Gauge Pressure Transmitter

What Is a Gauge Pressure Transmitter?

The measured quantity

Gauge pressure is the difference between absolute process pressure and local atmospheric pressure: p(gauge) equals p(absolute) minus p(atmosphere). At sea level the atmosphere presses at about 101.325 kPa, but the actual figure swings between roughly 97 and 104 kPa with weather and altitude. A gauge transmitter subtracts that moving reference automatically by venting the low side of the cell, so the instrument reads 0 kPa with the process open to air, regardless of whether the plant sits at sea level or 1,800 m up a mesa. That self canceling behavior is the entire point of the reference choice.

The output chain is standardized. A two wire loop powers the transmitter at 24 V DC and carries the measurement as a proportional current: 4 mA at the lower range value, 20 mA at the upper range value, with pressure mapped linearly between them. A living zero at 4 mA lets the host distinguish a true zero reading from a broken wire or a dead loop, which is one reason current loops, not voltage signals, dominate process plants. The same pair can also carry digital data, which Chapter 4 treats as a protocol decision rather than an afterthought.

Gauge, absolute, and differential: picking the reference

The reference choice is made once per measurement point, and it is usually obvious from the process. Above atmosphere, with pressures quoted in barg or psig, the gauge instrument matches every drawing, every pressure gauge on the skid, and every acceptance record already in the file. Deep sub-atmospheric duty is where the absolute pressure transmitter earns its price: a vacuum column running at 25 kPa absolute would show a daily wobble of 2 to 3 kPa on a vented sensor as weather fronts pass, several percent of the reading, while the sealed reference ignores the weather entirely.

The differential pressure transmitter is the third sibling and the busiest of the three. It compares two live process ports: across an orifice plate or venturi to compute flow, from the bottom to the top of a pressurized vessel to infer level, or across a filter housing to time cartridge changes. In flow duty the primary element and the differential cell together form the flow meter of record for the loop. Its two port architecture, three-valve manifold hardware, and equalizing practice are close cousins to gauge installation technique, which is why Chapter 6 borrows from them freely.

When the selection question arises

The gauge transmitter question shows up in three situations. First, a new skid or line needs pressure into the control system, typically one transmitter per pump discharge, header, and vessel. Second, a plant replaces local mechanical indication with signal: a field of dial instruments becomes a wired network feeding the DCS, adding trending, alarms, and history that a bourdon tube cannot hold. Third, an aging loop is upgraded: a purely mechanical pressure switch trip becomes a transmitter plus a software limit, buying adjustable setpoints and pre-alarm bands without touching the process piping.

The comparison with a switch deserves one line of honesty. A pressure switch is cheaper, simpler, and fully satisfies a trip function, but it reports one bit of information. A transmitter reports continuously, supports a PID controller acting on the measured value, and lets operations see a pump walking up its curve hours before an alarm sounds. Most plants end up with both: a transmitter for control and indication, and an independent switch or relief valve for protection, deliberately on separate taps so one plugged impulse line cannot blind both.

What you are actually deciding

Five decisions consume nearly all of the engineering time and all of the schedule risk. Range: the upper range limit and the calibrated span, sized so normal operation sits between 25 and 75 percent of span. Accuracy: 0.075 percent of span is the standard class, 0.04 percent is premium, and most loops do not need premium. Wetted parts: 316L stainless covers the majority of duties, and alloy C-276 covers sour and hot chloride service. Signal: plain 4 to 20 mA, HART 7, or a fieldbus, decided by the host and the maintenance bench. Approvals: the area classification drives Ex d or Ex ia construction. The rest of this page works each decision in turn.

Working Principle and Sensing Types

The piezoresistive oil-filled cell

The dominant architecture isolates the sensing element from the process with a thin metal diaphragm and a hydraulic link. Process pressure deflects a corrosion resistant diaphragm by a few hundredths of a millimeter, and incompressible fill fluid, almost always silicone oil, carries that deflection to a silicon chip etched with piezoresistors. The resistors are wired into a Wheatstone bridge, so diaphragm strain arrives as a small differential voltage, on the order of tens of millivolts at full span. The transmitter electronics amplify that signal, digitize it, linearize it against factory characterization, and temperature compensate it before driving the loop.

The oil fill buys three things. Chemical isolation: only the diaphragm sees the process, so the silicon chip never meets sour gas, chlorides, or slurry. Overload protection: a stop behind the diaphragm limits travel when the process spikes, letting the cell survive 2 to 3 times its upper range limit without measurable permanent shift. Manufacturing economy: one silicon design, tuned by fill volume and diaphragm thickness, covers a family of spans from kilopascals to tens of megapascals, which is why the oil-filled piezoresistive cell is the volume leader across every major catalog.

Capacitance and bonded strain gauge alternatives

Capacitance cells place a conductive diaphragm between fixed plates and measure the capacitance shift as it moves. They excel at low spans, down to a few hundred pascals, with strong linearity and low hysteresis, which is why the low span members of differential families often use them. Against piezoresistive parts they give up some shock tolerance and add internal mass, and for gauge spans above roughly 1 MPa they hold no clear advantage, so they appear less often on this page's duty.

Bonded foil strain gauges, resistive elements bonded to the back of a machined stainless diaphragm, are the older and cheaper route, and they still ship in economy industrial transmitters at 0.2 to 0.5 percent of span accuracy. Creep in the adhesive and the bonded layers shows up as long term drift, so the calibration calendar matters more for these cells. For a utility air header read once a day, that can be a fair trade; for a reactor or custody loop it is false economy, because recalibration labor swamps the purchase saving within two cycles.

When a catalog says piezoresistive without further detail, expect the oil isolated silicon cell. A pressure sensor built that way, in a compact stainless body with a small amplifier and a voltage or current output, is also the architecture behind the large mid market of compact units. The pressure transmitter differs from a bare sensor mainly in the electronics: configurable range, damping, self diagnostics, and a loop compliant output, packaged in a housing rated for the plant environment and the area classification.

Span, turndown, and the arithmetic of range setting

Range families are built around a sensor with a fixed upper range limit, then calibrated to whatever smaller span the application needs. A family rated at 100:1 turndown with a 0 to 25 MPa cell accepts calibrated spans down to 0 to 250 kPa. Setting range 0 to 1 MPa on the 0 to 25 MPa cell is a 25:1 turndown, comfortably inside the family but far enough in that the turndown terms in the error budget are no longer negligible.

Do the arithmetic explicitly. Reference accuracy of 0.075 percent of a 1 MPa span is plus or minus 0.75 kPa, which is excellent for most duties. But the datasheet adds a turndown effect and a zero stability term that grow as the calibrated span shrinks against the sensor limit; at 25:1, a realistic total performance is 0.1 to 0.2 percent of span, so plus or minus 1 to 2 kPa. Two rules fall out. Buy the smallest upper range limit that still covers your maximum pressure. And do not burn turndown for convenience, because each halving of span past about 10:1 costs measurable accuracy.

Digital ranging is free at commissioning. A HART communicator, or the local display option, rewrites the lower and upper range values in minutes without touching hardware. Plants exploit this for spare pooling: a handful of 0 to 25 MPa and 0 to 4 MPa cells, re-ranged on demand, can cover dozens of service points at the documented accuracy cost above. The discipline is to record the configured span and the turndown actually used in the instrument register, because the accuracy budget of a pooled spare is rarely the accuracy budget of the instrument it replaced.

The vent path

The gauge reference lives or dies on a small vent path from the low side cavity of the cell to atmosphere, usually a drilled passage in the housing closed by a vent plug or a breathable membrane. Open, the reading self zeros against the weather. Blocked by paint, an insect, or ice, the cavity seals and the instrument quietly becomes a pseudo absolute device: barometric shifts of 2 to 4 kPa now read as process changes, a drift that arrives with the weather map rather than with the process. Chapter 6 puts vent checking on the maintenance calendar for exactly this reason.

From sensor to signal

Inside the electronics, the bridge signal is amplified and digitized at 16 to 24 bits, corrected against an on-board temperature sensor using coefficients established at the factory, linearized, and damped with an adjustable time constant from near zero to tens of seconds, then converted back to current for the loop. Damping deserves attention: one second of damping smooths compressor discharge ripple nicely, while 30 seconds of damping on a trip input will hide a real excursion until after the fact. HART 7 superimposes 1200 baud FSK data on the same pair without disturbing the analog value, and it needs a minimum loop resistance near 230 ohms to develop the signal, a detail that matters when barrier and input card resistances are being totaled.

Proof, burst, and the pressure ladder

Catalogs quote a ladder of pressure limits, and misreading it causes the classic selection failure. Maximum working pressure is what the cell may see continuously, typically equal to the upper range limit. Proof pressure, commonly 2 to 3 times the upper range limit, is a temporary excursion the cell takes without permanent output shift beyond its accuracy band. Burst pressure, commonly 3 to 5 times the limit, is the structural failure envelope at reference temperature. A pump that can deadhead to 9 MPa against a 0 to 4 MPa cell sits above proof on paper: the correct answers are a higher range cell, a relief path, or an overpressure rated seal system, not optimism and a heavy damping setting.

Construction, Wetted Materials, and Duty Boundaries

Anatomy of the transmitter

A conventional two compartment transmitter separates the process from the wiring. On the process side sit the connection, usually 1/2 NPT or G1/2, the cell body carrying the isolation diaphragm and fill fluid, and the sensing module. The cast aluminum or 316 stainless housing keeps the amplifier compartment and the terminal compartment apart, so opening the electrical cover does not expose the cell side. An optional local display clips onto the terminal side and shows engineering units, loop current, and diagnostic codes where the technician actually stands. Everything on the wetted path is a corrosion decision; everything after it is an electrical one.

Wetted materials and where they stop

316L stainless is the default diaphragm and body material and covers water, steam, air, most hydrocarbons, and mild chemicals. It has two documented weaknesses. Warm chloride bearing brine above roughly 60 deg C attacks 316L with pitting and stress corrosion cracking. Wet hydrogen sulfide service is governed by NACE MR0175 / ISO 15156 material practice, and the routine answer is alloy C-276 diaphragms and wetted bodies. C-276 tolerates the sour and hot chloride envelope comfortably and typically adds 30 to 60 percent to the wetted parts cost; beyond C-276, tantalum liners and gold overlays handle aggressive chemical service, almost always built as remote seals rather than inline cells.

Hydrogen deserves a specific paragraph because it defeats simple stainless logic. Atomic hydrogen permeates thin diaphragms, dissolves into silicone fill fluid, and recombines as bubbles that shift the zero over days to weeks. Hydrogen duty is specified with gold plated or gold plated C-276 diaphragms and a fill fluid chosen for low permeation. It is an order code decision made when the instrument is bought, never a field retrofit, and a cell that has already soaked in hydrogen is a candidate for replacement rather than recalibration.

Fill fluid selection

The fill fluid is a quiet specification line with real consequences. Standard silicone oil spans about -40 to +120 deg C at the cell and suits nearly everything except strong oxidizers. Oxygen and wet chlorine service prohibit silicone, whose behavior in pure oxygen is a documented hazard, so those duties specify inert halocarbon fill together with cleaning and paperwork for oxygen service. High temperature silicone variants push the working ceiling toward 200 deg C for steam and hot oil, at some cost at the cold end of the range. The wrong fluid is not a nuisance: it is a reorder with the measurement loop down in the meantime.

Process connections

The connection list is short and standard driven: 1/2-14 NPT per ASME B1.20.1 as the North American default, G1/2 per ISO 228 for metric plants, 1/4 NPT on compact low pressure bodies, flanged interfaces per ASME B16.5 in class 150 through 1500 for seal and high pressure work, coplanar and planar modular faces that let one transmitter family share manifolds and seal hardware, and tri-clamp sanitary connections per ASME BPE for food and pharmaceutical lines. Two rules apply. Match the manifold and seal hardware family to the interface. And check that the instrument connection pressure rating, not just the sensor, covers the line rating.

Housings, ingress, and explosion protection

Electrical protection is a matching game between the housing and the area. For ingress, IP66 and IP67 per IEC 60529, or NEMA 4X per NEMA 250, are the working standards for outdoor and washdown mounting, with IP67 adding brief immersion tolerance. For explosion protection, flameproof Ex d per IEC 60079-1 contains an internal ignition inside a certified enclosure and demands certified cable glands. Intrinsic safety Ex ia per IEC 60079-11 limits loop energy below ignition thresholds and demands barriers and IS wiring discipline. Dust ignition protection Ex tb covers zone 21 and 22 powder handling. Enclosure material follows the site: low copper aluminum alloys are the economical Ex d choice, while 316 stainless wins offshore and on chloride laden coastal sites.

Duty boundaries: pulsation, heat, and clogging

Three process conditions retire standard installs, and each has a standard countermeasure. Pulsation from reciprocating pumps and compressors fatigues diaphragms and rattles the output; a piston or porous snubber at the tap, or a needle valve cracked open a quarter turn, drops the amplitude reaching the cell, while electronic damping only hides the ripple from the operator. Heat: above about 120 deg C at the diaphragm, standard silicone fill degrades and the cell drifts; steam duties take a pigtail siphon kept full of condensate, and hot duties take a standoff or up to 5 m of capillary to a transmitter mounted away from the source. Clogging: crystallizing, polymerizing, and solids bearing fluids do not belong on a narrow NPT port at all; a flush diaphragm seal on a flange, connected by capillary, moves the measurement to a wetted surface that sheds rather than traps media, at the price of fill fluid temperature effects on zero that Chapter 6 quantifies.

Key Specification Parameters

A pressure transmitter datasheet is a decode exercise: every line either sizes the cell, prices the loop, or sets the calibration calendar. The table below lists the ten lines that decide a gauge transmitter selection, with the catalog spread to expect across mainstream manufacturers and the selection consequence of each. The paragraphs after it work the three lines most often misread: accuracy, turndown, and drift.

Accuracy: percent of what, at what span

Accuracy quoted as a percent is meaningless until you ask percent of what. A 0.075 percent of span class on a 0 to 1 MPa calibrated span is plus or minus 0.75 kPa anywhere on the scale, because span referenced error is flat across the range. The 0.04 percent premium class cuts that to plus or minus 0.4 kPa, which matters only if the loop actually uses the band: a pump discharge alarm at 800 kPa plus or minus 10 does not. Pay for premium accuracy on custody transfer, tight ratio control, and safety relevant measurement, and spend the difference on manifolds and calibration discipline instead.

Turndown: the family rating is not free performance

The composite error after configuring a 100:1 rated cell at high turndown includes reference accuracy plus turndown effect plus zero stability over time, and each term grows as the configured span shrinks against the sensor limit. Practical guidance consistent with mainstream catalogs: stay at or below 10:1 when the measurement matters, treat 25:1 as the working limit for control loops, and reserve 100:1 for spare pooling and non critical indication. Wherever the configured turndown exceeds 10:1, commission with a two point verification and record the configured span in the instrument register.

Drift: the line that sizes the calendar

Drift, quoted as a percent of upper range limit per year, decides the calibration interval. Standard cells drift 0.1 to 0.25 percent of URL per year in the mainstream spread, so a 0.15 percent cell over a 25 MPa limit can accumulate tens of kilopascals of potential shift over a multi year interval, which is exactly why the 12 month default exists. Premium families publish combined long term stability figures an order better, which is what makes 3 to 5 year verification defensible outside regulated duties. Buy the drift figure you are willing to verify, not the one that reads best in the brochure.

Reference comparison at a glance

Before protocol, one comparison settles the most common early question: which reference does this point need? The table places the three siblings side by side on the attributes a specifier checks, including the signature symptom that appears when the wrong reference is installed.

Signal protocol options

Protocol choice locks wiring, cabinet hardware, and the maintenance toolset for the life of the plant, so it deserves a decision, not a habit. The table maps the five options a gauge transmitter catalog actually offers against the constraints that bite during design and commissioning.

On the host side, a 16 bit PLC analog input card resolves a 20 mA span into steps of about 0.3 microamp, roughly 0.002 percent of span, so the card is rarely the bottleneck; grounding, cable routing, and barrier quality usually are. Whether the PID controller executes in the PLC, in the DCS, or in a single loop controller in the field, the transmitter's job is identical: hold a linear, damped, trustworthy measurement so the control action stays sane. Fieldbus pays where segment infrastructure already exists or where device density makes home run wiring expensive; otherwise HART 7 over 4 to 20 mA is the default that every maintenance technician in the plant can already service.

Selection Decision Factors

Selection is a sequence, not a checklist race: each step prunes the option space for the next, and the order below stops the two classic mistakes, sizing span before overpressure and choosing alloys before chemistry. Work it top to bottom and the RFQ writes itself at the end.

1. Fix the span from the process, not the catalog. Size the calibrated span so normal operating pressure sits between 25 and 75 percent of span: enough headroom that excursions stay on scale, enough working band that control and trending use real resolution. A pump that walks between 600 and 900 kPa takes a 0 to 1.6 MPa span happily and wastes a 0 to 25 MPa cell.
2. List the overpressure events before the vendor does. Pump deadhead, valve slam, thermal expansion of a blocked liquid line, and hydrotest residue all reach the cell. Every credible event must sit below proof pressure, 2 to 3 times the upper range limit on mainstream cells. Anything above proof is a redesign input, not a tolerance.
3. Match wetted parts to the chemistry, temperature, and standards file. 316L by default; alloy C-276 for wet sour gas per NACE MR0175 practice and for warm chlorides; gold plated diaphragms where hydrogen permeates; inert fill with oxygen cleaning for oxidizer duty. Write the fluid in the RFQ, not just the alloy, and let vendors propose against it.
4. Distance the cell from heat. Above about 120 deg C at the diaphragm, budget a siphon, a standoff, or up to 5 m of capillary, and verify ratings at the cold end too: continental outdoor sites see -40 deg C ambient, and displays blank well before the electronics quit.
5. Let the area classification drive the enclosure concept. Ex d housings want certified glands and conduit practice; Ex ia loops want barriers and disciplined IS wiring. Put the certificate number in the purchase line, not the words explosion proof.
6. Choose protocol for the host and the maintenance bench, in that order. Plain 4 to 20 mA for simple indication into a PLC analog card; HART 7 as the mainstream default because ranging and diagnostics pay for themselves at first commissioning; fieldbus where the segment already exists.
7. Set the drift and diagnostics expectation that sizes the maintenance calendar. Premium stability claims support 3 to 5 year verification; standard cells want 12 months. Ask what device diagnostics are available on your protocol: plugged impulse line detection is now a mainstream software option on HART and fieldbus hosts.

Steps 1 and 2 interact: raising the span for overpressure headroom trades away working resolution, and the conflict is real whenever a pump can deadhead far above its normal discharge. Where it bites, the resolution is hardware rather than compromise: a cell with a higher proof rating, a snubber for slam, or a relief path protects the instrument while keeping the span honest. Buying a 0 to 25 MPa cell to survive a 9 MPa deadhead on a 900 kPa duty throws away eleven counts of resolution for every hundred used.

Budget structure is stable across the market. Purchase price is half the story; accessories and recurring calibration are the other half, and the cheapest transmitter on the quote is rarely the cheapest loop at year five. The table below gives planning ranges in US dollars for the variants and the recurring items a gauge loop actually carries.

Write the RFQ so vendors quote against measurable lines rather than a model number. A complete gauge transmitter inquiry contains the following eight items, and each one eliminates a class of misquotation on its own.

• Fluid and phase: liquid, gas, steam, or slurry, with the chemistry that drives wetted parts, including chloride and hydrogen sulfide content.
• Normal and maximum operating pressure with units, plus the credible overpressure event and its magnitude.
• Process temperature at the tap, and ambient extremes at the mounting point.
• Connection size and type: 1/2 NPT, G1/2, or flange class and face, plus manifold, siphon, or seal intent.
• Signal and protocol: 4 to 20 mA, HART 7, or fieldbus, with the host system named.
• Area classification: zone or division, gas group, temperature class, and the protection concept preferred.
• Accuracy target as a percent of calibrated span, with the reference conditions assumed.
• Verification requirement: calibration certificate scope, ISO/IEC 17025 accreditation if the QA file needs it, and the requested interval.

A vendor answering that list has no room to quote a general purpose cell against a sour service duty, and you have no room to be surprised by an accessory line at month two. The checklist is deliberately short enough to attach to a purchase requisition, which is where selection discipline actually survives contact with procurement.

Field Practice: Installation, Calibration, and Failure Modes

Impulse piping orientation

Orientation follows the fluid. In gas service, tap the top of the pipe and mount the transmitter above the tap, so condensate drains back to the process instead of collecting a hydraulic head on the cell. In liquid service, tap the side of the pipe and mount the transmitter at or below the tap, with lines flooded solid and vented at commissioning. In steam service, a siphon or condensate loop sits between tap and instrument and must stay full, because live steam reaching a silicone filled cell is a fill fluid failure, not a reading error.

Every vertical meter of liquid leg is a standing head error of about 10 kPa for water, more for dense fluids, and it is compensable at zero trim only if the leg holds a constant height and density. That is the quiet argument for close coupling: a compact transmitter mounted directly on the tap deletes the leg, the head arithmetic, and a winterization problem in one stroke, at the price of vibration exposure that supports and snubbers must then manage. Where the tap must be remote, keep lines short, sloped, and labeled with the head correction that was applied.

Valves and manifolds

Isolation practice on gauge duty is a block valve or a block and bleed combination sized to the connection. The three-valve manifold belongs to differential duty, where isolating and equalizing two ports without opening either to atmosphere is the whole game; on a gauge point its equalize valve does nothing, though plants often standardize the hardware anyway for spares and training. Whatever the arrangement, rate the valve train to the line, not to the sensor: a class 150 manifold on a class 300 line is an audit finding with a date already on it.

Zero trim and loop checkout

The commissioning sequence that survives audits is short. With the loop powered and stable, vent the cell to atmosphere and confirm that both the local display and the host read zero, then apply a zero trim. Next, apply a known pressure near 75 percent of span from a pressure calibrator, confirm the reported value against the applied value, and trim or re-range only if the deviation exceeds the loop budget. Record as-found and as-left at every point: the trend across calibrations is the real drift record, and it is the evidence that justifies stretching or shrinking the interval later.

The electrical half of the checkout mirrors the procedure used on any temperature transmitter: verify 4 mA at live zero, drive 20 mA at full span, confirm the PLC analog input reading against the local display, and log loop resistance and supply voltage at the transmitter terminals. Test terminals on the transmitter let a technician insert a milliammeter without breaking the loop; use them, because the loop you break to take a measurement is the outage you explain afterwards. Bond the housing to the local structure ground, land the cable shield at the cabinet end only, and keep the 24 V return grounded at one point.

Grounding and surge discipline close the electrical list. Split grounding paths on long runs are the classic invisible noise source, arriving as a wobble that no damping setting cures, and lightning season on exposed runs between buildings accounts for a steady share of so-called sensor failures. A surge protector at the cabinet boundary is cheap insurance against both, and a quarterly glance at the trend history catches the rest early.

Calibration intervals and drift management

Default intervals: 12 months for standard duty, 6 months where the measurement feeds custody transfer or fuel gas accounting, and 24 to 36 months defensible for premium cells with documented long term stability outside regulated duty. Set the interval from as-found history, not from the vendor's calendar: two consecutive as-found results inside the loop budget justify stretching the interval, and one outside justifies halving it. Every verification must state the calibrator class, because a field standard at 0.05 percent of reading cannot audit a transmitter claiming 0.04 percent of span; when the numbers are that close, the lab and the fixture decide, not the field kit.

Failure modes and how to specify against them

• Plugged impulse leg. Response slows first, then the reading freezes while the process moves on. Specify larger bore tubing on fouling service, slope lines so they self flush, and put the point on a routine exercise list; on supported digital protocols, plugged leg detection alarms on the signature change.
• Fill fluid loss through a pinholed diaphragm. The output drifts and grows noisy as the hydraulic link ingests gas. Specify the diaphragm alloy for the chemistry at order, inspect the diaphragm at every calibration, and treat unexplained zero shift on a corrosive duty as a seal investigation, not a trim.
• Vent blockage. The transmitter quietly becomes quasi absolute and tracks the weather, reading 1 to 3 kPa low as a low pressure front passes. Keep the vent plug paint free and clear, check it at every calibration, and prefer housings with a protected vent path on washdown and outdoor sites.
• Overpressure. Deadhead or water hammer above proof leaves a permanent zero shift that shows up later as an as-found step change. Compare every credible transient against the proof ladder at selection, and fit snubbers where slam is routine.
• Temperature induced zero shift on capillary systems. Fill fluid expands with ambient cycles, moving the zero by the equivalent of a few kilopascals per 10 deg C on long capillaries. Route capillaries out of direct sun and away from steam traces, insulate where the swings are large, and record the installation temperature at commissioning.
• Electronics and power. Switching supply ripple on the 24 V bus and moisture in poorly glanded housings account for a steady share of failures booked against the sensor. Gland to the actual cable diameter, torque covers to the housing spec, and verify loop voltage at the transmitter terminals at commissioning.

Maintenance calendar

A defensible annual list is short: clean and verify the vent path, exercise the isolation valves, inspect diaphragms and gaskets, replace soft parts on the schedule the chemistry dictates, verify as-found against a traceable pressure calibrator, trim to as-left, and file both numbers. Gasket and O-ring replacement at every calibration is cheap insurance against the slow leak that otherwise announces itself as a month long drift investigation. Between calibrations, the best maintenance is observation: a trend that wanders with the weather says vent, a trend that steps says process event, and a trend that ramps says fill fluid or chemistry, and each verdict sends a technician to a different paragraph of this chapter.