.
| Output | Convention | Where it is used |
|---|---|---|
| 240-33 ohm | 240 Ω empty, 33 Ω full | US aftermarket standard |
| 0-90 ohm | 0 Ω empty, 90 Ω full | GM vehicles, 1965 on |
| 73-10 ohm | 73 Ω empty, 10 Ω full | Ford before 1987 |
| 16-158 ohm | 16 Ω empty, 158 Ω full | Ford 1987 on |
| 10-180 ohm | 10 Ω empty, 180 Ω full | VDO and most European gauges |
| 0-5 V or 0-10 V | Voltage rises with level; 0-5 V on 12 V systems, 0-10 V on 24 V | Panel indicators, telematics inputs |
| 4-20 mA | Two-wire current loop, typical 250 Ω load | Industrial tanks, PLC and RTU inputs |
| RS485, J1939, CAN | Digital registers, polled or broadcast | Fleet telematics and fuel management |
Programmable senders exist; one published spec allows any resistance span from 3 to 500 Ω in 3 Ω steps.
The empty end of the US convention is 33 Ω rather than zero for a reason. A shorted signal wire reads 0 Ω, and the gap between 0 and 33 lets the gauge indicate a wiring fault instead of a plausible empty tank. Note also that the 240-33 aftermarket standard runs high-to-low as the tank fills, while VDO-style European gauges run low-to-high. Wiring one to the other produces a gauge that reads backwards, a fault that has confused many retrofits.
Reading the accuracy spec
Datasheets quote capacitive sender accuracy on at least three different bases, and the difference matters more than the headline number. Here is what each basis means on a 1,000 mm probe.
| Spec as written | Error on a 1,000 mm probe | Comment |
|---|---|---|
| ±0.5% of full scale | ±5 mm at any level | Common industrial spec |
| ±1% of full scale | ±10 mm at any level | Typical standard build |
| ±2% of probe length | ±20 mm at any level | Seen on cut-to-length automotive senders |
| ±0.5% of reading | ±0.5 mm at 100 mm, ±5 mm at full | Rare; much tighter at low level |
Temperature adds its own term. A published typical drift figure for this class of probe is ±0.05% of full scale per °C. That moves a 1,000 mm reading by up to ±10 mm over a 20 °C day-night swing, before the level changes at all.
One more caution from comparing published datasheets. The same sender is sometimes quoted with different temperature ranges or accuracy bases on different pages from the same maker. Ask for the basis in writing: percent of what, at which temperature, in which fluid.
Fuel changes the calibration
A capacitive sender is calibrated at two points, empty and full, in one specific fuel. The electronics then compute level as a fraction of that span. Change the fuel and the span changes underneath the calibration, because the signal per wetted millimeter scales with (εr − 1).
Response multiplier = (εr,new − 1) / (εr,calibrated − 1)
Ethanol is the worst case. Straight gasoline sits near εr = 2.1. One US patent measurement puts gasoline with 10 to 15 percent ethanol at an effective εr of about 5.3.
An ideal linear blend rule predicts about 4.3; polar blends do not mix linearly, which is why the measured value runs higher. Take the measured figure:
(5.3 − 1) / (2.1 − 1) = 3.9, so a tank one quarter full of E10 computes as 0.25 × 3.9 = 98 percent: the gauge shows full
Field reports match the direction of that arithmetic. Aircraft homebuilders running mixed 100LL and E10 auto fuel report gauge errors of 3 to 4 gallons (11 to 15 liters). In one documented build, seasonal changes in ethanol content broke a freshly completed calibration twice.
The remedies, in order of effort: keep the tank on one fuel, or recalibrate empty and full after every fuel change. The third option is a reference cell, a small always-immersed capacitor at the tank bottom that tracks the live dielectric and rescales the reading continuously. Biodiesel and kerosene sit close enough to diesel that the shift is far smaller, but any fuel change warrants a check against a dipstick.
Water in the tank
Water is the other large error source, and the arithmetic is the same formula pointed the other way. At εr of about 80, each millimeter of water against the probe adds (80 − 1) / (2.1 − 1) = 72 times the capacitance of a millimeter of fuel.
A 10 mm layer of water at the probe base adds as much signal as roughly 700 mm of diesel above it
On most probes that alone pushes the indicated level to the top of the scale. Free water is also conductive, and if it bridges the electrodes it drives many probes to a hard maximum. A reading pinned at full that refuses to move is the classic signature: drain and check for water at the probe base before condemning the electronics.
Condensation builds water in any vented fuel tank over time. That is one reason generator day tanks and storage tanks pair the level instrument with periodic water-bottom checks. For how water layers affect the other tank technologies, and how to pick one for diesel storage, see our diesel tank level gauge selection guide.
Aircraft fuel gauging
Aviation is where capacitive gauging is the default rather than the alternative. An aircraft capacitance fuel level sensor works on the same physics as a tank probe, packaged for flight. Aircraft use coaxial capacitive probes for two reasons. Nothing wears under constant vibration, and a single float cannot describe a fuel surface that tilts with pitch and roll.
A transport aircraft tank carries several probes at different stations. The gauging computer solves the fuel plane from their combined readings, so the indication holds through attitude changes. A compensator unit, an always-immersed capacitor low in the tank, corrects continuously for fuel dielectric and temperature, the production version of the homebuilder’s reference cell. With a density input the system indicates mass rather than volume, which is what the engines burn.
Application example
Aviation sector, Poland. A customer needed continuous gauging for a Jet A1 fuel tank at a fixed immersion depth. We proposed a 360 mm flanged capacitive level sensor with a 0-5 V output. Across the 360 mm active length, that signal resolves to about 13.9 mV per millimeter of fuel; operating pressure, temperature and the flange standard were left to be confirmed at order.
Selection and installation
Four checks cover most of the selection work. First, probe length. The probe must clear the tank bottom, and cut-to-length senders must be recalibrated after trimming because the empty baseline changes with the geometry.
Second, mounting. Keep the probe vertical, away from the fill stream and return lines so splash does not ripple the reading. Leave clearance below the tip for sludge and water to settle without touching it.
Third, the medium. Bare-electrode senders are for non-conductive fuels and oils only. Conductive, aggressive or high-temperature media call for an insulated industrial probe from the capacitance level sensor line.
Fourth, area classification. Gasoline vapor spaces are classified zones, so specify certified equipment where the authority requires it. Our guide to Class 1 Division 1 vs Division 2 covers the classifications themselves.
Published prices run from about US$110 for automotive cut-to-length senders to about US$470 for industrial 316L probes with 4-20 mA output. If the application needs tighter performance than a capacitive probe delivers, two alternatives are worth knowing. A magnetostrictive level transmitter reads to the millimeter for inventory-grade tank gauging. An external ultrasonic fuel sensor mounts on the outside of a truck tank with no penetration at all.
FAQ
How does a capacitive level sensor work?
Two electrodes, usually a rod inside a tube, form a capacitor whose dielectric is the tank contents. Air has a relative permittivity near 1.0 and fuel near 2.1, so capacitance rises in proportion to the wetted length as liquid replaces air. The electronics measure that capacitance and convert it to a continuous level signal.
What are the disadvantages of capacitive sensors?
The calibration is tied to one liquid, so a fuel change, an ethanol blend or a water layer shifts the reading. A quarter tank of E10 on a gasoline calibration can read full. Bare probes suit non-conductive liquids only, temperature adds drift, and the probe must physically reach through the tank.
How does a capacitive sensor work?
Any capacitive sensor measures the capacitance between two conductors, which changes when the material between or near them changes. Proximity switches and touch screens sense an approaching object this way. A level probe applies the same principle vertically: the liquid becomes the dielectric, and capacitance tracks how far up the probe it reaches.
What are the different types of fuel level sensors?
The common types are resistive float senders, capacitive probes, pressure-based sensors, magnetostrictive probes and external ultrasonic sensors. Float senders are the cheapest and the automotive standard. Capacitive probes read continuously with no moving parts, pressure sensors infer level from head, magnetostrictive probes give millimeter-grade gauging, and external ultrasonic sensors read through the tank wall.
Request a quote
Tell us the fuel, the tank depth, the output convention your gauge or telematics unit expects, and any hazardous-area requirement. We will match a capacitive probe or recommend a better technology for the tank. Reach our application engineers or use the form below.
Written and technically reviewed by Wu Peng and the Instranova engineering team.