Reliable empty-reservoir detection combines a sensor or process signal with validated thresholds, timing, plausibility checks, and a safe response because low level, an exposed pickup, inlet leak, blocked vent, bubbles, viscous fluid, blockage, and sensor failure can produce overlapping pump symptoms.

How to Detect an Empty Reservoir in a Micro Pump System

An empty source can cause missed doses, air entry, loss of prime, pump overheating, product waste, unreliable sampling, and damage in pumps that require liquid for lubrication or cooling. The safest detection method depends on the reservoir, pump type, fluid, dose size, environment, and consequence of continued operation.

No single indirect signal identifies every empty condition. Low flow can also result from a blocked filter, inlet leak, viscous fluid, closed valve, worn tube, or failed motor. Motor current can rise, fall, or remain similar depending on pump mechanism and fluid path.

Reliable detection combines the most direct available measurement with timing, plausibility checks, and a defined response and recovery process.

Define What "Empty" Means

The reservoir does not need to contain zero liquid before the system should stop.

Possible empty definitions include:

• Liquid below the pickup opening

• Minimum usable volume reached

• Remaining volume insufficient for the next dose

• Level below a warning threshold

• Air detected in the inlet

• Required flow cannot be maintained

• Container mass below a defined value

• Flexible bag fully collapsed

• Maximum expected withdrawal reached

Distinguish warning, refill, stop, and service thresholds. A product may warn early, complete the current dose, then prevent the next cycle.

Account for residual liquid that cannot be reached because of reservoir shape, tilt, pickup geometry, surface tension, or bag folds.

Define the consequence of an incorrect decision. A false empty alarm reduces availability; a missed empty condition may create an incorrect dose or damage.

Map Reservoir and Pickup Behavior

Understand how the source empties in every product orientation.

Record:

• Reservoir shape and capacity

• Minimum and maximum fill

• Pickup location and opening

• Usable and residual volume

• Venting or bag collapse

• Product orientation and tilt

• Agitation and return flow

• Foam, particles, and settling

• Refill or replacement procedure

A rigid reservoir, collapsible bag, cartridge, bottle, and open tank require different sensing strategies.

The liquid surface may not be flat during motion or pump operation. A narrow or irregular reservoir can show large height changes for small volumes.

Test the actual container and pickup rather than inferring usable volume from a CAD model alone.

Use Direct Level Sensing Where Practical

Level sensing can directly identify liquid position without relying on pump behavior.

Possible methods include:

• Float switch

• Capacitive sensing

• Optical sensing

• Ultrasonic or time-of-flight sensing

• Conductive electrodes

• Hydrostatic pressure

• Mechanical displacement

• External noncontact sensing

Selection depends on fluid conductivity, dielectric properties, optical clarity, foam, bubbles, vessel material, geometry, temperature, coating, and cleanliness.

A point sensor provides one threshold. A continuous sensor estimates level over a range but requires calibration to reservoir geometry.

Ask what happens with residue on the sensor, condensation, an empty but wet wall, tilted liquid, foam, or a changed fluid formulation.

Direct sensing is useful, but it should still be checked against pump flow and plausibility where empty failure has significant consequences.

Evaluate Float and Mechanical Level Switches

Floats and mechanical switches can provide simple threshold detection.

Check:

• Fluid density and buoyancy

• Mounting orientation

• Minimum vessel dimensions

• Hysteresis

• Mechanical travel

• Sticking from residue or particles

• Vibration and shock

• Material compatibility

• Wire and seal integrity

• Cleaning access

A float can remain high because of foam, surface tension, deposits, or mechanical interference. It may also oscillate near the threshold as the pump pulses or product moves.

Use software debounce or mechanical damping only after understanding the real level dynamics.

Test minimum and maximum temperature, fluid density, contamination, and product orientation. A switch tested in water may not float or release the same way in another liquid.

Consider whether the float creates a dead volume or traps product.

Evaluate Capacitive and Optical Level Sensors

Capacitive sensors detect changes in dielectric properties. They can operate through some nonmetallic vessel walls but are influenced by wall thickness, fluid, residue, humidity, nearby structures, grounding, and calibration.

Optical point sensors detect a change in light behavior when a surface contacts liquid. They can be affected by coating, bubbles, foam, color, opacity, condensation, and ambient light.

For either method, test:

• Every fluid formulation

• Clean and coated surfaces

• Full and empty but wet states

• Temperature and humidity

• Vessel tolerances

• Sensor position

• Bubbles and foam

• Cleaning cycles

• Cable and electromagnetic environment

Do not assume a noncontact sensor eliminates compatibility concerns. The vessel wall, adhesive, mounting, and product residue still influence performance.

Define startup self-test and behavior when the sensor becomes disconnected or saturated.

Use Reservoir Weight or Load Cells

Weighing the reservoir provides a direct measure of remaining mass when the container and mounting allow it.

Potential advantages include:

• Continuous remaining-quantity estimate

• Independence from fluid color or optical clarity

• Ability to detect consumption over time

• Support for dose verification

Challenges include:

• Product movement and vibration

• Tube and cable forces on the reservoir

• Variable container tare

• Mounting stress

• Temperature drift

• Multiple fluids or components on one scale

• User contact or refill force

• Fluid density when converting mass to volume

Flexible tubing can support part of the reservoir weight and create level-dependent error. Route connections to minimize changing force.

Calibrate tare and span with production containers and mounting. Define how partial refill, container replacement, and spilled fluid are handled.

Use filtering that preserves a timely empty decision without following every pump pulse.

Use Flow Measurement

A flow sensor can detect when pump command no longer produces expected flow.

Compare:

• Commanded flow or dose

• Measured flow

• Response delay

• Minimum detectable flow

• Pump startup and prime behavior

• Sensor zero and drift

• Bubbles

• Pressure and viscosity

Low flow is not unique to an empty source. It can indicate a blocked inlet or outlet, failed valve, collapsed tube, motor stall, high viscosity, or sensor fouling.

Use supporting signals such as inlet pressure, reservoir level, motor current, and bubble detection.

For short doses, sensor response may be too slow to prevent one incomplete dispense. The controller may need to reject the dose, re-prime, or notify the process.

Do not allow a low flow reading to drive unlimited pump speed or run time.

Use Pressure or Vacuum Signals

Inlet pressure can change when the source empties, but the direction and magnitude depend on reservoir design.

With a vented rigid reservoir, an exposed pickup may draw air and inlet vacuum can fall or become unstable. Before exposure, vacuum may rise as liquid level drops or a filter loads.

With a sealed rigid reservoir and blocked vent, vacuum may rise even though liquid remains.

With a flexible bag, vacuum may rise when the bag folds or stops collapsing.

Therefore, pressure thresholds should distinguish:

• Empty pickup

• Vent blockage

• Inlet restriction

• Air leak

• Viscous fluid

• Filter loading

Combine pressure with flow or level where possible. Use time patterns and pump state, not one instantaneous value.

Place the sensor near the pump inlet and validate its range, response, bubble behavior, and port cleanliness.

Use Bubble Detection

An optical bubble detector or other gas-liquid sensor can identify air entering the inlet after the pickup is exposed.

Advantages include direct detection near the failure mechanism. Limitations include:

• Bubbles from inlet leaks

• Dissolved gas release

• Foam

• Transparent-tube requirements

• Fluid color or opacity

• Tube dimensional variation

• Sensor alignment

• Residue and condensation

Define the smallest gas segment that matters and the detection response time.

A bubble signal may arrive after part of an incorrect dose has already been delivered. The controller must decide whether to stop, mark the dose invalid, close a valve, or re-prime after refill.

Test known bubble sizes, speeds, fluids, tube lots, temperatures, and orientations.

Do not treat all detected air as an empty reservoir without checking for leaks and outgassing.

Use Motor Current with Caution

Motor current reflects torque and electrical behavior, not liquid level directly.

An empty condition can produce different signals:

• Lower load when liquid resistance disappears

• Similar load because tube compression dominates

• Higher or irregular load from failed priming or bubbles

• Changed load only at certain speeds

Current is also affected by:

• Outlet pressure

• Inlet restriction

• Viscosity and temperature

• Peristaltic occlusion

• Diaphragm valve behavior

• Motor wear

• Supply voltage

• Driver control

Use current as one feature in a fault model, not a universal empty detector.

Record command, speed, voltage, current, flow, pressure, and reservoir state across multiple pump samples. Validate thresholds at cold startup and pressure extremes.

Use Run Time or Pump Motion as an Estimate

The controller can estimate remaining volume by integrating calibrated pump output or counting doses, revolutions, steps, or operating time.

This approach can be low cost but accumulates error from:

• Initial fill uncertainty

• Refill amount

• Tube wear

• Pressure and viscosity

• Bubbles

• Failed doses

• Calibration drift

• Manual fluid removal

• Leakage or siphoning

• Power or data loss

Use estimation as a prediction or plausibility signal unless accuracy is validated for the full process.

Reconcile the estimate after reservoir replacement, refill, or a direct level measurement.

Protect stored totals from corruption and interrupted writes. Define behavior when the estimate is missing or inconsistent.

Do not assume motor motion equals delivered liquid.

Combine Signals for Better Diagnosis

Signal combinations can distinguish overlapping faults.

Examples:

• Low level plus low flow supports an empty-source diagnosis

• High inlet vacuum plus low flow suggests restriction or vent blockage

• Bubble detection plus normal motor speed supports air entry

• Low flow plus high outlet pressure suggests downstream blockage

• No flow plus low motor speed suggests drive or stall fault

• Weight decrease without commanded flow suggests leakage or siphoning

Build the logic from measured data rather than assumptions.

Use fault persistence, state machines, and plausibility windows to avoid false decisions during startup, priming, valve switching, or natural pulsation.

Where consequences are significant, use independent signals that do not share the same failure mode.

Document the diagnostic confidence and which faults remain indistinguishable.

Define Warning and Stop Thresholds

Set thresholds based on usable volume and process needs.

A warning threshold should allow enough time or remaining volume for:

• Completing a safe operation

• Alerting the user

• Preparing replacement fluid

• Preventing air from entering a sensitive path

• Avoiding an incomplete next dose

Include sensor tolerance, reservoir variation, tilt, fluid movement, pickup geometry, pump delay, and production variation.

Use hysteresis or state control so a level near the threshold does not repeatedly change status.

Do not set hysteresis so wide that the system resumes without enough fluid.

For continuous processes, define how much volume can be delivered between detection and pump stop.

Test the threshold with minimum and maximum dose size and pump speed.

Define the Safe Response

The correct response depends on pump type, fluid, and process risk.

Possible actions include:

• Stop the pump

• Complete or abort the current dose

• Close a shutoff valve

• Mark the dose invalid

• Prevent automatic retry

• Reduce speed

• Move to a safe position

• Notify the user or upstream controller

• Log the event

• Request reservoir replacement

Stopping the motor may not stop siphoning, backflow, or stored-volume release.

Some peristaltic pumps can tolerate limited dry running, while other pumps rely on liquid for cooling or lubrication. Use exact supplier data and validation.

Do not repeatedly prime an empty source without a maximum time, attempt count, and waste or overflow control.

Plan Refill and Recovery

After refill or container replacement, the system may contain air and require re-priming.

Define:

• How replacement is detected

• Whether the user confirms the action

• Prime speed and timeout

• Valve positions

• Waste destination

• Bubble-clear criterion

• First-dose handling

• Calibration or tare update

• Alarm reset

Prevent a partially connected or incorrectly oriented reservoir from being accepted.

For load-cell systems, capture the new tare and expected fill range. For level sensors, confirm the signal changes through a plausible sequence.

For disposable bags or cartridges, control part identity and connection.

Test refill after complete empty, early warning, power loss, and interrupted prime.

Handle Sensor Failure

Every empty-detection method has failure modes.

Consider:

• Sensor disconnected

• Short circuit

• Frozen output

• Drift

• Coating or fouling

• Blocked pressure port

• Stuck float

• Bubble detector misalignment

• Load-cell overload

• Corrupted remaining-volume estimate

• Communication loss

Define valid range, update timing, plausibility, startup self-test, and diagnostic coverage.

A frozen normal reading can be harder to detect than an out-of-range failure. Compare the signal with pump operation and expected consumption.

Choose a safe fallback. The system may stop, limit operation, require service, or use a secondary estimate depending on risk.

Do not bypass a failed sensor indefinitely without a controlled degraded mode.

Account for Pump Type

Peristaltic Pumps

Tube compression creates much of the motor load even when no liquid is present, so current may change little. Flow or bubble sensing is often more direct.

Dry running may be mechanically possible for some configurations, but tube heat, speed, pressure, and service life still require review.

Diaphragm Liquid Pumps

An empty source introduces gas into the chamber. Prime may be lost, valves may respond differently, and the pump can fail to rebuild liquid pressure.

Some internal components may rely on fluid conditions. Use supplier-approved dry-run limits.

Gas Pumps

An "empty reservoir" may instead be loss of sample gas, disconnected line, or open inlet. Pressure, flow, gas composition, and process state are more relevant than liquid level.

Match the detection model to the exact pump and medium.

Validate across Fluid and Environmental Conditions

Sensor behavior changes with fluid and environment.

Test:

• Minimum and maximum temperature

• Viscosity and density range

• Foam and bubbles

• Fluid color and conductivity

• Particles and residue

• Full, low, and tilted reservoir

• Clean and loaded filters

• Minimum and maximum voltage

• Cold startup and thermal steady state

• Vibration and transport

• Humidity and condensation

Capacitive and optical sensors can shift with fluid formulation or residue. Float buoyancy changes with density. Pressure and flow thresholds shift with viscosity and temperature.

Use production containers, pickups, tubing, electronics, software, and mounting.

Include multiple pump, reservoir, and sensor samples.

Build a Representative Empty Test

Empty the real reservoir through normal operation while recording:

• Actual remaining mass or volume

• Level-sensor output

• Pump command and speed

• Flow or dose

• Inlet and outlet pressure

• Motor voltage and current

• Bubble signal

• Fluid and sensor temperature

• Warning and stop state

• Delivered volume after detection

Repeat for continuous flow and intermittent dosing. Test full-to-empty, partially filled, partially refilled, tilted, and flexible-container fold conditions.

Introduce non-empty faults such as blocked vent, inlet restriction, air leak, loaded filter, viscous fluid, and sensor failure to measure false diagnosis.

Verify safe stop, valve behavior, alarm, logging, refill, priming, and first valid dose.

Preserve raw synchronized data for threshold tuning.

Empty-Detection Checklist

• Empty, warning, refill, and stop states defined

• Usable and residual volume measured in all orientations

• Reservoir, pickup, vent, and flexible-container behavior mapped

• Direct level methods evaluated with fluid and residue

• Float, capacitive, optical, conductive, or other sensor limits tested

• Load-cell mounting, tare, tube force, vibration, and temperature validated

• Flow detection range, delay, bubbles, and false causes reviewed

• Inlet pressure patterns distinguished for empty, restriction, and vent blockage

• Bubble detection tested with leaks and outgassing

• Motor current used only with supporting evidence

• Run-time or pump-motion estimate includes drift and refill uncertainty

• Multiple signals combined where consequences require it

• Warning, stop, hysteresis, and remaining-volume margins validated

• Safe response handles siphoning and stored fluid

• Refill, prime, alarm reset, and first-dose sequence defined

• Sensor disconnection, frozen output, drift, and fouling tested

• Pump-specific dry-run behavior reviewed

• Fluid, temperature, orientation, vibration, and production variation included

• Real and false empty conditions tested end to end

Common Empty-Detection Mistakes

• Defining empty as zero physical liquid

• Using total reservoir capacity instead of usable volume

• Testing one upright orientation only

• Assuming a flexible bag always collapses completely

• Using motor current as the only signal

• Treating low flow as proof of empty source

• Ignoring blocked vent and loaded-filter faults

• Detecting bubbles without distinguishing inlet leaks

• Integrating pump run time as if motion equals delivered volume

• Setting thresholds from one prototype

• Stopping the motor without controlling siphoning or stored volume

• Retrying prime indefinitely

• Ignoring sensor coating, frozen data, and disconnected states

• Testing detection but not refill and first-dose recovery

Frequently Asked Questions

What is the best way to detect an empty pump reservoir?

The best method depends on container, fluid, dose, and risk. Direct level or weight sensing is often useful, while flow, pressure, bubble, and current signals can provide confirmation and fault diagnosis.

Can motor current detect an empty reservoir?

Sometimes it contributes, but current also changes with pressure, viscosity, temperature, occlusion, voltage, and wear. It should not be assumed to provide universal empty detection.

Can a flow sensor detect an empty source?

It can detect missing output, but blockage, leaks, bubbles, pump faults, and sensor fouling can also reduce flow. Combine it with other signals where needed.

How early should a low-level warning occur?

Set it early enough to complete the intended safe action and prevent an incomplete next dose, including sensor tolerance, pickup residual, tilt, process delay, and replacement time.

What should the pump do after empty detection?

It may stop, close a valve, reject the dose, log a fault, and request refill. The response should also control siphoning, stored pressure, dry-run exposure, and automatic retry.

How should refill recovery be tested?

Test container replacement, connection, sensor reset, priming, bubble clearance, waste handling, first-dose verification, power interruption, and incorrect or partial refill.

Kamoer Empty-Reservoir Detection Support

Kamoer can help evaluate pump behavior, reservoir and pickup design, flow and pressure signals, bubbles, motor current, dry-run limits, fault logic, and representative empty and refill tests for OEM systems.

Related Stories

Kamoer Fluid Tech (Shanghai) Co., Ltd.

pump@kamoer.com