Response Time vs Switching Frequency: What Is the Difference?
Response time is the delay before a sensor output changes. Switching frequency is how often detection can repeat under stated conditions. Use the first for trigger timing and the second for repeated events—then check target width, reset gaps and the controller input.
A suitable waveform measurement can reveal pulse durations and electrical levels that an indicator LED cannot show.
One transition is not the same as a complete cycle
Imagine a part passing a fixed sensor. Its leading edge enters the sensing area, the part remains detectable for a while, and its trailing edge leaves. The output has to respond to both changes before the next part arrives.
Response time follows an edge; switching frequency follows the repeating pattern. A third quantity, total system latency, follows the event all the way to the required controller or mechanical action.
| Specification | What it describes | The practical question |
|---|---|---|
| Response time ms or µs | Delay from a defined sensing change to the corresponding output transition. | How late will this output edge occur? |
| Switching frequency Hz or kHz | Complete detection cycles per second under the manufacturer's test conditions. | Can the sensor keep repeating this pattern? |
| Total system latency ms | Elapsed time from the physical event to the required system action. | When will the machine actually respond? |
OMRON's proximity-sensor terminology makes this edge-versus-repetition distinction. At 1 kHz, there are 1,000 complete cycles per second, so each cycle has a 1 ms period. That does not assign 1 ms to each edge or guarantee 1,000 counts per second in any controller.
ON and OFF delays can change the output pulse width
ON response is the delay before output activation; OFF response is the delay before release. The two need not be equal. For photoelectric devices, the definition relates light-input changes to control-output operation or reset, as shown in OMRON's photoelectric timing definition.
In the illustrative waveform, a 6 ms target window, fixed 2 ms ON delay and fixed 1 ms OFF delay produce a 5 ms active output pulse. The output starts later than the target and is narrower. In this fixed-delay model, output duration = target dwell + OFF delay − ON delay. This assumes both edges are produced, activation corresponds to target presence, and no pulse suppression or timer changes the behavior.
Physical target: detectable from 0 to 6 ms
Sensor output: active from 2 to 7 ms
Two separate maximum delay specifications are not enough to calculate a guaranteed minimum pulse width this way. Also check the output logic: with light-ON or dark-ON operation, target arrival may cause the opposite output transition.
Can response time be converted directly into Hz?
Only when the documentation defines the relationship. Frequency is the reciprocal of a complete period, not automatically the reciprocal of a target-to-output delay.
A 5 ms response-delay entry does not, by itself, mean either 200 Hz or 100 Hz. If a documented cycle explicitly requires 5 ms in each of two states, its 10 ms period corresponds to 100 Hz. But adding two propagation delays does not establish that they are the required target and gap intervals.
A documented example: 1 kHz does not imply equal edge delays
In OMRON FAQ01074, the E2E-X3D1-N example uses a 1 kHz response frequency and a 1:2 ratio to estimate approximately 0.3 ms ON delay and 0.7 ms OFF delay. OMRON also states that the actual delays are shorter than those estimates.
The lesson is not to apply that ratio to every sensor. It is that a manufacturer-specific method can support an estimate, while “1 kHz therefore means 0.5 ms on each edge” is an unsupported assumption.
Compare guaranteed values with guaranteed values
A higher number in brackets may be a typical or average result, not another operating limit. OMRON's EE-SX47/67 photomicrosensor example distinguishes an ensured 1 kHz response frequency from a 3 kHz measured average and directs system design to the ensured value.
For the sensor you are considering, read the rating together with its target pattern, sensing distance and selected mode. A response-mode change may also change sensing performance. Confirm the relevant limits rather than assuming that a faster setting changes only the delay.
Why can a modest event rate still hide a difficult target?
The sensor does not see a production total such as “40 parts per second.” It sees a sequence of target-present and target-absent intervals. The shorter interval can be the difficult one even when the complete cycles are well separated.
Case: two target patterns with the same 40 Hz rate
Illustrative comparison: two repeating patterns move at 2 m/s and have the same 50 mm pitch. Pattern A has a broad detectable feature; Pattern B has a narrow one. Assume the lengths below are the effective sensing windows at the installed position.
Dwell time (ms) = effective length (mm) ÷ speed (m/s)
Event rate (Hz) = 1,000 × speed (m/s) ÷ pitch (mm)
Here, pitch = detectable length + reset gap. The unit shortcut works because 1 m/s = 1 mm/ms. The calculation assumes constant speed and one target/gap pair per pitch.
| At 2 m/s | Pattern A: broad feature | Pattern B: narrow feature |
|---|---|---|
| Detectable length / gap | 25 mm / 25 mm | 4 mm / 46 mm |
| Target-present time | 12.5 ms | 2 ms |
| Target-absent time | 12.5 ms | 23 ms |
| Period / event rate | 25 ms / 40 Hz | 25 ms / 40 Hz |
A buyer comparing only 40 Hz with a sensor's 100 Hz rating might treat these patterns as equivalent. They are not: Pattern B gives the sensor much less time to recognize each feature.
If the candidate's documentation explicitly requires at least 5 ms of target presence, Pattern A meets that one requirement and Pattern B does not. That is not full application approval. If the only available value is a 5 ms maximum response delay, neither conclusion follows automatically: ask for the minimum detectable target duration or short-pulse test evidence.
What to do next: compare both physical intervals with the sensor's documented detection/reset requirements, then verify the resulting electrical pulse at the controller. Use the detectable feature—not the full carton, label or metal part. Beam geometry, material, distance and approach can change the effective window.
For rotary targets, count teeth or slots—not just revolutions
For one event per feature, event rate = RPM × features per revolution ÷ 60. At 6,000 RPM with 10 equally spaced teeth, the rate is 1,000 Hz. Equal effective tooth and gap intervals would each last 0.5 ms; a 1:4 ratio would give 0.2 ms of target time and 0.8 ms of gap time at the same rate.
Use actual tooth geometry, sensing distance and runout when evaluating the windows. For more on this part of selection, see switching frequency and missed signals.
A valid sensor pulse can still be missed by the PLC
A programmable logic controller (PLC) must first accept the electrical state and then make the event available to the program. A sensor that detects every part can therefore coexist with an incorrect controller count.
An input filter can remove the pulse entirely
Rockwell Automation's 1756 input-filter documentation gives a concrete example: with a 2 ms OFF-to-ON filter, a pulse that remains ON for only 1.8 ms is not reported as ON. The new state must persist long enough to qualify.
This is not simply a pulse appearing 2 ms late; the reported ON event never appears. A higher-frequency sensor will not, by itself, change that filter requirement. Check the actual pulse duration, electrical levels and channel configuration. Changing filtering also changes noise rejection, so validate the receiving path after an adjustment.
Pulse catch is not the same as counting every pulse
A brief pulse may fall entirely between normal input observations. A pulse-catch function can hold an event until the controller reads it, but several events can still collapse into one indication.
For the S7-1200 G2, Siemens' pulse-catch description places this function after the input filter and states that only the first pulse is read when multiple pulses occur in one scan. Use the exact controller/channel documentation to choose event capture or a high-speed counter for the required event pattern.
A high-speed counter (HSC) still needs a valid signal within its voltage, frequency and minimum HIGH/LOW-time limits. It cannot recover a target that the sensor did not detect. If the terminal waveform is incorrect, resolve PLC input compatibility and wiring before assuming that software is the limiting factor.
Correct counting does not guarantee the right action position
Counting asks whether an event was recorded. Printing, cutting or rejecting also asks where the moving product is when the action happens. This is where response delay and total system latency matter more than a frequency rating alone.
Case: every part is counted, but the action happens late
Illustrative constant-speed scenario: assume a line moves at 1.5 m/s, the relevant sensor edge has a fixed 2 ms delay, and distinct downstream controller, output and actuator stages add another 6 ms. The assumed delays are for explanation, not measurements from a customer installation.
Travel during delay (mm) = speed (m/s) × delay (ms)
The sensor contribution is 1.5 × 2 = 3 mm. The complete 8 ms path corresponds to 1.5 × 8 = 12 mm of travel.
If the count is correct but the action is late, choosing a larger Hz rating does not identify or remove the cause. Separate the delay stages and establish the timing reference used by the machine's position or tracking logic.
The useful conclusion is that “fast enough to count” and “early enough to act” are different requirements. Use defined start/end points when adding delays; do not count a communication or input delay twice if it is already included in an end-to-end figure.
Predictable lag and timing variation need different treatment
A stable offset may be compensated in a validated tracking design. Event-to-event variation—often called jitter—changes the offset from one event to the next. At a fixed 1.5 m/s, an assumed ±0.2 ms timing variation contributes ±0.3 mm of position variation before other errors are included.
If speed changes during the delay, use the motion profile or encoder-based tracking rather than one nominal speed. Compare repeatability and timing variation as well as response delay. A smaller average delay does not, by itself, prove tighter positioning.
Choose the timing specification for the job
You do not need the highest switching frequency for every application. You need sufficient detection and reset capability, a receiving path that preserves the event, and timing that fits the required result.
- For repeated counting: start with the shortest target and gap intervals, then the guaranteed cycle rating and controller capture limits.
- For a position-sensitive trigger: also examine the relevant edge delay, its variation and the complete sensor-to-action path.
- For a slowly changing presence check: sensing stability and the allowed sequence delay may be more important than an unnecessarily high Hz rating.
When the machine behaves differently from the specification
Follow the event through three observation points: the physical target, the electrical output at the PLC terminal, and the recorded count or action.
No valid terminal pulse points you toward sensing conditions or the electrical path. A valid pulse with a missing count points toward input qualification, capture or count handling. A correct count with a misplaced action points toward timing and motion tracking.
The indicator LED alone cannot separate these cases. Use suitable instrument resolution and an independent physical-event reference when measuring response delay.
Work safely: isolate hazardous energy before changing wiring or probe connections, and use circuit-appropriate rated equipment and qualified personnel. Follow the instrument's grounding precautions: never defeat protective earth to float a grounded oscilloscope or attach its ground clip to an unknown circuit point. These process-sensing examples are not personnel-protection or machine-stopping-time validation.
The distinction to remember: response time tells you when an edge arrives; switching frequency describes repeated detection under defined conditions. Reliable machine behavior depends on preserving the actual pulse and acting on it at the right time.
For an xsz sensor selection discussion, share the target, speed, detectable feature and gap, sensing distance, controller input, and whether the requirement is counting or action position. Those details are more useful than asking for a “fast sensor” alone.