A Speed Indicator is a mechanical or electromechanical instrument that measures the rotational speed of a shaft, usually in RPM. It tackles the basic problem of knowing how fast a prime mover or driven machine is actually turning, which you cannot judge accurately by eye or ear above about 200 RPM. Most designs work by pressing a tip against a shaft centre and reading either a centrifugal flyweight, a chronometric counter, or a magnetic drag cup. Pair the reading with torque and you get shaft power — the basis of every dynamometer test from a 3 hp Lister diesel to a 400 kW marine genset.
Speed Indicator Interactive Calculator
Vary RPM, flyweight mass, radius, and spring rate to see centrifugal force, sleeve lift, and dial deflection update live.
Equation Used
This calculator models the article's centrifugal flyweight speed indicator. Shaft speed is converted to angular velocity, each flyweight produces centrifugal force Fc = m omega^2 r, and the two weights lift a sleeve against a linear spring. Dial deflection is shown as proportional to RPM squared on a 3000 RPM full-scale dial.
- Ideal two-flyweight centrifugal speed indicator.
- Sleeve load is approximated as the sum of two flyweight centrifugal forces.
- Spring is linear and friction, slip, and linkage losses are ignored.
- Dial scale is normalized to the article's 0-3000 RPM dial.
Operating Principle of the Speed Indicator
A Speed Indicator turns shaft rotation into a number you can read. The simplest hand-held type — think of a vintage Hasler or a Smiths chronometric — has a rubber tip you press into the centre dimple of a rotating shaft. Inside the case, that input drives one of three mechanisms: a centrifugal flyweight pulling against a hairspring, a chronometric escapement that counts revolutions over a fixed time gate, or a magnetic eddy-current drag cup whose deflection is proportional to RPM. Each one converts angular velocity into a needle position on a calibrated dial, typically with full-scale ranges of 500, 3000, or 10000 RPM.
Why three different mechanisms? Because each handles a different speed band and accuracy requirement. Centrifugal types are cheap and rugged but drift with temperature and spring fatigue — accuracy is usually ±2% of full scale. Chronometric tachometers are the most accurate hand instruments ever made, holding ±0.25% because they decouple the measurement from spring force and instead time a fixed count of revolutions. Magnetic drag cups sit in the middle and dominate automotive dashboards from the 1950s through the 1980s. If you press the tip off-axis by more than about 5°, all three types under-read because the rubber drive cone slips against the dimple. If the hairspring weakens with age, a centrifugal indicator over-reads at low RPM and under-reads at high RPM — a classic field symptom on a 50-year-old Smiths instrument.
The failure mode that bites people most often is dimple wear. The shaft centre needs a clean 60° conical recess about 3 mm deep. Once that dimple wallows out to a flat or skewed crater, the rubber tip walks, the contact slips, and the reading drops 5-15% below true. You then chase a phantom power problem that does not exist.
Key Components
- Drive Tip: A hardened rubber or fibre cone that engages the shaft centre dimple. The cone half-angle is typically 60° to match a standard centre-drilled hole, and the tip must run true to within 0.1 mm or it slips at speeds above 3000 RPM.
- Input Spindle: Carries the rotation from the tip into the gearbox. Runs in two jewelled or bronze bushings with a clearance of 0.02-0.05 mm. Excess clearance lets the spindle wobble and corrupts the reading by up to 3%.
- Reduction Gearbox: Steps the shaft RPM down to a useful range for the indicating element. A 10:1 or 60:1 ratio is common — a 60:1 train converts 6000 RPM input into 100 RPM at the flyweight, keeping the centrifugal force inside the elastic range of the hairspring.
- Sensing Element: The flyweight, escapement, or drag cup that converts RPM into a torque or angular deflection. On a chronometric type the escapement gates exactly 3 seconds of revolution counting before resetting — that gate time sets the fundamental accuracy.
- Calibrated Dial: Reads angular deflection as RPM. Quality instruments use a mirrored arc to eliminate parallax error, which is worth roughly 1.5% of reading on a 75 mm dial viewed at 30° off-axis.
- Hairspring or Restoring Element: Provides the counter-torque against which the flyweight or drag cup deflects. Made of phosphor bronze or Ni-Span-C — the latter is temperature compensated and holds calibration within ±0.1% across 0-50 °C.
Industries That Rely on the Speed Indicator
Speed indicators show up anywhere a rotating shaft needs verification — engine test cells, machine tool spindles, generator commissioning, pump-curve checks, and historical equipment surveys. The reason is simple: torque measurement on its own tells you nothing about power until you multiply it by RPM. Every dynamometer, brake horsepower test, and pump efficiency calculation needs a trustworthy speed reading first.
- Wind Energy Maintenance: A wind farm service crew in Jutland uses a Hasler Tel hand tachometer to verify low-speed-shaft RPM on a Vestas V52 turbine during a gearbox borescope inspection — confirming the rotor turns at 26 RPM nominal before the planetary stage.
- Industrial Pump Commissioning: A Grundfos field engineer commissioning a CR 95 multistage centrifugal pump in a Norwegian fish farm uses a Compact Instruments AT5 contact tachometer to confirm the 2950 RPM nameplate before signing off the pump curve.
- Diesel Generator Testing: A standby genset contractor running a 250 kW Cummins QSL9 acceptance test uses a chronometric speed indicator on the crankshaft pulley to verify governed speed of 1800 RPM ±0.25% before declaring the unit fit for hospital backup duty.
- Machine Tool Calibration: A toolroom calibration tech checking spindle speed on a 1972 Bridgeport Series 1 mill uses a Smiths centrifugal speed indicator to confirm the back-gear range delivers the dial-marked 80 RPM at lowest setting — a classic check before rebuilding a worn variable-speed head.
- Marine Propulsion Survey: A Lloyd's Register surveyor on a North Sea trawler uses a hand tachometer on the propeller shaft of a Caterpillar 3406 to confirm 1200 RPM under full load during a sea trial, cross-checking against the engine room's flyweight tachometer.
- Heritage Engine Restoration: A volunteer at the Anson Engine Museum in Cheshire uses a Smiths chronometric tachometer to verify a rebuilt 1934 Crossley DD4 horizontal oil engine holds its rated 350 RPM under load on the museum's Heenan & Froude rope brake.
The Formula Behind the Speed Indicator
When you pair a speed indicator with a torque measurement — typically from a Prony brake, rope brake, or strain-gauge dynamometer — you get shaft power. This is the headline number every engine test produces. At the low end of the typical operating range (say 300 RPM on a slow-speed oil engine) the RPM term shrinks the power output for a given torque, so the engine has to make far more torque to hit nameplate kW. At the high end (3600 RPM on an automotive petrol engine) modest torque numbers translate into big power figures, but indicator accuracy matters more because a 1% RPM error becomes a 1% power error directly. The sweet spot for hand-tachometer-based measurement sits between 800 and 3000 RPM, where chronometric instruments hold ±0.25% and torque arms are still long enough to read cleanly.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| P | Shaft power | W | ft·lbf/s (÷550 for hp) |
| N | Rotational speed measured by the indicator | RPM | RPM |
| T | Torque on the shaft | N·m | ft·lbf |
| 2π / 60 | Conversion factor from RPM to rad/s | rad/(s·RPM) | rad/(s·RPM) |
Worked Example: Speed Indicator in a small hydro turbine acceptance test
A microhydro contractor in the Scottish Borders is doing the acceptance test on a freshly installed 22 kW Gilkes Turgo turbine driving a belt-coupled induction generator. The crew uses a Smiths chronometric speed indicator on the generator shaft and a 0.6 m torque arm with a calibrated load cell on the brake during a load bank test. They want to verify shaft power at three operating points: part load at 900 RPM, nominal at 1500 RPM, and overspeed runaway check at 1800 RPM. Measured torque at nominal is 140 N·m.
Given
- Nnom = 1500 RPM
- Tnom = 140 N·m
- Nlow = 900 RPM
- Nhigh = 1800 RPM
- Indicator accuracy = ±0.25 % of reading
Solution
Step 1 — at nominal 1500 RPM and 140 N·m, compute shaft power directly:
That hits the turbine's 22 kW nameplate within the chronometric tachometer's ±0.25% RPM band — about ±55 W of uncertainty from the speed reading alone, which is comfortably below the load cell's ±1% torque uncertainty.
Step 2 — at the low-end operating point of 900 RPM, with measured torque dropping to 95 N·m under reduced jet flow:
This is part-load running. The induction generator is barely above its slip-induced cut-in, and the indicator is still in its accurate band. You can hear the change — the turbine note drops noticeably and the brake squeal eases.
Step 3 — at the high-end runaway check, 1800 RPM with the brake released and torque approaching zero:
At runaway the speed is well above nominal but there's almost no shaft torque, so power collapses. This is exactly what you want to confirm before trusting the over-speed trip — the speed indicator is the primary instrument that tells you whether the trip set point of 1.8× nominal actually works. Above about 2000 RPM the chronometric escapement starts to tick visibly between counts and the reading gets jumpy, so 1800 RPM sits at the upper edge of where you trust the instrument absolutely.
Result
Shaft power at nominal is 22. 0 kW, matching the Gilkes nameplate. The three operating points span 9.0 kW at part load, 22.0 kW at rated, and 0.9 kW at runaway — that range tells you the speed-indicator-plus-torque-arm method tracks the machine cleanly through its whole envelope, with the sweet spot for trustworthy data between 1200 and 1700 RPM. If your measured power comes out 5-10% below predicted, the three most common causes are: (1) the rubber drive tip on the speed indicator has hardened or glazed and is slipping in the shaft dimple, costing 3-8% of reading, (2) the torque arm is not square to the shaft axis — even 5° of skew costs roughly 0.4% per degree of cosine error, or (3) the load cell has not been re-zeroed against the dead weight of the brake arm itself, which on a 0.6 m arm with a 4 kg brake assembly hides about 24 N·m of pre-load.
Choosing the Speed Indicator: Pros and Cons
A speed indicator is one of three families of RPM measurement. The choice between contact hand tachometer, optical/laser tachometer, and permanently fitted shaft encoder depends on accuracy, access, and how often you take a reading.
| Property | Contact Speed Indicator (chronometric) | Optical/Laser Tachometer | Shaft Encoder |
|---|---|---|---|
| Accuracy | ±0.25% of reading | ±0.05% of reading | ±0.01% of reading |
| Speed range | 50-10,000 RPM | 1-100,000 RPM | 0-30,000 RPM (typical) |
| Cost | £200-600 for a Smiths or Hasler | £150-500 for a handheld optical | £300-2000 installed plus signal conditioner |
| Shaft access required | Direct contact with centre dimple | Reflective tape, line of sight only | Permanent coupling to shaft end |
| Reliability over 20 years | High — purely mechanical, repairable | Medium — LED and optics degrade | High if sealed, but electronics obsolete in ~15 years |
| Best application fit | Field surveys, engine tests, heritage machinery | High-speed spindles, hard-to-reach shafts | Continuous monitoring, control feedback |
| Setup time per reading | 10-20 seconds | 30-60 seconds (apply tape) | Zero — always live |
Frequently Asked Questions About Speed Indicator
The drive tip is a 60° rubber cone sized to bottom in a standard centre-drilled hole. On a small shaft (under about 12 mm diameter), the centre dimple is usually shallower than 2 mm, so the cone bears partly on its sides instead of seating cleanly at the apex. That side contact slips slightly under load, particularly above 1500 RPM, and the indicator under-reads.
The fix is to use the smaller-diameter conical adapter that came in the original instrument case — most Smiths and Hasler kits include three tip sizes for exactly this reason. If the dimple itself is shallow or worn, re-cut it with a centre drill before testing.
Always re-verify on commissioning and every major service. Permanently fitted tachometers — especially flyweight types driven off the camshaft via a flexible cable — drift. The cable develops backlash, the cable lubricant dries out, and the flyweight hairspring fatigues. A 25-year-old genset panel tachometer can easily read 4-6% off true RPM, which on a 1800 RPM nominal machine means you're actually running at 1700 or 1900 — both of which give the wrong output frequency.
The discipline on every Lloyd's or DNV survey is to land a calibrated chronometric hand instrument on the crankshaft pulley dimple and compare. If the panel reads more than 1% off, the panel gets adjusted or replaced, not the engine governor.
Pick optical when shaft access is poor, when speeds exceed roughly 6000 RPM, or when the shaft has no usable centre dimple. Above 6000 RPM the rubber tip on a contact instrument heats up and glazes within seconds, and the slip error climbs fast.
Stick with contact when you need ±0.25% accuracy on a slow-to-medium-speed machine and you have a clean centre hole. Optical instruments are vulnerable to ambient light, oily reflective tape that loses contrast, and shaft surfaces that throw stray reflections — all common in a real engine test cell. For a heritage 350 RPM oil engine survey, contact wins every time.
Three likely causes, in order of frequency. First, the centre dimple is worn into a wallowed crater, so the tip walks around and the input rate fluctuates. Re-cut the centre with a 60° centre drill and re-test.
Second, you're holding the instrument off-axis. Anything more than about 3° of misalignment makes the rubber cone skip at the contact point. Brace your wrist against the engine block and watch the needle settle.
Third — and this catches people — the engine genuinely is hunting at ±50 RPM because the governor is reacting to a torsional vibration the human ear can't pick up. Single-cylinder oil engines and two-cylinder diesels often hunt this way. Try the same reading on a multi-cylinder machine and see if the jitter disappears. If it does, the indicator is fine and you've found a governor symptom.
Drive it from a known reference. The cheapest field check is a synchronous AC motor running off mains — a 4-pole motor on 50 Hz turns at exactly 1500 RPM minus a small slip (typically 1450 RPM under no load, 1500 RPM unloaded synchronous). A spindle motor on a lathe with a known belt ratio works too.
If the chronometric reads within ±0.5% of the calculated reference at three different speeds (say 500, 1500, and 3000 RPM), it's still good for field work. If it's off by more than 1% at any point, the escapement needs a service — the pivot jewels and the hairspring are the usual culprits, and a competent instrument repairer can restore it for less than the cost of a new one.
About 100 RPM is the practical lower limit for most centrifugal hand instruments. Below that, the centrifugal force on the flyweights is so weak it falls into the same magnitude as the hairspring's hysteresis and the bearing friction. The needle wobbles because each flyweight pulse from the rotating shaft transfers through the gear train as a discrete kick rather than a smoothed force.
If you're trying to measure something genuinely slow — a 26 RPM wind turbine main shaft, or a 30 RPM pottery wheel — use a chronometric instrument instead. The escapement counts whole revolutions over a fixed gate time, so it doesn't care about instantaneous force. Or simply count revolutions against a stopwatch by eye, which gives you ±1 RPM accuracy at low speeds without any instrument at all.
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.
