How to Select a Pressure Sensor Core: A Complete Selection Guide

What You Are Actually Buying

When you buy a pressure sensor core, you are not buying a finished instrument. You are buying the sensing element and the mechanical boundary around it, and you are agreeing to build everything else yourself: the housing, the signal conditioning, the connector, the calibration, and the documentation that your own customer will eventually ask for.

That is a deliberate trade. A core gives you a high degree of freedom over the finished product, and a corresponding degree of responsibility. You decide what the product looks like, what it is called, and where it is sold. In exchange, every interface the core presents to the outside world becomes your problem to solve, and some of those problems are significantly easier to solve at the specification stage than after tooling exists.

This guide works through the decisions that define a core. For each one it explains what the choice actually changes, what happens downstream if the choice is wrong, and what a sensor manufacturer needs from you in order to quote accurately.

Pressure Sensor Core Types

Selection starts on two axes that are independent of each other: how the element senses pressure, and how it is ported and mounted. They are orthogonal. The same sensing family can be built as a single-port element or as a differential element, and it can be supplied in a threaded body or with a flush diaphragm. Treating them as one flat list of "types" is the first place buyers get confused, so they are separated here.

By sensing family

By port and mounting configuration

WTsensor builds cores across both axes. Diffused silicon elements cover measuring spans from −100 kPa to 100 MPa, with body diameters from Φ10 mm to Φ19 mm and wetted materials including SS316L, SS304, gold-plated, Hastelloy C, titanium and tantalum. Where the accuracy requirement is tighter, monocrystalline silicon elements are used instead. Differential cores are available in both families — a diffused silicon differential core at ±0.25 %FS, and a monocrystalline silicon differential core reaching ±0.075 %FS — so within the differential line the choice between them is an accuracy decision, not a constraint imposed by the port configuration.

The Eight Decisions at a Glance

These are ordered by how much they constrain everything after them. Decisions 1 to 3 define what the core has to measure. Decisions 4 and 5 define its electrical interface. Decisions 6 to 8 define its mechanical interface. Skipping ahead to the mechanical questions before the measurement questions are settled is an easily overlooked path to an expensive redesign.

Decision 1: Pressure Reference

A pressure reading is meaningless without knowing what it is measured against. This is one of the decisions that is difficult to correct later, because the reference is built into the element and the port arrangement, not into software.

The choice between gauge and sealed gauge is the one buyers often leave implicit, and the one that becomes a field problem if they do. A gauge element needs a path to atmosphere — through a vent tube, a cable, or a dedicated vent in the housing. If your finished product has no such path, or if the core will be potted or encapsulated, a gauge element will read something other than what you intended, and it will do so consistently enough to look like a calibration error rather than a reference error.

Sealed gauge and absolute are not the same thing. Sealed gauge uses an internal reference that is not a vacuum, so it reads zero at the pressure it was sealed at, while absolute reads the true pressure including the atmospheric component. For most industrial measurement the difference is small, but for altitude work, vacuum duty, or any application where the reading feeds a calculation, it matters.

Decision 2: Range and Overpressure

Two numbers matter, and buyers frequently specify only the first. The measuring range is the span the core is calibrated over. The overpressure rating, sometimes called burst pressure or proof pressure, is what the element survives without damage or permanent offset.

Selecting a range is a balance. A wide range gives headroom and protects against transients, but it spreads the accuracy over more span, so the same percentage figure represents a larger absolute error. A narrow range gives finer resolution at the working pressure, but leaves less margin for the events that happen during commissioning, cleaning, or a blocked line.

A practical approach is to size the range on the working pressure, allowing for normal fluctuation, and then check the overpressure rating against the highest pressure the element can see — including the pressure spike from a valve closing, a pump starting, or a line being isolated. Those events are usually brief and are frequently omitted from the specification. Where the spike is severe relative to the working pressure, a higher-range element with a correspondingly wider tolerance is often the more robust commercial choice than a better-specified narrow-range element that will be replaced under warranty.

Two further points belong here. First, range and accuracy are not independent: most suppliers quote accuracy as a percentage of full scale, so a 1 MPa range and a 100 MPa range with the same percentage figure are not equivalent in absolute terms. Second, if the process can be cleaned, flushed or sterilised, the peak pressure during that step may exceed normal operation by a wide margin, and it is worth stating explicitly rather than assuming the supplier will allow for it.

Decision 3: Accuracy and Thermal Compensation

Accuracy figures are the easiest part of a specification to compare incorrectly, because the headline number describes only one of several error sources, and the conditions it applies under are often stated in a footnote.

A pressure sensor accuracy specification normally combines several contributions: non-linearity, hysteresis, repeatability, zero offset, and span error. Some suppliers quote these separately, others quote a combined figure. The combined figure is more convenient and less informative, because two cores with the same combined number may behave quite differently in service depending on which term dominates.

Beyond the basic figure, three further specifications change the real-world result:

Together these are what pressure sensor temperature compensation is for. Compensation is applied over a stated temperature range — a compensated range, not the full operating range — and outside that range the specified temperature coefficients no longer describe the behaviour. Some elements are compensated over a narrow band around ambient, others over a range wide enough to cover outdoor and unheated installations. A core rated for a wide operating temperature and a core rated for a wide compensated temperature are not the same product.

This is worth stating plainly, because the distinction is easy to miss when reading a datasheet: When comparing pressure sensor accuracy specifications, always confirm whether the quoted value includes thermal effects or applies only at reference temperature. A figure described as "includes temperature effects" over a stated compensated range is directly comparable. A figure quoted at reference temperature only is the starting point, and the thermal terms have to be added on top to estimate field performance.

Long-term stability deserves the same treatment. A stability figure expressed per year compounds, and over a five- or ten-year service life the accumulated drift may exceed the initial accuracy several times over. Where the instrument is inaccessible, or where recalibration means taking a line out of service, the stability figure is the more important of the two numbers.

For reference: WTsensor core accuracy spans from ±0.5 %FS in the general-purpose range to ±0.075 %FS at the top of the monocrystalline silicon range, with long-term stability figures quoted per year for the higher-stability models.

Decision 4: Output Signal

The output determines how much work you do downstream. There is no answer that applies to every case, only a trade between integration effort, noise performance, and distance.

The decision follows from distance and environment. A core mounted next to its own signal-conditioning board can use a raw mV output, because the lead length is short and the conditioning can be designed for the specific element. A core at the end of several metres of cable in an electrically noisy plant needs either a conditioned output at the core or careful attention to shielding, grounding and the input stage of whatever reads it.

Raw mV output has a further consequence worth knowing before the specification is frozen: the output is proportional to the excitation, so the excitation and the amplifier have to be treated as one design. Any variation in excitation appears directly in the measured value unless the reading is ratiometric to that same excitation. This is why raw-element specifications state the excitation the output was measured at — without it, the millivolt figure has no meaning.

A further practical point: the output choice interacts with the accuracy discussion in Decision 3. A raw mV element that has been temperature compensated at the element is a different starting point from an uncompensated element, and the difference shows up as additional work on your board rather than as a different number on the core datasheet.

Decision 5: Electrical Connection and Supply

Two things are decided here, and they are usually specified together: how the core is wired, and how it is powered.

Excitation and supply

A raw mV element may be specified for constant-current excitation or for a defined voltage supply, and the two are not interchangeable — a circuit built for one will misread the other. Constant-current excitation is used because the bridge output is then largely independent of the absolute bridge resistance, which reduces the effect of temperature-induced resistance changes. Driving the same element from a constant voltage leaves that dependence in place, and the resulting error appears as a temperature effect on span.

The required excitation is set by the manufacturer for the specific element and is confirmed against the core datasheet. On the WTsensor diffused silicon core range, the standard supply options are 1.5 mA constant current, 1.5 mA constant current combined with a 10 V or 5 V constant voltage supply, 5 V DC constant voltage, and a 0.5–4.5 V ratiometric supply with an optional I²C digital output. The correct option depends on the element and on the output format chosen in Decision 4, and the value quoted for one does not carry over to another.

Conditioned elements instead accept a defined voltage supply, and the output is read directly. In that case the specification to confirm is the supply range the element tolerates, including the behaviour below the minimum supply — some elements produce a misleading output rather than an obviously invalid one when the supply sags.

Wiring

Wiring is where a specification that looked complete turns out not to be. The number of wires, the colour code, the lead length, the lead material, the pin pitch and the pin plating all have to be fixed, and any of them may need to change for an OEM variant. Gold plating on the pins is the usual choice where long-term contact resistance matters or where the assembly is exposed to humidity during storage.

Two details cause a disproportionate share of integration problems. The first is lead length: a specification that names a wire count but not a length will be built to a nominal value, and if your assembly needs a different one, the change is easy to make at the factory and awkward to make on the line. The second is the wiring convention for differential and pressure-reference variants, where the sense of the output relative to the two ports has to be stated rather than assumed — a differential core wired the wrong way round produces a reading that is internally consistent and points the wrong direction.

Decision 6: Diameter and Mounting Form

This is where the specification meets the physical product, and where an otherwise ideal core turns out not to fit. Three dimensions are involved, and they constrain each other: the body diameter, the thread or fitting, and the sealing face.

Body diameter sets how much space the core needs inside your housing and, in many designs, how large the housing has to be. A smaller diameter element is easier to package, but usually carries a narrower measuring range or a lower overpressure rating, because the diaphragm area available to act on the element is smaller. Common body diameters in the diffused silicon range run from Φ10 mm up to Φ19 mm, with options at Φ12.6 mm, Φ15 mm and Φ15.8 mm in between.

Two mounting details are frequently left out of a first specification and cause rework later. The first is the thread standard: a nominal size such as G1/4 is not unique once the standard, the sealing form and the thread length are taken into account, and the same nominal size can be supplied in forms that do not interchange. The second is the wrenching or clamping feature — whether the assembly is tightened by a hex, a spanner flat or a dedicated tool, and whether that feature survives the tightening torque in a production environment.

If the core is to be screwed into a manifold or a machined block, it is worth confirming that the mating thread and the sealing face are defined by the same drawing, rather than assuming compatibility between two specifications that each look standard on their own.

Decision 7: Wetted Materials

The wetted materials are the diaphragm, the body or the housing that contacts the process, and any seal in the wetted path. This decision is driven entirely by the medium, and getting it wrong produces failures that appear months later and are expensive to diagnose.

Two practical rules apply here. The first is that compatibility is a property of the medium, the concentration, the temperature and the material taken together — not of the material alone. A grade that performs well at ambient temperature in a dilute solution may behave differently at elevated temperature or higher concentration, so the operating conditions belong in the specification alongside the medium name.

The second is that the seal is often the limiting item rather than the metal. An O-ring in the wetted path may have a narrower temperature range, or a different chemical compatibility, than the diaphragm and body around it, and it is the component that ages. If the medium is aggressive, the sealing material needs the same scrutiny as the diaphragm.

Decision 8: Sealing and Integration

This decision determines whether the core becomes a reliable product or a warranty problem, and it is the one often treated as a detail to be settled at assembly time.

A core presents several joints that all have to remain leak-tight: the joint to the process, the joint between the element and its body, and the electrical feedthrough that carries the signal out through the housing wall. The first is a sealing decision you make, using a gasket, an O-ring or a weld. The second is normally made at the factory. The third sits in between, and it is the one buyers ask about least and reliability engineers care about most, because a leak path through the electrical feedthrough puts the medium in contact with the electronics regardless of how well the process joint is made.

The temperature range of the finished assembly is also set here rather than by the element alone. The core may be rated over a wide range, while the seal, the potting compound or the cable has a narrower one, and the assembly rating is the narrower of the two. Where the finished product is qualified as a whole, the qualification is against the assembly's rating, not the core's.

Integration choices that belong in the specification include:

• The mounting torque the assembly is designed for, and whether the core is intended to be installed with a torque wrench or a fixture.

• Whether the joint is to be welded, brazed, crimped or sealed with an elastomer, and if welded, whether the heat input is acceptable to the element.

• The leak-test method and acceptance criterion for the finished assembly — helium leak testing, pressure decay, or bubble test — and at what stage it is performed.

• Whether the cable or lead exit is strain-relieved, and how much pull the assembly is expected to survive.

• Whether the core will be potted, overmoulded or encapsulated, and whether the process temperature involved is acceptable to the element.

The last item deserves emphasis, because potting and overmoulding are the steps where a correctly specified core is often damaged. The temperature and pressure of the moulding process, and the shrinkage of the compound as it cures, both act on the element. A core that performs correctly on the bench may show a permanent offset after encapsulation, and the cause is the process rather than the core. If the finished product involves either step, it belongs in the original enquiry, so the manufacturer can supply an element suited to it.

Core, Module, or Complete Transmitter?

A core is one of several forms in which a sensing element is supplied, and the three options carry very different amounts of work.

The transition between these forms is a commercial decision as much as a technical one. A core gives the lowest unit cost and the greatest control over the finished product, and it is the right choice where you have the engineering resource to own the signal chain and the calibration. A module or a complete transmitter costs more per unit and reduces the work on your side, which is the right choice where volume does not justify the engineering, where lead time matters more than unit cost, or where the finished product has to carry a certification you would rather inherit than obtain yourself.

A frequent and easily avoided mistake is to specify a core for a product that has no resource to develop the signal chain around it. The result is a longer development cycle than a module would have needed, and a higher total cost than the more expensive component. It is worth deciding which of the three forms you are buying before working through Decisions 1 to 8, because the answers change depending on the answer to this question.

Where Pressure Sensor Cores Are Used

Core selection is not specific to one industry, and it is worth seeing the range of duties, because requirements from one area often transfer to another.

Two patterns are worth noticing in that table. The first is that differential measurement appears across several areas for the same underlying reason — a pressure difference is a direct indication of flow across a restriction or of the state of a filter — so the differential decisions carry across industries more than the others. The second is that the hygienic and aggressive-media applications are the ones where the mechanical and material decisions, rather than the measurement decisions, determine whether the product succeeds.

Worked Specification

A specification is easier to write against an example. The following describes a core for an OEM pressure transmitter intended for general industrial use, in a housing of the customer's own design.

Several entries in this table are conditions rather than absolute values, and that is deliberate. The overpressure requirement is expressed as a multiple that has to be confirmed against the supplier's datasheet, because the ratio varies between models. The excitation is expressed as "as specified by the manufacturer", because the correct value depends on the element and the output format. Writing a specification this way makes the dependencies visible instead of hiding them behind numbers that look precise but were not verified.

Common Selection Mistakes

The mistakes below appear repeatedly in enquiries, and most are corrected cheaply at the specification stage and expensively afterwards.

Selecting a range that is too wide "for safety". A wider range protects against transients, but if accuracy is quoted as a percentage of full scale, the absolute error grows with it. A core with the right range and a suitable overpressure rating often serves better than an oversized one.

Treating gauge and sealed gauge as interchangeable. They are not, and the difference appears as a consistent offset rather than as an obvious fault. If the finished product has no vent path, gauge is the wrong reference regardless of how the specification reads.

Comparing accuracy figures without comparing their conditions. A figure quoted at reference temperature and a figure that includes thermal effects over a compensated range are not comparable, even when the numbers look similar.

Leaving the electrical connection partly undefined. A wire count without a length, a colour code or a pin plating is an incomplete specification, and the gap is filled at the factory by a nominal value that may not match the assembly.

Specifying the wetted material without the medium conditions. Compatibility depends on concentration and temperature as well as on the material, so a material name on its own does not establish that the core is suitable.

Omitting the assembly process from the enquiry. Welding, potting and overmoulding all act on the element. Where they are part of the finished product, they change which element is appropriate.

Confusing an operating temperature range with a compensated one. A wide operating range describes what the element survives. The compensated range describes where the specified accuracy holds. They are different figures.

Ordering a core for a product with no signal-chain resource. This produces a longer development cycle than specifying a module would have needed, at a higher total cost.

What to Send a Sensor Manufacturer: RFQ Checklist

A complete enquiry produces an accurate quotation quickly. An incomplete one produces either a request for clarification or a quotation with assumptions built in, and the assumptions are where the surprises come from. Quotation delays often come from missing inputs rather than from the manufacturing side.

Measurement

• Pressure reference: gauge, sealed gauge, absolute or differential

• Measuring range, in the units you actually work in, with the working pressure stated separately if it differs from the range

• Required overpressure or proof pressure, if you know it

• Accuracy requirement, and whether it includes thermal effects

• Compensated temperature range, and operating temperature range

• Long-term stability expectation, if the recalibration interval matters

Electrical

• Output format, and the value it refers to

• Excitation or supply, if you have a preference

• Number of wires, lead length, colour code, pin plating and pitch

• Whether the assembly must carry any certification, and which

Mechanical

• Body diameter and any envelope constraints

• Mounting form and thread standard, with the sealing method

• Wetted materials, including the seal

• Whether the core is welded, brazed, potted or overmoulded into the assembly

• Leak-test method and acceptance criterion, if defined

Commercial

• Annual or first-order quantity, and whether the requirement is expected to grow

• Marking, labelling and traceability requirements

• Documentation required with the shipment, such as a test certificate or a material declaration

• Required delivery date and destination

• Any sample or qualification process the order has to pass before volume supply

If some of these are not yet decided, saying so is more useful than leaving them out. A manufacturer who knows which items are open can propose options and quote them as alternatives, which is a faster route to a decision than a specification that appears complete but contains assumptions on both sides.

Where to Go Next

The decisions above connect to more detailed material elsewhere on this site. Depending on which one is currently open, the following pages go deeper.

Sensing technologies

• Diffused silicon pressure sensors — the mainstream family, and the range of bodies, ranges and diaphragm materials available in it.

• Monocrystalline silicon pressure sensors — where the tighter accuracy and stability ceiling applies, and what it costs.

• Internal structure of a diffused silicon pressure sensor — what sits behind the diaphragm, part by part.

• Piezoresistive pressure sensors: diffused silicon vs monocrystalline silicon — how the two silicon families differ in principle and in application.

Differential measurement

• Monocrystalline silicon differential pressure sensors — the higher-accuracy end of the two-port range.

• Differential pressure transmitters — the finished-instrument side of the same measurement.

Finished instruments

• Pressure transmitters — the range built around these cores, where the integration has already been done.

• Level sensors and transmitters — hydrostatic level measurement, which follows Decisions 1 and 2 closely.

• Submersible level transmitter selection guide — the full selection process applied to a specific product family.

Construction, sealing and the sensing boundary

• Glass-to-metal seal manufacturer guide — the hermetic feedthrough that keeps the electrical side dry.

• Hermetic sealing — sealed connectors, headers and feedthroughs for instrument assemblies.

• Metal bellows guide — the separate sensing element used where a diaphragm is not the right boundary.

About WTsensor

WTsensor (Nanjing Wotian Technology) has manufactured pressure sensors since 2005. Its core business is the sensing element itself: diffused silicon and glass micro-fused elements supplied both as standalone cores for OEM integration and built into its own finished instrument range. Quality systems are certified to ISO 9001 and IATF 16949, and the core range is covered by CE, RoHS, REACH and EMC documentation.

Core options span gauge, absolute, sealed-gauge and differential references, measuring spans from −100 kPa to 100 MPa, body diameters from Φ10 mm to Φ19 mm, and wetted materials including SS316L, SS304, gold-plated, Hastelloy C, titanium and tantalum. Differential cores are available in both diffused silicon and monocrystalline silicon constructions, the latter reaching ±0.075 %FS. Because these are elements rather than finished instruments, most requirements are built to order and configured against the customer's own product.

For OEMs looking for a pressure sensor core manufacturer, WTsensor can configure core diameter, output, excitation, wiring, materials and marking around the finished product requirements. Final specifications are confirmed against the current datasheet at quotation.

FAQ

What is the difference between a pressure sensor core and a pressure transmitter?

A core is the sensing element and its mechanical boundary, with a raw or lightly conditioned output. A transmitter is a finished instrument that adds signal conditioning, a defined output, an electrical connection and a housing. Choosing between them is a question of how much of the signal chain you intend to own.

Can I use a gauge core in a sealed housing?

A gauge element needs a path to atmosphere to work as intended. In a sealed housing with no vent, it will read relative to whatever pressure was trapped inside at assembly, which appears as a constant offset. Where the housing is sealed, a sealed-gauge or absolute element is the appropriate choice.

Should I choose diffused silicon or monocrystalline silicon?

Both are silicon piezoresistive. In practice the choice usually follows accuracy and stability requirements: diffused silicon covers the mainstream range, while monocrystalline silicon is selected where the reference accuracy and long-term stability need to be tighter. Within the differential range both are available, so the choice there is an accuracy decision rather than a question of what the port configuration allows.

How do I compare two accuracy specifications that quote the same number?

Establish what each figure covers and under what conditions. A combined figure that includes thermal effects over a stated compensated range is directly comparable with another figure of the same kind. A figure quoted at reference temperature only is not, because the thermal contributions have to be added to estimate performance in service.

What excitation does a raw mV core need?

The excitation is specified by the manufacturer for the individual element, and the output figure is only meaningful at the excitation it was measured at. Constant-current excitation is used where the output should be largely independent of bridge resistance. Confirm the requirement against the datasheet for the specific core rather than assuming a value carries across models.

How long does a typical core order take?

Lead time depends on the configuration, the quantity and whether the requirement is a standard model or a customer-specific variant. Because much of the core range is built to order, it is worth including the required delivery date in the enquiry so it can be quoted against directly.