A Centrifugal Speed Indicator is a mechanical tachometer that measures shaft RPM by sensing the centrifugal force generated by spinning flyweights against a calibrated spring. Unlike modern optical or magnetic tachometers, it works without electronics — purely on Newton's laws and a pointer linkage. The flyweights swing outward as speed rises, lifting a sleeve that drives the pointer across a dial. You get a direct, real-time RPM reading on machines like steam engines, mill drives, and diesel generators where electrical sensors can't survive the heat, oil mist, or vibration.
Centrifugal Speed Indicator Interactive Calculator
Vary a reference RPM error and target speed to see how the flyweight force law magnifies tachometer error with RPM.
Equation Used
The calculator applies the article's quadratic centrifugal-force relationship. If a spring calibration error produces e1 RPM error at RPM1, the equivalent force-scale error at RPM2 is e2 = e1*(RPM2/RPM1)^2. Positive error means the tachometer reads high.
- Calibration error is dominated by spring force scaling, not pointer friction.
- Flyweight mass and radius stay constant between speeds.
- Positive error means the indicated RPM is higher than the true RPM.
The Centrifugal Speed Indicator in Action
The principle is simple: spin a mass on a pivoted arm and centrifugal force pulls it outward proportional to the square of the rotational speed. Two or more flyweights mount on hinged arms attached to a vertical spindle that you couple to the shaft you want to measure. As the spindle turns, the weights fly outward and lift a sliding sleeve through a bell-crank linkage. A calibrated spring opposes the motion, so at any given RPM the sleeve sits at one specific height. A lever and rack translate that height into pointer rotation on the dial.
The spring is the heart of the calibration. Get the spring rate wrong by even 5% and the dial reads wrong across the entire range — and the error isn't linear, because centrifugal force scales with ω². That means a 100 RPM error at 500 RPM becomes a 400 RPM error at 1000 RPM if the calibration is off. The flyweights themselves must be matched to within roughly 0.1 g of each other, otherwise the spinning assembly develops a wobble that adds vibration noise to the pointer and accelerates pivot wear.
Common failure modes are predictable. Pivot pins gall and stick, causing the pointer to lag on rising speed and overshoot on falling speed — you'll see a hysteresis band of 20-50 RPM. The return spring relaxes over years of cyclic loading, biasing the reading high by a few percent. And if the drive coupling between the engine shaft and the indicator spindle slips, the reading drops without warning. A flyweight tachometer that hasn't been re-calibrated in a decade is almost certainly reading 3-8% optimistic.
Key Components
- Flyweights: Two or four pivoted masses, typically 10-50 g each on a 30-60 mm arm. They generate the centrifugal force that drives the indication. Mass-matching to within 0.1 g is mandatory or the spindle wobbles.
- Calibration Spring: A coil or leaf spring that opposes flyweight outswing. Its rate sets the dial scale and must hold rate within ±2% across the indicator's working temperature range. Spring relaxation is the dominant long-term error source.
- Sliding Sleeve and Bell-Crank: Translates radial flyweight motion into axial sleeve motion, then into pointer rotation through a rack and pinion or lever. Pivot clearances should sit at 0.02-0.05 mm — tighter and it sticks, looser and you get pointer flutter.
- Drive Spindle and Coupling: Connects the indicator to the measured shaft, typically through a flexible coupling or a thumb-tip on a hand-held unit like a Hagner pocket tachometer. Slip here gives a low reading with no other symptom.
- Dial Face and Pointer: Calibrated against the spring rate and flyweight geometry. Scales are typically non-linear, with the lower third compressed because force scales with ω². A typical engine indicator reads 0-3000 RPM with ±2% full-scale accuracy.
Where the Centrifugal Speed Indicator Is Used
Centrifugal speed indicators dominated rotating-machinery instrumentation from the 1860s through the 1970s, and they're still the right answer wherever electronics struggle. You'll find them on heritage steam plant, large diesel gensets, marine propulsion shafts, and industrial mill drives. The hand-held version — a Hagner or Smiths pocket tachometer with a rubber tip you press against a centre-drilled shaft end — is still the fastest way to spot-check a rotating shaft when you don't trust the installed sensor.
- Marine Propulsion: Wheelhouse RPM indicators on tugboats and fishing vessels driven by Cummins KTA-19 or Caterpillar 3406 main engines, where vibration and salt mist destroy electronic tachs.
- Heritage Steam: Engine room speed indicators on preserved Lancashire boilers and mill engines at Ellenroad Engine House, reading 50-80 RPM on the flywheel shaft.
- Stationary Diesel Generation: Local-panel RPM gauges on Lister-Petter and Rolls-Royce stationary gensets in remote pumping stations where panel electronics aren't economical.
- Field Service and Maintenance: Hagner Mechanical Pocket Tachometer for spot-checking pump shafts, fan drives, and gearbox outputs during commissioning walk-arounds.
- Locomotive Engineering: Cab speed indicators on heritage diesel-hydraulic locomotives like the Class 52 Westerns at the Severn Valley Railway, geared to the cardan shaft.
- Industrial Mill Drives: Permanently mounted indicators on rolling mill main drives and cement kiln drives where shaft access is rough and continuous reading matters more than ±0.1% accuracy.
The Formula Behind the Centrifugal Speed Indicator
The governing equation links pointer deflection to shaft speed through the flyweight mass, arm length, and spring rate. At the low end of the typical operating range, around 20-30% of full scale, the dial is cramped and small RPM changes barely move the pointer — this is why centrifugal indicators are poor at idle measurement. At the high end of full scale the pointer moves fast per RPM, giving you good resolution but you're also closest to the mechanical stops. The sweet spot sits between roughly 40% and 80% of full scale, where the ω²-driven force is large enough to swamp pivot friction but the linkage isn't bottoming out.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Fc | Centrifugal force per flyweight | N | lbf |
| m | Mass of one flyweight | kg | lb |
| r | Effective radius from spindle axis to flyweight centre of mass | m | in |
| ω | Angular velocity of spindle (= 2π × N / 60 for N in RPM) | rad/s | rad/s |
| kspring | Calibration spring rate | N/m | lbf/in |
| xsleeve | Sleeve lift (drives pointer through linkage) | m | in |
Worked Example: Centrifugal Speed Indicator in a vintage tractor restoration shop
A vintage tractor restoration shop in Iowa is calibrating a flyweight RPM indicator for a 1952 John Deere Model B fitted with a PTO-driven centrifugal tachometer. The flyweight mass is 0.025 kg per weight, the effective radius at zero deflection is 0.030 m, and the spring rate is 850 N/m. The tractor's PTO runs from idle around 600 RPM up to a rated 1500 RPM, with the operator typically working the engine at 1100 RPM for belt-pulley work.
Given
- m = 0.025 kg
- r = 0.030 m
- kspring = 850 N/m
- Nnom = 1100 RPM
Solution
Step 1 — convert nominal speed to angular velocity:
Step 2 — compute the centrifugal force per flyweight at nominal speed:
Step 3 — convert force to sleeve lift through the calibration spring:
That 11.7 mm of sleeve lift drives the pointer to roughly mid-scale on the dial — the sweet spot where the operator gets crisp resolution and the linkage isn't fighting the stops.
Step 4 — at the low end of the operating range, idle at 600 RPM:
At idle the pointer barely lifts off the peg — about 30% of the nominal lift even though speed is 55% of nominal. That ω² scaling is why you can't trust a centrifugal indicator below about 25% of full scale; pivot stiction swamps the small force.
Step 5 — at the high end, rated 1500 RPM:
At rated speed the pointer sits near the top of the dial with high RPM-per-mm resolution, so you can spot a 20 RPM governor hunt clearly. Push another 10% over rated and the sleeve will physically bottom against its upper stop, pinning the pointer.
Result
At 1100 RPM the indicator gives a sleeve lift of 11. 7 mm, parking the pointer in the middle of the dial — exactly where the operator wants it during sustained PTO work. Compare across the range: 3.5 mm at idle (cramped, unreliable), 11.7 mm at working speed (crisp), 21.8 mm at rated (good resolution, near the limit). If you measure 9 mm of lift instead of 11.7 mm at the same 1100 RPM input, three failure modes are likely culprits: (1) the calibration spring has relaxed by 20-25% — common after 30+ years of service and easy to confirm by bench-testing the spring against a known mass; (2) the flyweight effective radius has changed because the pivot pins are worn and the arms now hang at a different rest angle; or (3) one of the flyweights has lost mass through corrosion of the lead inserts used in some older Stewart-Warner units, putting the pair out of balance and reducing the average outswing.
Centrifugal Speed Indicator vs Alternatives
Centrifugal indicators compete with magnetic pickup tachometers, optical (laser) tachometers, and Hall-effect sensors. Each technology wins on different axes — the centrifugal mechanism is rarely the most accurate option, but it's often the most survivable one.
| Property | Centrifugal Speed Indicator | Magnetic Pickup Tachometer | Optical Laser Tachometer |
|---|---|---|---|
| Accuracy (% full scale) | ±2% to ±3% | ±0.1% to ±0.5% | ±0.05% |
| Useful RPM range | 50 to 10,000 RPM | 1 to 100,000 RPM | 0.5 to 99,999 RPM |
| Survives oil mist, vibration, heat | Excellent — no electronics | Moderate — sensor cable is the weak link | Poor — optics foul quickly |
| Power required | None — driven by shaft | 12-24 VDC | Battery (handheld) or 12-24 VDC |
| Typical cost (panel mount) | $80-$300 | $150-$600 plus pickup | $200-$1500 |
| Calibration drift over 10 years | 3-8% (spring relaxation) | <0.5% | <0.1% |
| Application fit | Heritage engines, marine, hand-held spot checks | Industrial control loops, gensets | Lab measurement, balancing, commissioning |
Frequently Asked Questions About Centrifugal Speed Indicator
Centrifugal force scales with ω², so the force generated at 600 RPM is only 16% of the force at 1500 RPM. At idle, that small force is fighting the same pivot stiction and linkage friction as at full speed — the friction-to-signal ratio is roughly 6× worse. Below about 25% of full scale, stiction dominates and the reading becomes unreliable.
If you need accurate idle measurement, you've got the wrong instrument. A magnetic pickup or optical tach holds accuracy to zero RPM. The centrifugal mechanism is a mid-to-high-range tool by physics, not by design choice.
200-400 RPM is genuinely awkward for a centrifugal unit. You'd need to size the flyweights and spring so that 400 RPM lands at about 80% of full scale, which means the indicator is purpose-built for low-speed marine work — Stewart-Warner and Smiths both made these but they're scarce now. A custom rebuild is realistic if authenticity matters.
If the shaft has a centre-drilled end accessible from the engine room, the centrifugal route gives you a period-correct, electronics-free reading that survives bilge water and diesel fumes. If you're willing to lose visual authenticity, a Hall-effect sensor on a 60-tooth wheel will give you ±1 RPM and run for decades. For a heritage launch I'd specify centrifugal in the wheelhouse and an inconspicuous Hall sensor on a hidden engineer's panel as a reference.
Three usual suspects, in this order. First, drive coupling backlash — if the flexible coupling between the shaft and the indicator spindle has a worn rubber spider or sloppy dog teeth, torque pulses transmit as angular jitter into the spindle. Second, dial linkage clearance — when bell-crank pivots wear past about 0.1 mm clearance, the pointer wanders within a dead band. Third, mechanical resonance — the flyweight-spring system has a natural frequency typically in the 5-15 Hz range, and if the engine's firing frequency hits that band you get sympathetic oscillation.
Quick diagnostic: lock the indicator drive temporarily with the shaft stationary and tap the case. If the pointer wobbles freely, the linkage is worn. If it sits dead, you've got a coupling or resonance issue.
No, and this is the most common amateur restoration mistake. The dial is calibrated specifically for one spring rate paired with one flyweight mass and arm geometry. Substitute a spring with a different rate and the entire scale is wrong — and wrong non-linearly because the force-versus-RPM relationship is quadratic.
If the original spring has relaxed, you have two options: source an exact OEM replacement (many British and German indicator springs are still listed by Smiths Industries and VDO restoration suppliers), or have the dial face re-printed to match a known new spring rate. Bench-test any replacement against a calibrated mass before installation — measure deflection at 5 N and 20 N and confirm linearity within 2%.
A consistent low reading across the whole scale is unusual for spring relaxation, because relaxed springs typically over-read (a softer spring lets the sleeve travel farther for the same force). A consistent low reading points instead to a slipping drive — most often a worn rubber drive coupling, a loose set screw on the spindle, or in hand-held units a glazed rubber tip skating against the centre-drilled shaft end.
Check the coupling first. Mark the shaft and the spindle with a felt pen, run the engine for 30 seconds, and inspect for relative rotation. If the marks no longer line up, you've found your slip. The fix is either a new coupling or, for hand-helds, dressing the rubber tip with fine sandpaper to restore grip.
Because force scales with ω², the sleeve lift between 0 and 500 RPM is much smaller than the lift between 2500 and 3000 RPM, even though both intervals span 500 RPM. Manufacturers print the dial to match the actual sleeve travel, so the lower numbers crowd together and the upper numbers spread out.
Some premium indicators — Jaeger and certain Smiths aviation units — use a cam linkage between the sleeve and pointer to linearise the scale. These cost roughly 3× a standard unit and the cam adds another wear item. For most industrial use the non-linear scale is fine because the operator works in the upper two-thirds where resolution is best anyway.
References & Further Reading
- Wikipedia contributors. Tachometer. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
