Level Measurement Technologies Compared: Float, Capacitive, Hydrostatic, Ultrasonic, Radar, Guided-Wave

A cement plant swaps in an 80GHz radar transmitter on a 30-meter silo, expecting the sharpest accuracy on the market. Instead, the display throws a "full level" alarm at 50 percent, then drifts and locks up entirely every time the pneumatic filler kicks on.

The cause isn't a bad unit. Dense airborne cement dust scatters the narrow 80GHz beam badly enough to fool the echo processor, a failure mode called Mie scattering (NotebookLM, Level Measurement notebook, 2026). The fix was a lower-frequency 26GHz radar with a wider beam built to punch through dust, plus a compressed-air purge to keep the lens clean.

That case sums up the whole problem: the highest spec on a datasheet isn't automatically the right sensor for your tank. Six technologies solve overlapping but distinct problems, and picking wrong costs real money.

TL;DR: Six technologies cover almost every level application: float switches, capacitive probes, hydrostatic/DP cells, ultrasonic, non-contact radar, and guided-wave radar (GWR). Match the technology to your medium's dielectric constant, foam/vapor exposure, and vessel pressure first; radar and GWR carry the highest upfront cost but the lowest five-year total cost of ownership thanks to near-zero maintenance (NotebookLM, Level Measurement notebook, 2026).

For the narrower question of radar versus ultrasonic specifically, see our deep dive on radar vs ultrasonic level measurement, which covers beam angle and dead zone math this article won't repeat.

Which Level Technology Fits Your Tank? Full Comparison Table

Six technologies split cleanly by contact method, media tolerance, and cost. Hydrostatic and DP cells run the lowest capital cost at 20-40 percent of an equivalent radar system, while guided-wave radar runs the highest at $2,000-8,000 depending on probe length (NotebookLM, Level Measurement notebook, 2026). The table below lines up all six side by side.

Citation capsule: Six level technologies cover the field: float/displacer, capacitive, hydrostatic/DP cell, ultrasonic, non-contact radar, and guided-wave radar, ranging from $10 float switches to $8,000 guided-wave radar probes, with accuracy spanning ±25 mm down to ±1 mm (NotebookLM, Level Measurement notebook, 2026).

How Do Float and Displacer Level Sensors Work?

Float switches use buoyancy: a float rides the liquid surface and trips a switch or moves a mechanical linkage as level changes. They're simple, cheap, and reliable in clean liquids, with basic point-level models starting around $10-30 (NotebookLM, Level Measurement notebook, 2026).

Simplicity comes with a mechanical cost. Moving parts bind, wear, and coat in viscous or sticky fluids, and heavy turbulence causes floats to bounce, producing relay chatter, rapid on-off cycling that can hammer downstream equipment (NotebookLM, Level Measurement notebook, 2026). Continuous mechanical floats hold ±2-5 mm accuracy, but point-level switches drift to ±25-50 mm because of mounting and liquid-density tolerances.

From the field: Float switches are the sensor I never worry about on a clean water tank and never trust in a sump with rag or grease buildup. Once a float starts sticking, it fails quiet, no alarm, just a pump that never kicks on.

Floats max out around 6-20 meters of stem or cable length, so they're a poor fit for tall process vessels regardless of medium (NotebookLM, Level Measurement notebook, 2026). Protective cages or stilling wells cut down on turbulence-induced chatter, but they don't solve coating or wear.

Citation capsule: Float switches deliver ±2-5 mm accuracy in continuous mode and ±25-50 mm as point-level switches, running $10-400 depending on materials, but their moving parts bind, wear, and chatter under turbulence (NotebookLM, Level Measurement notebook, 2026).

How Does Capacitive Level Measurement Work, and Where Does It Fall Short?

Capacitive sensors turn an electrode probe and the tank wall into a capacitor. As liquid, with a higher dielectric constant than air, rises past the probe, capacitance changes measurably, and the sensor converts that change into a level reading (NotebookLM, Level Measurement notebook, 2026).

It has no moving parts and holds up well at high temperature and pressure, a solid alternative to float switches once conditions get harsh. Standard capacitive probes hold ±0.25-1% of span, and specialized high-precision models tighten that to roughly ±1 mm (NotebookLM, Level Measurement notebook, 2026).

Two things break capacitive measurement. First, the math assumes a stable dielectric constant; any shift from temperature, moisture, or composition change forces a full recalibration. Second, a conductive coating building up on the probe creates parasitic capacitance that mimics a higher level than what's there (NotebookLM, Level Measurement notebook, 2026). Modern guard-electrode designs drive part of the probe at the same potential to route current around the buildup, but they don't eliminate the problem.

Capacitive probes need a dielectric constant above roughly 1.5 to register reliably, which rules them out for some light solvents and hydrocarbons (NotebookLM, Level Measurement notebook, 2026). For a related contact-sensing comparison outside of level applications, see our guide to inductive vs capacitive proximity sensors.

Citation capsule: Capacitive level sensors hold ±0.25-1% of span accuracy and run $565-1,100, but they require a stable dielectric constant above 1.5 and recalibration whenever material composition or moisture content shifts (NotebookLM, Level Measurement notebook, 2026).

How Accurate Is Hydrostatic and DP Cell Level Measurement?

Hydrostatic and differential pressure sensors measure the pressure at the bottom of a liquid column, which is directly proportional to height and density. DP transmitters run 20-40 percent the cost of an equivalent radar system, making them the most economical continuous option on the market (NotebookLM, Level Measurement notebook, 2026).

Accuracy holds at ±0.075-0.25% of full scale, tight enough for most inventory and process control needs, as long as one condition holds: the fluid's density has to stay stable (NotebookLM, Level Measurement notebook, 2026). Temperature swings that change density will skew every reading, silently, since the sensor has no way to tell density drift from a real level change.

Flush-mount hydrostatic sensors reach 0.1-50 meters of measurement height; submersible probe versions go far deeper, up to 250 meters of water column (NotebookLM, Level Measurement notebook, 2026). That range makes hydrostatic the default for deep wells and tall tank farms where radar's line-of-sight requirements get expensive.

Hydrostatic sensors also ignore surface conditions entirely. They read from the bottom of the tank, so foam, vapor, and agitation never enter the measurement (NotebookLM, Level Measurement notebook, 2026). If you're weighing hydrostatic against other pressure-based instrumentation, our pressure sensor types compared guide covers the broader family this technology belongs to.

Citation capsule: Hydrostatic and DP cell level sensors hold ±0.075-0.25% of full scale accuracy at 20-40 percent of radar's cost, immune to foam and vapor, but only as long as fluid density stays constant (NotebookLM, Level Measurement notebook, 2026).

When Does Ultrasonic Level Measurement Still Make Sense?

Ultrasonic sensors fire a sound pulse and time its round trip to the liquid surface, at 30-50 percent of radar's price for comparable range (NotebookLM, Level Measurement notebook, 2026). They're affordable, easy to install, and a strong fit for clean, atmospheric, stable-temperature tanks.

The same acoustic principle is the weakness. Foam absorbs and scatters the pulse instead of reflecting it. Heavy vapor and steam change the speed of sound the pulse travels through, and condensation on the transducer face can block it outright (NotebookLM, Level Measurement notebook, 2026). None of those failure modes have a reliable fix short of switching technology.

What gets missed: Ultrasonic's failure modes aren't rare edge cases confined to exotic process media. A wastewater plant tried ultrasonic on turbulent, foaming tanks and got constant false readings, since grease coated the transducer face and foam scattered the pulse before it ever reached the water (NotebookLM, Level Measurement notebook, 2026). Switching to non-contact radar fixed it completely.

Ultrasonic also cannot operate below roughly 0.1 bar absolute, since sound needs a gas medium to propagate; radar has no such floor. We covered the full accuracy, dead-zone, and cost breakdown between these two technologies in our radar vs ultrasonic level measurement guide, which is worth reading before you spec either one on a new tank. For the underlying sensing mechanics outside of level applications, see how ultrasonic sensors work.

Citation capsule: Ultrasonic level sensors cost 30-50 percent of comparable radar but fail predictably in foam, vapor, and vacuum, three documented failure modes with no fix short of a technology switch (NotebookLM, Level Measurement notebook, 2026).

What Makes Non-Contact Radar the Default for Harsh Tanks?

Non-contact radar sends microwave pulses that reflect off the liquid surface and reads the time-of-flight back, unaffected by temperature swings, pressure changes, heavy vapor, or dust (NotebookLM, Level Measurement notebook, 2026). That immunity to the exact conditions that break ultrasonic is why radar has become the harsh-tank default.

80GHz radar reaches ±1-3 mm accuracy and 100-150 meters of range; some tank-gauging models push accuracy under 0.5 mm (NotebookLM, Level Measurement notebook, 2026). It costs more upfront, averaging $1,000-1,200 for a mid-range unit, but its near-zero maintenance need pushes total cost of ownership below ultrasonic's over roughly a four-year horizon.

The one hard requirement is dielectric constant. Standard radar needs a value above 1.8 to bounce back a usable echo; water at roughly 80 reflects almost all of the signal, while hydrocarbons around 2 reflect only 2-4 percent (NotebookLM, Level Measurement notebook, 2026). Below that threshold, radar's accuracy advantage evaporates fast.

Is 26GHz Radar Just a Cheaper 80GHz Radar?

No, and treating them as interchangeable is exactly what caused the cement-silo failure at the top of this article. 26GHz radar runs a wider beam and lower accuracy, around ±5 mm against 80GHz's ±1-3 mm, but its longer wavelength punches through heavy dust and steam that would scatter an 80GHz signal (NotebookLM, Level Measurement notebook, 2026).

Frequency alone doesn't determine which is right. 26GHz maxes out around 70-80 meters range, while 80GHz reaches up to 150 meters, so the choice depends on tank height, media cleanliness, and whether dust or steam density is the dominant environmental factor (NotebookLM, Level Measurement notebook, 2026). Dusty solids applications, cement, grain, fly ash, tend to default to 26GHz for that reason alone.

Citation capsule: Non-contact radar holds ±1-3 mm accuracy at 80GHz and reaches 150 meters range, immune to vapor, dust, and pressure swings, but it needs a dielectric constant above 1.8 to produce a usable echo (NotebookLM, Level Measurement notebook, 2026).

When Should You Choose Guided-Wave Radar Instead?

Guided-wave radar sends the same microwave pulse down a physical probe, rod or cable, that extends into the media, concentrating signal energy along that path instead of letting it spread through open air (NotebookLM, Level Measurement notebook, 2026). That concentration is what non-contact radar can't match in three specific situations.

Low-dielectric media is one case: GWR maintains reliable readings down to a dielectric constant of 1.4, versus non-contact radar's 1.8 floor, which matters for LNG, LPG, solvents, and light hydrocarbons (NotebookLM, Level Measurement notebook, 2026). Heavy foam is another, since the probe gives the pulse a continuous path to punch through a layer that would scatter a free-space signal. And interface measurement, the boundary between two immiscible liquids like oil and water, is where GWR is the standard choice.

Interface measurement has a specific requirement worth knowing before you spec it: the two layers need a dielectric constant difference of at least 6 to 10 for the sensor to distinguish them reliably (NotebookLM, Level Measurement notebook, 2026). Below that gap, the reflection off the interface gets too weak to trust.

GWR's tradeoff is physical. The probe touches the media, so it's exposed to coating, corrosion, and mechanical wear that non-contact radar never sees, and its range tops out at 30-60 meters because it's capped by the physical probe length rather than signal strength (NotebookLM, Level Measurement notebook, 2026). It's also the priciest option on the table, $2,000-8,000 depending on probe construction.

Citation capsule: Guided-wave radar holds ±1-3 mm accuracy down to a dielectric constant of 1.4, the standard choice for oil-water interface measurement, but its probe is exposed to wear and its range caps at 30-60 meters based on physical probe length (NotebookLM, Level Measurement notebook, 2026).

How Do You Choose Between These Six Technologies?

Four questions settle most selections: is the vessel under vacuum or pressure, is foam or heavy vapor present, what's the medium's dielectric constant, and what's the budget per measurement point. Vacuum or extreme pressure rules out ultrasonic and mechanical floats immediately, leaving radar or GWR as the only realistic options (NotebookLM, Level Measurement notebook, 2026).

If foam or vapor is a daily condition rather than an occasional nuisance, default to radar or GWR before cost enters the conversation. Below a dielectric constant of 1.5, standard radar and capacitive sensors both struggle; GWR's 1.4 floor or ultrasonic's acoustic reflection become the workable paths instead (NotebookLM, Level Measurement notebook, 2026).

Cost still matters once the physics rules a technology in or out. A 40-tank retrofit on a fixed budget rarely justifies switching every tank to radar; upgrading only the tanks with documented drift, foam, or vapor history and leaving stable, clean tanks on hydrostatic or ultrasonic is usually the better financial call.

One factor overrides every branch above: hazardous-area certification and, on regulated storage tanks, functional safety requirements. API 2350 governs overfill prevention on above-ground petroleum storage tanks over 1,320 gallons, and it explicitly bars using the same sensor for routine filling and for the High-High overfill alarm (NotebookLM, Level Measurement notebook, 2026). Unattended, Category 3 tanks need a fully independent Automated Overfill Prevention System on top of the primary level instrument.

That independence requirement sits inside a broader functional-safety framework: IEC 61508 governs how manufacturers build and certify the sensor hardware, while IEC 61511 governs how plant owners design, install, and maintain the resulting safety instrumented system across its lifecycle (NotebookLM, Level Measurement notebook, 2026). SIL ratings, from SIL 1 through SIL 4, quantify how much risk reduction a given sensor and logic solver combination actually delivers.

Field data point: A petroleum refinery processing 200,000 barrels a day can lose up to $500,000 a year purely from inventory discrepancies caused by inadequate level-measurement accuracy, with a precision radar upgrade typically paying for itself within 6-12 months (NotebookLM, Level Measurement notebook, 2026). That's not a safety failure, it's a measurement accuracy problem hiding in plain sight on the accounting side.

Citation capsule: API 2350 requires independent sensors for overfill alarms versus routine filling, and unattended tanks need a separate automated overfill prevention system; IEC 61508 and IEC 61511 govern the underlying functional safety of the instruments used to meet that requirement (NotebookLM, Level Measurement notebook, 2026).

What Are the Most Common Level-Measurement Failure Modes?

Foam, vapor, and agitation cause more level-measurement failures than any hardware defect, and each technology responds to them differently. Ultrasonic is the most exposed: foam absorbs its acoustic pulse, vapor changes the speed of sound it depends on, and condensation blocks the transducer face outright (NotebookLM, Level Measurement notebook, 2026).

Non-contact radar handles vapor and foam far better since microwave propagation doesn't depend on the atmosphere, but it isn't immune to everything. High-pressure steam condensing on the antenna absorbs energy and can trigger false near-zone echoes; drop-shedding antenna designs and compressed-air purges are the standard fix (NotebookLM, Level Measurement notebook, 2026).

Agitation causes a different failure per technology. Floats and displacers chatter or physically strike agitator blades and tank walls. A GWR probe installed too close to a pump intake can scar, bend, or break from turbulent flow. Ultrasonic scatters its signal off an inconsistent surface angle, producing noisy readings and slow response (NotebookLM, Level Measurement notebook, 2026). Non-contact radar's FMCW processing and Doppler compensation handle surface turbulence the best of any non-contact option, and mounting a contact sensor inside a stilling well isolates it from agitation entirely.

A second documented case follows the same pattern. A boiling chemical reactor lost its radar signal completely, dropping to zero, because the antenna sat directly above the vapor outlet and steam condensation absorbed the microwave energy; relocating the mounting nozzle and adding a dry-air purge solved it (NotebookLM, Level Measurement notebook, 2026). Each failure traces back to a mismatch between the medium's behavior and the technology's known blind spot, not a random hardware fault.

Citation capsule: Foam, vapor, and agitation are the three dominant level-measurement failure modes; radar resists all three better than ultrasonic or float sensors, but even radar needs antenna purging or repositioning around condensing steam (NotebookLM, Level Measurement notebook, 2026).

Frequently Asked Questions

What is the difference between guided-wave radar and non-contact radar level sensors?

Non-contact radar sends microwave pulses through open air from an antenna at the tank top; guided-wave radar sends the same pulses down a physical probe that touches the media. The probe concentrates signal energy, so GWR handles foam, low dielectric constant, and turbulence better, while non-contact radar reaches farther and avoids mechanical wear (NotebookLM, Level Measurement notebook, 2026).

What dielectric constant do you need for radar or capacitive level measurement to work reliably?

Standard non-contact radar needs a dielectric constant above 1.8, capacitive sensors need above 1.5, and guided-wave radar can work down to 1.4. Water at roughly 80 reflects nearly 80 percent of a radar signal; hydrocarbons around 2 reflect only 2-4 percent (NotebookLM, Level Measurement notebook, 2026).

Why do ultrasonic level sensors fail in foam or heavy vapor?

Foam absorbs and scatters the acoustic pulse instead of reflecting it, and heavy vapor changes the speed of sound the pulse travels through, throwing off the time-of-flight math. Both are documented failure modes with no reliable fix short of switching to radar or guided-wave radar (NotebookLM, Level Measurement notebook, 2026).

What does API 2350 require for tank overfill level instrumentation?

API 2350 requires an Overfill Prevention Process with defined alarm levels, and it bars the sensor used for the High-High alarm from also being used for routine filling. Unattended, Category 3 tanks need a fully independent automated overfill prevention system on top of the primary level instrument (NotebookLM, Level Measurement notebook, 2026).

Which level measurement technology has the lowest total cost of ownership?

Non-contact radar and guided-wave radar carry the lowest five-year total cost of ownership because they have no moving parts and need almost no recalibration, even though they cost more upfront than float or ultrasonic sensors. Ultrasonic and float switches carry the highest ongoing OPEX from cleaning, recalibration, and mechanical wear (NotebookLM, Level Measurement notebook, 2026).

Conclusion

No single level technology wins every tank. Float switches and hydrostatic sensors are still the right call on clean, budget-constrained jobs. Ultrasonic earns its keep on stable atmospheric tanks. Radar and guided-wave radar own the harsh-condition, high-reliability end of the spectrum, at a price.

Start with the medium, not the sensor: dielectric constant, foam or vapor exposure, and vessel pressure narrow the field before cost ever enters the decision. If your tank falls into the radar-versus-ultrasonic gray zone specifically, our radar vs ultrasonic level measurement guide walks through that exact tradeoff in more depth.