Pressure Sensor Response Time: Rise Time, T63 & T90 Explained

Respuesta rápida:

Pressure sensor response time is how long a sensor takes to show a pressure change at its output. Manufacturers publish the same figure three ways — rise time (10–90%), T63 and T90 — and T90 is roughly 2.3 times T63, so two identical sensors can look 2.3 times apart on paper. Dead time, damping and your PLC add more. This guide defines each term, converts between them, and shows what to specify for hydraulic, water and digital pressure sensors. Simply Buy supplies them across multiple brands.

What will this guide give you?

Three quotations for the same hydraulic power unit come back with numbers you cannot rank as written. One datasheet promises ≤ 1 ms, a second publishes T90 = 10 ms, a third headlines 90 ms — a ninety-fold spread nobody can defend in a design review. Each answers a different question, and that gap is where pressure sensor response time becomes an argument instead of a specification.

Here is what you get: plain-English definitions of response time, rise time, T63 and T90, the arithmetic that puts any supplier’s figure on a comparable basis, and two tables: manufacturer-published specifications, and a selection table matching response-time class to application. It is written for B2B distributors, system integrators, EPC buyers and maintenance teams, and every figure below is the manufacturer’s own number with the definition that vendor publishes.

Where does a sensor’s response time actually come from?

A pressure sensor is a chain, not a component. KELLER Pressure, testing its own step response, is blunt about the dominant link: “In most cases, it is not the electronics, but the mechanical-fluidic coupling.”

From the process outward: the connection and impulse line, which add volume and restrict flow; the diaphragm, oil fill and sensing element, where pressure arrives mechanically before electronics see it; the conditioning, damping and output stage, whatever the signal path; and finally your own system — input filter, PLC scan time, HMI refresh.

The most useful field finding there: “air bubbles in the system are the biggest killer of dynamics,” and in oil-filled systems “even a small amount of residual air has a significant effect on the step response.” A 1 ms sensor on a cavity that has not been bled is not a 1 ms sensor.

Still choosing between measuring principles? Our overview of the main types of pressure sensors explains where each one fits, and the pressure sensors category shows the models we stock.

What do rise time, T63 and T90 mean in practice?

A step response applies a sudden pressure change and records how the output climbs. It never jumps, so where you stop measuring decides the number you get.

WIKA defines response time as “the interval required by the output signal of a pressure sensor to display a change in the applied pressure,” and calls the rise time “of greatest practical relevance.” Its worked example: the interval in which a step “from 10 to 90% of the full span (e.g. from 60 to 540 bar for a 0 … 600 bar pressure sensor) leads to a defined change in output signal of 10 to 90% of the final value.”

Yokogawa defines the same words differently: response time is “the amount of time (in seconds) it takes for the output signal to reach 63.2% of an instantaneous full-scale pressure change,” because transmitters “respond to pressure changes as a first-order system.”

Both are correct; they are not interchangeable, and that is why pressure sensor response time causes so many purchasing arguments.

How do T63 and T90 relate to each other?

The 63.2% figure comes from the physics. A first-order system approaches its final value along output(t) = 1 − e^(−t/τ), where τ (tau) is the time constant. Solving that equation gives the numbers worth carrying:

Two consequences matter. T90 is about 2.3 times T63 on the same sensor, and a 10–90% rise time sits within about 5% of T90 — close enough that vendors treat the labels as interchangeable, which is why WIKA writes “T90 or 10 … 90%”.

There is also a bridge to bandwidth: since τ = 1 / (2π × f_c), a 10–90% rise time works out to roughly 0.35 / f_c, so a sensor with T63 = 1 ms is about 159 Hz of bandwidth, a T90 of about 2.3 ms and a T99 of about 4.6 ms.

How do you convert T63 to T90 before comparing quotes?

One datasheet prints “response time: 1 ms (T63)”; another “2.3 ms (T90)”. Identical hardware; on paper, the first looks twice as fast. Convert before you rank:

• T63 → T90: multiply by 2.3
• 10–90% rise time → T90: multiply by about 1.05
• T90 → T99: multiply by 2.0

State the basis in the purchase order. If a supplier will not say which basis their figure uses, that is your answer.

What is the difference between dead time and time constant?

Yokogawa splits total response time as RTt = Td + Tc, where Td is dead time, “the amount of time for the transmitter to initially respond to the change in pressure,” and Tc is the time constant — mechanical plus electronic. Its own example: transmitters with Hastelloy C-276 diaphragms “have a response time of 90 ms (including a dead time of 45 ms).” Half of that figure is pure delay before the output moves at all.

Dead time is a transport delay. No controller tuning or filter removes it; it consumes phase margin and caps how fast a loop can run.

Should you compare bandwidth or sample rate?

Analogue outputs are governed by the limiting frequency — the −3 dB point up to which pressure changes pass with no more than 3 dB attenuation, about 70% of amplitude. Digital outputs are governed by the sample rate: “how often a signal is recorded and processed internally per second.” KELLER states the rule plainly: “The limiting frequency is decisive for analogue outputs and the sampling rate for digital outputs,” and notes that on a digital sensor “the dead time is part of the response time.” Its own fast analogue electronics reach “typically less than 10 µs,” while its digital X-Line needs “around 2 to 4 ms, depending on the selected output signal.” Our note on digital versus analog sensors covers the practical differences.

Which “fast” numbers are not response times?

Four neighbours on the datasheet get mistaken for pressure sensor response time.

• Switching frequency. The SICK IME12-04BPSZC0S, an inductive proximity sensor stocked by Simply Buy, publishes 2,000 Hz. That is a discrete-output repetition rate at a 4 mm sensing range, not the time for a pressure output to reach 90% of a step; switching sensors and analogue transmitters do not share a scale.
• Time delay before availability. The same sensor lists ≤ 100 ms — how long after power-up before its output is valid. Critical on a machine that powers down between cycles; irrelevant to tracking a pressure change.
• Damping and filter settings. A filter that steadies a noisy reading also slows the sensor down. Ask which setting it was measured at.
• Display refresh and PLC scan. The display updates far more slowly than the analogue signal behind it, and your input filter may be slower still. A pressure switch reports a threshold, not a continuous value.

What response times do manufacturers actually publish?

Each row is a manufacturer-published figure with its vendor’s own definition. Read the third column before the second.

The legitimate span of pressure sensor response time runs across four orders of magnitude, and the fastest number is not automatically the right purchase. If you would rather send the application than argue the datasheet, email or message your pressure range, output type and the response-time class you need — a full BOM works as well as a single line. We return matched models with each manufacturer’s published figure, stock status and lead time.

Sensores de presión OMRON E8F2

Voltaje de la fuente de alimentación: 12 a 24 V CC
Consumo actual: 70 mA máx.
Rangos de presión:
0 a 100 kPa (E8F2-A01C/A01B)
0 a 1 MPa (E8F2-B10C/B10B)
0 a -101 kPa (E8F2-AN0C/AN0B)

Ver detalles del producto

What should you ask before you trust a response time figure?

Five questions to run against any figure before it enters a tender.

1. Which basis — T63, T90, T99 or 10–90%?
2. Measured on which output? OMRON’s 5 ms applies to the E8F2’s ON/OFF output, not to every signal it produces.
3. What was the test? Step size, share of span, test volume and fittings — the cavity and capillary change the answer more than the electronics do.
4. At what temperature, with which medium, and with which damping setting? SUCO publishes a compensated range of 0 … +70 °C alongside its ≤ 2 ms response time; check both.
5. To which standard, and does the number include dead time? SUCO defines accuracy “according to IEC 61298-2”; if dead time is not included, ask separately.

What response time do hydraulic pressure sensors need?

Response time in hydraulics is usually a control-loop requirement, not a monitoring one. WIKA’s guidance is that “in applications with high load cycles, such as in mobile hydraulics, short rise times are recommended,” and KELLER adds that this is where analogue transmitters with high bandwidth and minimal dead time belong. That is why 4–20 mA remains the default; our wiring guide for 4–20 mA pressure transmitters covers the installation side.

Typical hardware runs from analogue transmitters such as the Danfoss MBS 3000 at 0–10 bar, to IO-Link units like the ifm PN3594 when you want diagnostics on the same cable, and the ifm PN7092 at 0 to 100 bar for higher-pressure circuits.

Two warnings. Cavitation and pressure spikes punish the installation, not the datasheet: ifm recommends “a damping element or restrictor” for dynamic pressures and cavitation — a fitting that deliberately slows the pressure the diaphragm sees, because a faster sensor on a cavitating circuit just reports the damage sooner. And speed is only useful if something downstream can use it: a 1 ms transmitter feeding a PLC input filtered at 20 ms is a 20 ms measurement.

Why do water pressure sensors sometimes need to be slow?

Water inverts the problem. Pump control, level and pressure monitoring want stability, and the dominant fast event — water hammer after a sudden valve closure — is a transient most control systems should not act on. So the same vendors who advertise microseconds elsewhere publish slow figures here. WIKA notes that submersible pressure transmitters “may also show clearly higher values of ≥ 100 ms on purpose,” and that for level measurement “long rise times are usually advantageous.” A ten-times-faster transmitter just hands the PLC more noise to average.

For water and hygienic duty, the mechanical specification usually outweighs the dynamic one: a flush diaphragm avoids the crevice a standard port leaves, and the certificate set matters more than the millisecond number. The IFM PG2799 flush pressure sensor with analogue display is a reference point for that class — −1…10 bar relative, 4–20 mA plus two switching outputs, IO-Link 1.1, a fully welded 316L stainless-steel diaphragm, IP67/IP69K and EHEDG, 3-A and FDA listings. For mbar-level work the ifm PI2798 is a flush alternative, and our guide to choosing a pressure sensor for a water pump covers the application side.

How fast are digital pressure sensors really?

Digital sensors add a display, often a bus, and a new set of clocks. The OMRON E8F2 digital pressure sensor shows how narrow one published figure can be: response time (ON/OFF output) 5 ms max, on rated ranges of 0–100 kPa, 0–1 MPa and 0 to −101 kPa gauge, with a 1–5 V linear output, a 3.5-digit LED display, IP50 and a 0 to 55 °C ambient limit.

Three things follow. The 5 ms belongs to the switching output, not to the linear output, the display or any bus update. The IP50 rating and 0–55 °C window make this a cabinet-mounted device, not a washdown one. And the family is discontinued: OMRON’s notice of March 2023 lists every model with “No recommended replacement” — a sourcing problem rather than a technical one, and a common one in this category.

How do you choose a pressure sensor by application?

Only two of the six rows improve with a smaller millisecond figure. Response time is one column in a selection, and for several of these applications a smaller number is the wrong specification entirely. The broader range is on our OMRON E8F2 digital pressure sensor and the industrial sensor hub.

If you want that comparison run against your own list, send the BOM — pressure range, output type and the response-time class you need. We quote matched models with each manufacturer’s published figure and its basis, plus stock status and lead time.

How do you verify and source a response-time-critical sensor?

Simply Buy is a multi-brand distributor and supply-chain integrator: we source more than 100 industrial brands through 300+ suppliers rather than manufacturing sensors ourselves.

• Model matching. Most response-time mistakes are order-code mistakes. The E8F2 family splits by range and output — A01 is 0–100 kPa, B10 is 0–1 MPa, AN0 is 0 to −101 kPa, and a final C means NPN while B means PNP. A quote that returns “E8F2, 1 MPa” has not matched your part. And when a family ends without a replacement, as the E8F2 did, remaining stock is what exists — ask for the date code, the quantity available, and a documented alternative with its response-time basis stated.

• Genuine-product verification and certificates. Order-code match against the manufacturer’s published data, label and serial capture at dispatch, country of origin and date code on request, and approval marks where they exist — the SICK IME12-04BPSZC0S carries UL File No. NRKH.E181493, which is checkable. For UL-listed items, we hold UL documentation, and hygienic listings such as EHEDG and 3-A travel, where the manufacturer issues them for that model.

• Stock and lead time. More than 5,000 industrial control products in stock, with stock points in the US and Germany. In stock versus factory order is the largest driver of lead time, so ask in writing.

• DDP delivery. Delivered Duty Paid moves export, import, duty and final delivery to us, which shortens the administrative tail on an urgent spare. It changes who clears customs — not what the sensor measures.

If response time is why you are replacing a sensor, put the requirement in the request: the basis you will accept, whether dead time must be stated, the output it applies to, and the damping setting active at that figure.

What should you do next?

To sum up: pressure sensor response time is only comparable once you know the basis. Convert every figure to T90 or T63, ask whether dead time is included, and check which output the number belongs to. Buy speed only as far as your PLC can use it; for water and level, a deliberately damped sensor is usually the better engineering choice.

Message us on WhatsApp or email your part numbers or full BOM and add three lines: the response-time class you need, the basis you will compare on, and the date the parts must be on site. We return matched models, each manufacturer’s published figure with its definition, stock status, lead time and delivery terms. Our published quote turnaround is 12 hours, and the página de contacto has the direct details.

Still deciding which class you need? Send the application and let the data settle it. Response time is only expensive when it is specified by accident.

What else do buyers ask about pressure sensor response time?

1. Is rise time the same as response time?

Not exactly. Response time is the general term for how long the output takes to reflect an input change; rise time is the specific measurement WIKA calls “of greatest practical relevance,” usually 10% to 90% of the final value.

2. What are T63 and T90?

T63 is the time for the output to reach 63.2% of its final value after a step change — for a first-order system, exactly one time constant, and the basis Yokogawa publishes as response time. T90 is the time to 90%, about 2.3 times T63 on the same basis: T63 = 1 ms and T90 ≈ 2.3 ms describe the same sensor.

3. Can a digital pressure sensor be as fast as an analogue one?

Not at the top end — KELLER’s published figures run from under 10 µs for fast analogue electronics to 2–4 ms for its digital X-Line, with dead time folded in. Choose digital for the display, bus and diagnostics; analogue when the loop itself is fast.

4. Does a long response time mean a bad sensor?

No. Submersible level transmitters are built with rise times of ≥ 100 ms on purpose, because a fast output on a moving water surface reports turbulence instead of level.