Capacitance Level Transmitter

What is a Capacitance Level Transmitter

A capacitance level transmitter is a contacting level instrument that infers material level from a change in electrical capacitance. The sensing probe, usually a rod or a flexible cable lowered into the vessel, forms one plate of a capacitor. The metallic vessel wall, or a concentric reference tube on non-metallic vessels, forms the second plate. The process material between them is the dielectric. As level rises, the dielectric between the plates changes, the capacitance changes proportionally, and the electronics translate that into a continuous 4-20 mA or digital output. The same physics also serves point-level switches, where the electronics simply trip a relay when capacitance crosses a threshold.

The governing relationship is the parallel-plate capacitor equation, C equals k times epsilon-zero times A divided by d, where C is capacitance, k is the relative dielectric constant of the material between the plates, epsilon-zero is the permittivity of free space at 8.85 times ten to the minus twelve farads per metre, A is the effective plate area, and d is the separation. In a real probe the geometry is cylindrical rather than flat, but the principle is identical: the probe area and spacing are fixed by mechanical design, so the only variable that moves the capacitance during operation is the height of high-dielectric product covering the probe.

Because the dielectric constant of air and most vapors is close to 1, while water sits near 80, light hydrocarbons near 2, and many powders between 1.5 and 4, the contrast between covered and uncovered probe is what the instrument resolves. The transmitter does not measure capacitance with a DC bridge; it applies a fixed radio-frequency excitation, typically in the tens to hundreds of kilohertz range, which is why the technology is also called RF capacitance level measurement. The radio-frequency drive keeps the measurement stable against the bulk resistance of the medium and allows the more advanced admittance variant described in Chapter 2.

Capacitance level measurement is one of the oldest electronic level transmitter techniques, predating radar and ultrasonic in process plants, and it remains widely deployed because it has no moving parts, tolerates high temperature and pressure, and works on conductive media and foams that confuse non-contact methods. AMETEK Drexelbrook is historically associated with commercialising and refining the radio-frequency admittance form of the technology, and Endress+Hauser, Vega, KROHNE, Emerson, and Magnetrol all offer capacitance or admittance level lines today. The technology spans simple two-wire transmitters on small tanks up to high-temperature insulated probes rated beyond 200 degrees Celsius.

It is worth being clear about where capacitance sits among contacting and non-contacting level methods. Non-contacting techniques, such as through-air radar and ultrasonic, never touch the medium and are ideal where contact is undesirable, but they can be defeated by heavy vapor, dust, foam, and turbulence. Contacting techniques, including capacitance, guided wave radar, hydrostatic pressure, and float displacement, put hardware into the process and gain immunity to those airborne effects in exchange for wetted-part compatibility and cleaning concerns. Capacitance occupies a specific niche within the contacting group: it is comparatively inexpensive, mechanically simple, and uniquely good at conductive and coating-prone liquids when built as RF admittance, but it depends on the dielectric or conductivity of the medium in a way that radar and hydrostatic methods do not, which is the trade-off the rest of this guide quantifies.

Four engineering realities decide whether capacitance is the right choice: the dielectric constant of the medium and how stable it is, whether the medium is conductive or non-conductive, whether the medium coats the probe, and the temperature and pressure at the process connection. Misjudging any one of these is the usual cause of a capacitance loop that reads erratically or drifts. The chapters that follow address each in turn so the decision can be made on data rather than vendor preference.

Capacitance and RF Admittance Variants

Within this single technology there are two electrically distinct measurement schemes, plain RF capacitance and RF admittance, plus a split between continuous transmitters and point-level switches. The distinction between capacitance and admittance is the most consequential selection decision, because it determines whether the instrument can survive coating and conductive media. The table below summarises the variants and where each fits.

RF capacitance measures only the capacitive component between probe and wall. It is simple and inexpensive, but it has a fundamental weakness: a conductive coating that bridges the probe to the vessel wall is electrically indistinguishable from real product. The result is a false high reading, a stuck output, or premature switch trips. Plain capacitance is therefore best limited to clean, non-coating, non-conductive media where the probe stays dry above the true level.

RF admittance is the engineered answer to that weakness. Admittance is the reciprocal of impedance and combines a resistive part, conductance, with a reactive part, susceptance. By driving a short guard section of the probe near the process connection to the same potential as the active element, the instrument shunts the leakage current that flows through any coating straight back to ground before it reaches the measuring circuit. The measurement then responds only to the genuine level beyond the guarded zone. AMETEK Drexelbrook markets this driven-guard coating-rejection scheme under the Cote-Shield name on its Universal series, and the same principle appears under other vendor names. RF admittance is consequently the standard choice for latex, glue, sugar syrups, slurries, and other build-up-forming products.

It helps to be precise about why admittance survives where capacitance fails. A dry probe in vapor presents a near-pure capacitance to ground. When a conductive film of product wets the probe and reaches up the wall, it adds a parallel conductance path: current can now leak through the film even where there is no real product. A capacitance-only circuit cannot separate that leakage conductance from the capacitance it is trying to read, so the film inflates the apparent level. An admittance circuit measures both terms but uses the driven guard to drain the film current to ground at the top of the probe, so only the admittance of the genuinely covered length below the guard is reported. This is the electrical reason coating rejection is a property of admittance instruments and not of plain capacitance ones, and it is why a coating-prone duty should never be ordered as plain RF capacitance to save cost.

The second axis is continuous versus point level. A continuous transmitter reads the analog capacitance over the full insertion length and outputs a proportional level signal, suitable for inventory and closed-loop control. A point-level switch ignores the analog value and only changes a relay or transistor state when the probe is covered or uncovered at one height, used for high-high alarms, low-low pump protection, and overfill prevention. Both forms can be built on either the plain capacitance or the admittance circuit, so a project specification should state both the variant and whether continuous or point output is required to avoid an ambiguous order. For overfill-prevention switches in particular, the functional-safety rating (SIL capability) and the fail-safe direction of the relay must also be stated, because a point switch is frequently the last line of protection against a tank overflow.

Probe Configurations and Mounting

The probe is where most of the engineering lives. Its geometry, insulation, and material decide the measurable range, the temperature and pressure ceiling, and the chemical compatibility with the medium. Four configuration choices recur across vendors: rod versus cable, insulated versus bare, the wetted material, and the mounting and ground-reference scheme. The table below compares the two dominant probe forms before the text examines each choice.

Rod versus cable. A rigid rod probe is stiff, self-supporting, and easy to clean, which suits shorter vessels, sumps, and bypass chambers; for example the Endress+Hauser Liquicap FMI51 rod covers measuring ranges from about 0.1 to 4.0 metres. For tall tanks and silos, a flexible rope or cable probe, such as the companion Liquicap FMI52, spans much greater heights and is anchored or weighted at the bottom to keep it tensioned and centred. The cable form trades the rod's mechanical simplicity for reach.

Insulated versus bare. For conductive liquids the probe is jacketed in PTFE or PFA insulation. The insulation itself sets a fixed capacitance and the conductive liquid behaves as the outer plate, which is why an insulated probe in a conductive medium becomes largely independent of the medium dielectric constant. Endress+Hauser specifies this independence once liquid conductivity exceeds roughly 100 microsiemens per centimetre, so the same probe measures many different conductive liquids without recalibration. A bare metal probe is cheaper and avoids insulation wear, but it relies on the medium dielectric constant and so suits non-conductive, clean service only.

Wetted material. The wetted metal parts are most commonly 316L stainless steel, with PTFE or PFA as the insulation, an arrangement that handles water, acids, solvents, and many process chemicals. Aggressive or abrasive media may call for alternative alloys or a fully fluoropolymer-jacketed probe. Hygienic versions for food, beverage, and pharmaceutical service add electropolished surfaces and approvals such as 3-A and EHEDG, both offered on the Liquicap line.

Mounting and ground reference. The vessel wall completes the capacitor, so a conductive, well-grounded metal tank is the simplest installation: top-mount the probe and the wall is the reference. In non-metallic tanks, or where the wall is too far from the probe, a concentric ground tube or a separate ground rod must be supplied so a defined second plate exists. The probe should not touch the wall, agitator, or internal structures, and turbulent or splashing zones near agitators and inlets should be avoided or shielded, since stray capacitance from nearby metal and from product splashing changes the reading.

Dielectric Constant and Media Behaviour

The single most important property in capacitance level selection is the relative dielectric constant of the medium, often written as the dielectric constant k or the relative permittivity epsilon-r. It determines how much the capacitance changes per unit of level, and therefore whether the measurement has enough signal to be stable. The behaviour splits sharply between conductive and non-conductive media, and both depend on the value being stable across the operating temperature. The table below lists representative dielectric constants used in initial sizing.

Conductive media. When the liquid conducts, the bare metal probe would short to the medium, so the probe must be insulated. The conductive liquid then acts as the outer capacitor plate at the insulation surface, and the measured capacitance is governed by the insulation geometry rather than by the medium dielectric constant. This is the most robust case: as noted above, Endress+Hauser quotes dielectric-constant independence once conductivity exceeds about 100 microsiemens per centimetre, with a linearity error below 0.5 percent of range for conductive liquids on the insulated Liquicap rod. Where that conductivity floor is uncertain it can be confirmed with a conductivity meter. Water, brines, acids, bases, and most aqueous process streams fall in this category.

Non-conductive media. Oils, fuels, solvents, and many powders do not conduct, so the measurement depends directly on the medium dielectric constant through the parallel-plate relationship. Here the practical problem is low contrast: a hydrocarbon at k near 2 produces only a small capacitance change over the vapor baseline near 1, which limits resolution and makes the reading sensitive to any drift in the dielectric value. A common practical floor for stable capacitance measurement is a relative dielectric constant of roughly 1.5 to 2.0; below that, guided wave radar with low-dielectric processing is the better technology.

Temperature and composition stability. Dielectric constant is not a fixed number; it falls as temperature rises and shifts with composition, water content, and dissolved gas. A loop calibrated on cold product can read low when the same product is hot, and an oil whose water content varies will show an apparent level offset. For non-conductive service the engineer should obtain the dielectric constant at the actual operating temperature and bound its variation; for conductive insulated service this sensitivity largely disappears, which is one reason insulated-probe conductive applications are the most successful use of the technology.

Two further media effects matter. Coating on the probe adds parasitic capacitance or, if conductive, bridges the probe and is read as level, which is precisely why RF admittance with a driven guard exists. Buildup of solids in bins behaves similarly: bridging or rat-holing changes the effective dielectric path. For sticky, conductive, or build-up-prone media the admittance variant from Chapter 2 should be specified rather than plain capacitance, regardless of the nominal dielectric constant.

Key Specification Parameters

Reading the datasheet correctly prevents most field failures. A capacitance level transmitter datasheet typically lists a dozen or more parameters, but a smaller set drives the selection: measuring range, accuracy and repeatability, process temperature, process pressure, probe and insulation material, dielectric or conductivity requirement, output signal and supply, and hazardous-area approvals. Each is explained below with representative values from published datasheets.

Measuring range. Continuous capacitance probes commonly span from about 0.1 metre on a short rod up to roughly 25 metres on a long cable probe; the Liquicap FMI51 insulated rod, for example, is specified from 0.1 to 4.0 metres. The range is bounded by probe mechanical stability and by the signal-to-noise of the smallest resolvable capacitance change, so very long bare probes in low-dielectric media are the least favourable case.

Accuracy and repeatability. For an insulated rod in conductive liquid, a published linearity error is below 0.5 percent of the measured range with repeatability around 0.1 percent, as quoted for the Liquicap FMI51. Simpler or non-conductive applications are coarser, often near 1 percent of range, because they ride on a stable dielectric constant. Capacitance does not reach the few-millimetre absolute accuracy of guided wave radar, which specifies about 3 millimetres on the Rosemount 5300; the trade-off is capacitance tolerance of coating, foam, and conductive media that radar may find harder.

Process temperature and pressure. High-temperature insulated capacitance probes are a genuine strength of the technology. The Liquicap FMI51 is rated for a process temperature from minus 80 to plus 200 degrees Celsius and a process pressure from vacuum up to 100 bar (1450 psi), with a PTFE and PFA insulation system on 316L wetted parts. These figures should be read at the process connection, and the electronics-housing ambient limit is a separate specification.

Output, supply, and protocol. Continuous transmitters output 4-20 mA, almost always with HART overlaid for configuration and diagnostics, and some vendors also offer PFM or fieldbus options; the Liquicap line is a two-wire 4-20 mA HART loop powered at roughly 12 to 36 V DC. Point-level versions output a relay, transistor, or PFM switch state instead of an analog level. The output choice should match the host PLC or DCS input cards and the plant signalling standard.

Two remaining parameters complete a responsible specification:

• Dielectric or conductivity requirement: for insulated probes, the conductivity floor (about 100 microsiemens per centimetre for dielectric-constant independence); for bare probes, the minimum dielectric constant (roughly 1.5 to 2.0) and its temperature stability.
• Hazardous-area and hygienic approvals: ATEX, IECEx, FM, CSA, and NEPSI for explosive atmospheres per the IEC 60079 series, plus 3-A and EHEDG for sanitary service and SIL functional-safety ratings where the loop is part of a safety function.

A recurring datasheet pitfall is confusing the reference basis of the accuracy figure. A linearity error expressed as a percentage of the measured range behaves very differently from one expressed as a percentage of full scale on a long probe, and a repeatability figure does not bound absolute accuracy on its own. Following ISA-51.1 terminology and the IEC 61298 evaluation conventions, linearity, hysteresis, and repeatability are separate properties and should not be silently summed into a single headline number. When two quotations appear to differ on accuracy, the first check is whether they share the same reference basis and the same test conditions, because a 0.5 percent of range figure on a conductive insulated probe and a 1 percent figure on a non-conductive bare probe describe genuinely different measurement situations rather than simply better or worse hardware.

Finally, confirm the electronics-housing rating and environmental robustness alongside the wetted specification. Ingress protection to IP66 or IP67, the ambient temperature window for the housing as distinct from the process temperature at the probe, and electromagnetic-compatibility compliance to the IEC 61326 series for measurement and control equipment all bear on field reliability. A probe that is mechanically perfect for the medium can still fail in service if the housing is undersized for a washdown area or an outdoor installation, so these chrome specifications deserve the same scrutiny as the sensing parameters.

Selection Decision Factors

To turn the preceding five chapters into a purchase, follow the decision sequence below. Most capacitance level failures trace to skipping the media-property steps and jumping straight to mechanical dimensions, so the order matters. These eight steps double as a fixed RFQ template.

1. Characterise the medium first: is it conductive or non-conductive, what is its dielectric constant at operating temperature, and how much does that value vary. This single step decides whether capacitance is even viable and whether an insulated or bare probe is required.
2. Coating and buildup: if the medium is sticky, conductive, or build-up-forming, specify RF admittance with a driven guard such as Cote-Shield rather than plain capacitance, otherwise the loop will read false high.
3. Continuous or point level: decide whether you need a proportional 4-20 mA level for inventory and control or only a threshold switch for alarm and pump protection, and state it explicitly on the order.
4. Range and probe form: derive the insertion length from the vessel height, then choose a rigid rod for shorter vessels (up to about 4 metres) or a tensioned cable probe for tall tanks and silos.
5. Wetted material and insulation: match 316L, alternative alloy, or full fluoropolymer to the chemistry, and select PTFE or PFA insulation for conductive media; add electropolish, 3-A, and EHEDG for hygienic duty.
6. Temperature and pressure at the connection: confirm the process figures (for example, minus 80 to plus 200 degrees Celsius and vacuum to 100 bar on a high-temperature insulated probe) and the separate ambient limit for the housing.
7. Output, supply, and certifications: 4-20 mA HART is the default for continuous loops; confirm two-wire supply range, then add ATEX, IECEx, FM, CSA, or NEPSI per IEC 60079 and SIL ratings where required.
8. Total cost of ownership: weigh purchase price against calibration effort, coating-related false trips on plain capacitance, and the cost of a wrong-technology choice. Where coating or low dielectric makes capacitance marginal, budget for guided wave radar instead.

One commonly overlooked dimension is serviceability: local spare-probe inventory, field calibration and ground-reference support, the availability of HART device descriptions for the host system, and clear guidance on minimum dielectric and conductivity so future media changes do not silently break the loop. AMETEK Drexelbrook, Endress+Hauser, Vega, KROHNE, Emerson, and Magnetrol maintain regional service and documentation, which is why they are dependable choices for plant-wide standardisation even when a cheaper local probe would pass the initial datasheet check.