AutomateDirectUSAMotion LibraryShop the parts catalog

How-to

Measuring RPM With a Proximity Switch on a Pulley

Sensors, encoders and machine vision on the machine Published 7 min read

Short answer

One bolt pass per revolution at 6,000 RPM is a 100 Hz signal any standard inductive proximity switch can produce and a scanned PLC input cannot reliably count, so route it to a high-speed input and verify the ON pulse clears the input's documented minimum — 100 µs on the DL05's X0–X2 inputs — with a scope at full speed.

On this page
  1. Turn the speed into a frequency first
  2. Match the sensor to the frequency and the target
  3. Compute the ON time the geometry delivers
  4. Choose the input before writing logic
  5. Commission in this order
  6. The call for each case
  7. Questions asked

One bolt pass per revolution at 6,000 RPM is a 100 Hz signal, and that one calculation settles the design before anything is bought. A standard inductive proximity switch outruns 100 Hz easily; a scanned PLC input cannot count it; and each pulse must stay ON long enough to clear the input's documented minimum — 100 µs on the DL05 high-speed inputs X0–X2, which exist only on DC-input models. The ON width is set by geometry: bolt-head width against pulley circumference. Size the chain in order — frequency, sensor, pulse width, input — then verify with a scope at full speed before trusting the RPM reading.

Turn the speed into a frequency first #

The sensor sees target passes, not revolutions, so convert before selecting anything:

Pulse frequency (Hz) = (RPM ÷ 60) × target passes per revolution

At 6,000 RPM with one bolt head passing once per turn: 6,000 ÷ 60 = 100 revolutions per second, × 1 target = 100 Hz, a 10 ms period. Every check that follows is made against those two numbers.

What the pulley is bolted to decides what the number means. A pulley on the crank nose reads engine speed directly. A belt-driven accessory pulley reads its own speed, and the relationship is pulley speed × pulley diameter = driver speed × driver diameter; belt slip adds error that no sensing can remove.

Decision: mount the target on the pulley whose speed you actually want displayed. If that pulley is belt-driven, put the diameter ratio into the PLC scaling and expect slip to appear as noise in the reading.

Match the sensor to the frequency and the target #

The target is a steel bolt head, so the sensor is inductive: it detects metal without contact and ignores the oil and dirt around a running machine. The rating to check first is switching frequency — the fastest rate at which the sensor repeatably resolves each individual passing target. In current manufacturer datasheets, inductive ratings run from 150 Hz on the slowest examples through 700 Hz for one 4 mm-range model to 8 kHz at the fast end, with the type typically reaching about 5 kHz. Every one of those clears the calculated rate, so at this speed the sensor's frequency rating is not the binding constraint — still read the value from the datasheet of the exact part you buy, because the rating belongs to the model, not the type.

Target fit deserves the attention instead. Rated sensing distance is established on a standard mild-steel target of a set size; a bolt head smaller than that standard does not reach the full rated distance, so take the derating from the sensor's datasheet and set the gap from the derated value. An unshielded (non-flush) body switches from farther away, which eases the gap to a rotating head. Body sizes of 8 mm and 12 mm are the usual starting points for this job — a rule of thumb from installation practice, not a datasheet limit. A quick-disconnect connector changes nothing electrically; it only lets you swap a damaged sensor without re-terminating cable.

The output-side wiring checks are the same ones covered in replacing a mechanical limit switch with a proximity sensor.

Decision: select the inductive sensor on target fit, gap, and body style; any model whose documented switching frequency clears the calculated rate with margin passes the frequency test.

Compute the ON time the geometry delivers #

Frequency alone does not prove the input will see the pulses — each one has to stay ON long enough. Bolt width against pulley circumference sets that:

ON time (s) = (60 × target width) ÷ (π × pulley diameter × RPM)

Calculated example, not a measured result — 8 in pulley, 1/2 in bolt head, 6,000 RPM:

  • Pulley circumference: π × 8 in = 25.1 in.
  • Fraction of one turn the head occupies: 0.5 ÷ 25.1 = 0.020, about 2%.
  • Time per turn: 60 ÷ 6,000 = 10 ms.
  • ON time: 10 ms × 0.020 = 0.20 ms = 200 µs.

The sensor's effective zone is wider than the head, so the real pulse runs slightly longer than this floor; amplifier response and hysteresis smear the edges back the other way. That is why the confirming check is a scope, not the formula.

Bolt head width ON time at 6,000 RPM on an 8 in pulley Against a 100 µs input minimum
1/4 in 100 µs at the limit — no margin
5/16 in 124 µs 24% margin
3/8 in 149 µs workable
1/2 in 199 µs comfortable

Both diameter and speed move the result in proportion: twice the RPM or half the diameter halves the ON time.

Decision: if the computed ON time is not at least twice the input minimum, widen the effective target — a larger head or a steel block bolted to the pulley — or move the counting to hardware with a lower minimum.

Choose the input before writing logic #

A scanned input is read once per scan. A 200 µs pulse is shorter than any scan, so most pulses fall between samples and the count reads low no matter what the ladder logic does. The counting has to happen in input hardware built for pulses.

On the DL05, that means the high-speed inputs X0–X2, which exist only on DC-input models — an AC-input model has none, so confirm the model before planning around them — and the documented minimum ON pulse is 100 µs. Dedicated counter modules run lower: the T1H-CTRIO specification lists a 5 µs minimum pulse width, and one 1769 compact counter module documents a 250 ns minimum pulse width with a 1 MHz maximum input frequency.

Decision for the example case: 100 Hz at roughly 200 µs clears the DL05 X0–X2 on both counts — the rate is far below what a high-speed input tracks, and the width is twice the minimum. Land the signal on X0–X2 and verify at speed. Had the width computed at 100 µs, the call would be a wider target or a counter module documented at 5 µs.

When one pulse per turn is too coarse for the application, the next step is an encoder on the shaft; timing belt versus direct mount for a rotary encoder works through how to drive it.

Commission in this order #

  1. With the machine locked out, mount the sensor so the head passes the sensing face squarely once per turn, set the gap from the derated distance for the head size, and confirm the bracket and cable cannot reach the rotating pulley.
  2. Confirm the controller actually has the input you counted on — a DC-input DL05 for X0–X2 — and land the sensor output on that terminal, not on a scanned one.
  3. With guards back in place, run at the highest intended speed and scope the signal at the PLC terminal; the shortest ON time must clear the documented minimum, 100 µs here, with margin. Measure at the terminal, never by probing near the rotating assembly.
  4. Check the reading against an independent tachometer: expected frequency = RPM ÷ 60 × target passes per revolution, which is 100 Hz for this case.
  5. Sweep the speed up to the maximum; a low-speed check proves nothing about pulse width at the top end.

The call for each case #

  • 6,000 RPM, one 1/2 in head on an 8 in pulley: 100 Hz at about 200 µs — standard inductive prox into DL05 X0–X2, verified at full speed. Run it.
  • Same shaft with a 1/4 in head: about 100 µs — no margin against the DL05 minimum; widen the target or count on a module documented at 5 µs.
  • Reading correct at low speed, low or erratic at high speed: scope the input before replacing anything. Clean pulses with a low count point to the input or its configuration; missing or narrow pulses point to the target, the gap, or the sensor.
  • Sensor on a belt-driven pulley but engine RPM wanted: apply the diameter ratio in the scaling, expect slip to show as noise, and move the target to a crank-driven pulley when the number must be exact.

Questions asked

How fine a resolution does one pulse per revolution give?

With one target and a one-second count gate, one count equals 60 RPM, so the reading resolves to about 60 RPM and updates once per second — 1% of full scale at 6,000 RPM. A longer gate improves resolution but slows the update; adding targets improves both, but raises the frequency and shortens each ON pulse, so re-run the width calculation before bolting on more heads.

The RPM reads right at low speed and drops out at high speed — sensor or PLC?

Scope the sensor output at the PLC input terminal at the failing speed. Clean pulses at full width with a low count point to the input or its configuration. Missing, ragged, or narrow pulses point to the target presentation, the mounting gap, or the sensor rating. Fix the branch that fails; replacing the sensor first fixes nothing when the input is the limit.

Does an 8-inch pulley diameter change the pulse frequency?

No — frequency depends only on speed and targets per revolution, so 6,000 RPM with one bolt is 100 Hz on any diameter. Diameter sets the ON time instead: the same 1/2-inch head on a 4-inch pulley covers twice the fraction of a turn and stays ON about 398 µs instead of 199 µs. Smaller pulleys give longer, safer pulses at the same speed.

Should the proximity switch be shielded or unshielded here?

Unshielded (non-flush) suits a rotating target: it switches from farther away, so the bolt head needs less precision in the mounting gap. The tradeoff is sensitivity to surrounding metal, so keep it clear of the pulley rim and bracket. A shielded (flush) body tolerates nearby metal but at shorter range. Either way, set the gap from the datasheet's rated distance for your actual target size.

Keep reading

  • How-to

    Replacing a mechanical limit switch with a proximity sensor

    Yes — an inductive proximity sensor on a 24 V DC supply, switching a mechanical interface relay whose contact reproduces the original circuit logic, replaces a worn mechanical limit switch. Verify first whether the normally closed contact is truly closed at rest or held open by the cam. Keep two-wire sensors away from solid-state relays: their off-state leakage can hold the relay on.

    Sensors, encoders and machine vision on the machine

  • Selection guide

    Timing Belt vs Direct Mount for a Rotary Encoder

    A guarded direct or hollow-shaft mount on the 5/8 in rear shaft is the better choice; 200 RPM is trivial for a belt drive, but debris-packed grooves wreck encoder bearings and loose setscrews cause slip. Build the belt drive only with matched pulleys, a rigid bracket, minimum tension and a full cover, and verify 1:1 with a marker and a speed check.

    Sensors, encoders and machine vision on the machine

  • How-to

    How to Identify Geared Resolver and Encoder Feedback on Old Servos

    Treat the resolver plus its geared 1:4 and 1:80 channels as one integrated coarse-and-fine, multi-turn position package: the resolver reads absolute angle within one shaft revolution, and the geared channels count revolutions above it. Identify each channel by resistance checks, ratio counting and waveform capture before choosing a replacement drive, because the resolver model number alone establishes neither commutation nor multi-turn position recovery.

    Sensors, encoders and machine vision on the machine