Liquid-Level Senders: Distinguish a Stuck Float from a Genuinely Stable Level

System boundary

The diagnostic spans liquid change, float motion, contact position, card output and receiver processing. The resistor card alone cannot prove that liquid level or the float should have moved.

Integration interfaces

Integration risks

Recognize the two states that produce the same signal

Consider a tank at rest with neither filling nor withdrawal. A healthy float can remain at one position for hours. A jammed float at that same position can produce an identical resistance. Requiring movement every few minutes would condemn a normal stationary system. The missing information is not finer ADC resolution; it is evidence that the physical input changed.

This is an observability problem. A fault can only be distinguished when the available inputs and outputs behave differently under the competing explanations. Record which independent signals exist: a controlled fill event, a measured withdrawal, another sensing principle, a position witness during service, or a known mechanical stimulus. A pump command alone may be insufficient if flow can be blocked or returned to the same tank.

Define an independently supported level change

For a controlled bench test, add or remove a known liquid quantity while observing the actual free surface. The expected height change depends on tank geometry. In a constant-area vessel, 200 cubic centimetres across 100 square centimetres corresponds to a 2 centimetre change. Select a change large enough to exceed the combined uncertainty of quantity, area and observation, while remaining inside safe operating conditions.

In an installed system, the evidence may be less direct. A flowmeter, verified valve state and inventory balance can support a predicted change, but leakage, return flow and trapped compartments can weaken that prediction. Label such an event inconclusive when the input evidence is insufficient. Do not strengthen the diagnosis by borrowing the level change from the very sender being evaluated.

Compare expected motion with a justified detection threshold

The expected electrical change follows the float linkage and card transfer curve, including any nonlinear region. A two-centimetre level change does not imply the same voltage change at every position. Use the local transfer relationship and receiver loading to estimate an interval of expected response. The interval must include mechanical free play and normal repeatability, not just nominal resistance.

A candidate fault event occurs when independently expected motion exceeds the detection threshold yet the observed response remains inconsistent with that motion. Keep a timing allowance for liquid equalization and normal float response. If the allowed window is shortened merely to obtain faster detection, the false-alarm rate can rise. Confirm the window using the actual assembly and fluid conditions rather than an arbitrary software timeout.

Use electrical diagnostics for the faults they can reveal

An input circuit can apply a controlled electrical stimulus to help distinguish an open wire or a short to a rail from a connected resistance. Such checks are valuable because a broken harness can otherwise resemble an endpoint or stationary value. The stimulus and sampling sequence must be designed for the actual receiver; it must not overstress the card or unintentionally command equipment.

Passing that test establishes something about the electrical path, not that the float is free. A mechanically jammed float can present a perfectly plausible impedance. Likewise, a software display can remain frozen while the raw ADC value moves normally. During investigation, preserve values at successive boundaries: card terminals, receiver input, digitized sample, filtered result and displayed quantity.

Choose the next test from the available evidence

A useful diagnostic table includes an inconclusive outcome. Forcing every event into pass or fail encourages assumptions about flow and motion that were never measured. The following distinctions keep the first response proportional to the evidence and prevent an electrical repair from being prescribed for a hydraulic problem.

Separate a jammed float from a disconnected linkage

When access is safe, observe the liquid, float body, arm and wiper as separate objects. A float that follows the surface while the wiper remains stationary points to a different fault than a float held against the enclosure. Deposits, geometric interference, abnormal buoyancy and assembly damage can each prevent the intended motion, but a photograph alone does not establish which caused the recorded event.

Preserve the as-found state before cleaning or manually freeing the mechanism. A test performed only after touching the float may remove the evidence. Any intentional mechanical stimulus should be defined by the assembly designer and conducted under safe, de-energized conditions where required. This page does not prescribe field manipulation inside a fuel tank or another hazardous vessel.

Validate detection with healthy stationary cases as well as faults

A diagnostic that detects every deliberately locked float may still be unusable if it alarms on normal service. Include healthy long stationary intervals, small level changes below resolution, interrupted filling, tilt changes and delayed equalization. Then introduce controlled faults whose physical state is independently known. Keep these populations distinct so that detection and false alarms can both be evaluated.

Do not train and judge the rule on the same few traces. Reserve different assemblies or operating sequences for confirmation. Report missed faults, false alarms and inconclusive events with their triggering conditions. A result outside the tested fluid, geometry or receiver configuration remains outside the established diagnostic scope, even if its electrical card belongs to the same product family.

Specify a diagnostic interface without assigning invented capability

The card drawing should identify the contact path, terminal functions and electrical operating conditions. The sender assembly specification adds float and linkage motion. The receiver specification adds timestamps, raw-data access, electrical tests and diagnostic states. Keeping these interfaces explicit makes it possible to investigate a stationary signal without silently changing the resistance curve to mask an unresolved mechanism.

ChipSimple can review the thick-film card requirements and drawing-defined interfaces. The system integrator owns the fault-detection logic, safe response and validation in the intended liquid-level application. Request a clear record of what independent stimulus must exist before a stuck-float conclusion is permitted; that requirement is often more useful than a more aggressive timeout.