The Forrester rotary engine is a 19th-century rotary-piston steam engine in which an eccentrically mounted piston rotates inside a cylindrical casing, sealed against a sliding spring-loaded abutment that separates the high-pressure inlet from the exhaust. Steam pressure acting on the offset piston face produces direct rotary torque without a crank, connecting rod, or flywheel-stored reciprocation. The design eliminated the dead-centres and reciprocating mass of beam and horizontal engines, giving compact direct drive for fans, pumps, and small mill duties. A typical Forrester unit ran 200-400 RPM at 40-80 psi delivering 2-15 indicated horsepower.
Forrester Rotary Engine Interactive Calculator
Vary steam pressure, effective piston area, eccentricity, and speed to see direct rotary torque, force, and power.
Equation Used
The Forrester rotary engine torque estimate treats steam pressure as a tangential force on an effective area, acting through the piston eccentricity. Force is P times A, torque is force times e, and horsepower is found from torque and rpm.
- Pressure is effective working pressure acting on the unbalanced face.
- Area is the effective steam-loaded piston face area.
- Eccentricity is the offset lever arm from casing center to piston center.
- Friction, leakage, cutoff, and expansion losses are not included.
How the Forrester Rotary Engine Works
The Forrester rotary engine works by mounting a solid cylindrical piston off-centre inside a slightly larger cylindrical casing. The piston touches the casing wall along one line — the sealing line — and rotates around its own axis, which sits offset from the casing axis. A radially sliding abutment, spring-loaded or steam-loaded, presses against the piston surface and divides the crescent-shaped working space into a high-pressure side and an exhaust side. Steam admitted on the high-pressure side acts on the unbalanced piston face area between the sealing line and the abutment, producing a net tangential force and continuous torque. There is no reciprocation, no dead-centre, and no crank — the output shaft turns directly with the piston.
Geometry is everything on this engine. The piston-to-casing eccentricity sets the working volume per revolution, typically 8-15% of the casing radius. The abutment must track the piston surface within roughly 0.05 mm of clearance over its full stroke, otherwise blow-by between high and low pressure sides destroys efficiency. If you notice indicated horsepower dropping 30-40% below predicted on an indicator card, the abutment tip seal is the first place to look — followed by the end-cover packing, which is where most Forresters leak after 500 hours of steaming. Wear on the abutment guide slot causes the abutment to cock and lift on one edge, which scores the piston surface in a characteristic helical pattern you can feel with a fingernail.
Why use a Forrester at all when a horizontal mill engine delivers more power per cubic foot? Because the Forrester gives smooth direct rotary output at modest speed, with no reciprocating masses to balance, no crosshead to align, and a footprint roughly a third of an equivalent reciprocating engine. The trade is sealing — abutment blade clearance, end-cover packing, and piston circularity all have to be held to tolerances tighter than a comparable reciprocating engine, and that is what limited Forresters to small duties and ultimately killed the type when steam turbines and electric motors arrived.
Key Components
- Eccentric Piston (Rotor): Solid cylindrical drum mounted on the output shaft, offset from the casing centre by typically 6-12 mm on a 150 mm casing bore. The piston outer surface must be ground concentric with its own axis to within 0.02 mm and held circular within 0.03 mm — wear or out-of-round causes the sealing line to hunt, dropping volumetric efficiency.
- Cylindrical Casing: The fixed bore the piston runs inside. Surface finish must reach Ra 0.4 µm or better on the working portion. Cast iron is standard; the casing is bored after the end covers are fitted to keep the abutment slot perpendicular to the piston axis within 0.1°.
- Sliding Abutment (Blade): Radially sliding bar housed in a slot in the casing wall, pressed against the rotating piston by a coil spring or by steam pressure on its back face. Tip clearance against the piston must hold 0.03-0.05 mm; the abutment stroke equals twice the eccentricity. Wear here is the single biggest killer of efficiency.
- Steam Admission Port: Drilled through the casing immediately on the high-pressure side of the abutment. Port timing is fixed by geometry — there is no valve gear. Port area sets the maximum mass flow and therefore the upper RPM limit; under-port a Forrester and it chokes above 250 RPM.
- Exhaust Port: Sits on the opposite side of the abutment, leading to atmosphere or a condenser. Generous port area matters more on exhaust than admission — a restricted exhaust raises back-pressure on the working face and you see it as a fat exhaust loop on the indicator card.
- End Covers and Packing: Bolted plates closing each end of the casing, carrying the shaft bearings and the piston-end gland packing. End-cover face flatness must be within 0.02 mm to seal against the piston end faces; this is the second-most-common leak path after the abutment.
- Output Shaft and Bearings: Plain white-metal journals of 40-60 mm diameter carry the shaft directly out to the driven load. No flywheel is strictly needed because torque is continuous, though a small inertia helps smooth the impulse as the abutment crosses the inlet port.
Real-World Applications of the Forrester Rotary Engine
Forrester rotary engines found a niche between roughly 1870 and 1910 wherever a small, smooth, direct rotary drive was wanted and the duty did not justify a full reciprocating engine. They drove fans, centrifugal pumps, small line shafts, and dynamos in mills, ships, and gasworks. Most surviving examples are now in heritage collections, where their compact form makes them easy to display in working order.
- Textile Mills: Driving auxiliary fans and small line shafts in carding rooms — examples ran at the George Forrester & Company works in Liverpool through the 1880s.
- Marine Auxiliaries: Cabin ventilation fans and stokehold blowers on late-Victorian merchant steamers, where the compact footprint suited cramped engine-room corners.
- Gasworks: Direct drive for exhauster fans pulling town gas through purification beds at municipal works such as the Beckton Gasworks ancillary plant.
- Heritage Pumping Stations: Restored auxiliary drives at sites like Crossness Pumping Station in London, where rotary engines powered small water and oil pumps independent of the main beam engines.
- Small Electrical Generation: Driving early DC dynamos at country houses and isolated workshops in the 1890s, typically 1-3 kW sets running at 300-400 RPM.
- Brewery and Distillery: Wort transfer pump drives and fan duties at heritage breweries, where smooth rotary output avoided the pulsation a reciprocating engine puts into a centrifugal pump.
The Formula Behind the Forrester Rotary Engine
The indicated horsepower of a Forrester rotary engine comes straight from mean effective pressure, swept volume per revolution, and shaft speed. What matters to a practitioner is how this scales across the working range. At the low end of typical operation — say 150 RPM — the engine loafs and sealing losses dominate, so indicated and brake power diverge sharply. At nominal 250-300 RPM the engine sits in its sweet spot, where port flow, abutment tracking, and bearing friction all balance out. Push past 400 RPM and the abutment starts skipping on the piston surface, MEP collapses because steam can't fill the working space fast enough through the fixed admission port, and you lose power even though shaft speed climbs.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| IHP | Indicated horsepower developed in the working space | W (× 745.7) | hp |
| Pm | Mean effective pressure averaged over one revolution | kPa | psi |
| Vs | Swept volume per revolution = 2 × π × e × L × r, where e is eccentricity, L is piston length, r is mean radius | m³ | ft³ |
| N | Shaft rotational speed | rev/min | RPM |
| 33,000 | Conversion constant for ft·lbf/min to horsepower (imperial form) | — | ft·lbf/min/hp |
Worked Example: Forrester Rotary Engine in a restored Forrester rotary fan drive
You are sizing the indicated horsepower of a recommissioned 1894 Forrester rotary engine being returned to demonstration steaming at a heritage gasworks museum in the West Midlands, where it will drive a small purifier-bed exhauster fan off saturated steam at 60 psi inlet exhausting to atmosphere. The engine has a 200 mm casing bore, a 250 mm piston length, 10 mm eccentricity, and you want to know how it will behave across the demonstration speed range of 150 to 400 RPM with the indicator card showing roughly 35 psi MEP at the nominal operating point.
Given
- Casing bore D = 200 mm
- Piston length L = 250 mm
- Eccentricity e = 10 mm
- Mean radius r = 95 mm
- Pm at nominal = 35 psi
- N nominal = 275 RPM
Solution
Step 1 — compute swept volume per revolution from the eccentric-piston geometry:
Step 2 — at nominal 275 RPM with 35 psi MEP, compute indicated horsepower in imperial form:
That is the demonstration sweet spot — the indicator card closes cleanly, the abutment tracks the piston without chatter, and the fan pulls a steady 200 Pa across the purifier bed.
Step 3 — at the low end of the demonstration range, 150 RPM, MEP typically falls to around 28 psi because port-throttling losses ease but sealing losses dominate at low speed:
At 150 RPM the engine sounds asthmatic — you can hear the abutment tick once per revolution as it crosses the inlet port, and steam consumption per IHP·hr nearly doubles compared to nominal because end-cover and abutment leakage take a fixed slice of the steam regardless of speed.
Step 4 — at the high end, 400 RPM, the fixed admission port starts to choke and MEP drops to roughly 22 psi:
So power actually falls past 350 RPM despite the higher shaft speed — classic port-choking behaviour on a fixed-port rotary. You will also hear the abutment start to skip above 380 RPM as its spring loading struggles to hold the blade against the piston through the inlet-port crossing.
Result
Nominal indicated horsepower at 275 RPM and 35 psi MEP works out to roughly 0. 73 hp — modest, but exactly the right size for a demonstration purifier-bed fan and well within the original 1894 nameplate rating. At 150 RPM you get about 0.32 hp with audibly poor sealing economy, and at 400 RPM you only reach 0.67 hp because port-choking and abutment skip both bite hard above 350 RPM, so the sweet spot sits clearly around 250-300 RPM. If your measured IHP comes in 25% below predicted at the nominal point, check three things in order: (1) abutment spring preload, because a tired spring lets the blade lift on the inlet-port crossing and dumps live steam straight to exhaust; (2) end-cover gland packing, where graphited cotton older than 200 service hours typically leaks 10-15% of mass flow past the piston ends; and (3) piston circularity on a dial gauge — anything beyond 0.05 mm out-of-round on a 200 mm bore creates a hunting sealing line and shows up as a ragged top edge on the indicator card.
Choosing the Forrester Rotary Engine: Pros and Cons
Forrester rotary engines compete in the same niche as small horizontal mill engines and, on the museum bench today, against modern electric drives. The honest comparison sits on speed range, sealing reliability, and how forgiving each is to imperfect maintenance.
| Property | Forrester Rotary Engine | Small Horizontal Mill Engine | Electric Motor |
|---|---|---|---|
| Typical operating speed | 200-400 RPM direct | 60-200 RPM, often geared up | 750-3000 RPM |
| Power range | 1-15 IHP | 5-150 IHP | 0.1-500 kW |
| Sealing reliability between overhauls | 300-500 hours before abutment service | 2000-5000 hours before piston-ring service | 20,000+ hours bearing-limited |
| Footprint per IHP | ~0.3 m²/hp | ~0.9 m²/hp | ~0.05 m²/hp |
| Tolerance sensitivity | Critical — 0.05 mm abutment clearance | Moderate — 0.1-0.2 mm ring gap | Sealed — not user-serviceable |
| Smoothness of rotary output | Continuous, no dead-centre | Pulsating, requires flywheel | Continuous, electrically smoothed |
| Restoration cost (heritage context) | £8-20k typical | £25-80k typical | £200-2,000 typical |
Frequently Asked Questions About Forrester Rotary Engine
The vibration above 300 RPM is almost always abutment skip, not rotor imbalance. As speed rises, the abutment has less time to follow the piston profile through the inlet-port crossing, and if the spring preload or steam back-loading is marginal, the blade momentarily lifts off the piston. That dumps live steam to exhaust for a few degrees of rotation and produces a sharp pressure pulse you feel as vibration through the casing.
Diagnostic check: take an indicator card at the speed where vibration starts. If the top of the diagram develops a sawtooth or ragged edge that wasn't there at lower speed, abutment skip is confirmed. Fix is usually a stiffer abutment spring or, on steam-loaded designs, opening up the back-loading port slightly.
If the duty is below 5 hp, the original engine bed will take a Forrester, and you have a competent fitter who can hold 0.05 mm tolerances on the abutment slot, the Forrester wins on footprint and historical authenticity. Above 5 hp or where the engine will run more than 200 hours a year, a small horizontal engine almost always wins on operating cost — its sealing tolerance budget is roughly 4× more forgiving and ring renewals come at 2,000-hour intervals rather than 300-500 hours for an abutment service.
The other deciding factor is steam supply. Forresters are unforgiving on wet steam — water hammer through the abutment slot will cock the blade and score the piston in a single bad start. If your boiler doesn't deliver consistently dry steam, lean horizontal.
That gap is unusually large for a Forrester. Mechanical efficiency on a healthy rotary should sit around 80-85%, so a 35% loss means you're losing roughly 20% extra somewhere mechanical. The two suspects in order of likelihood are: shaft gland friction (Forresters use through-shaft glands that are often over-tightened during reassembly — back the gland nuts off until you see a faint steam wisp at the gland, then nip up just enough to stop it), and end-cover face contact, where if the end covers were bolted with the piston not centred, the piston end faces rub on the cover and consume real torque.
Run the engine off-load on air at 20 psi and listen. A healthy Forrester is nearly silent on air; rubbing or hissing localises the loss immediately.
You'll get more power, but not proportionally, and you'll shorten abutment life sharply. Higher inlet pressure raises the unbalanced steam load on the back of the abutment, which on most Forresters is the same surface the spring acts on. Spring deflection increases, blade-tip contact force on the piston rises, and abutment wear scales roughly with the square of contact force.
Rule of thumb on heritage Forresters: stay within 110% of the original nameplate pressure unless you've replaced the abutment with a hardened modern equivalent and re-profiled the spring. Going from 60 to 90 psi is a 50% increase and will eat an abutment in maybe 80-100 service hours.
That spike is the abutment crossing the inlet port with the wrong timing. On a Forrester there's no valve gear — port timing is fixed by where the holes are drilled in the casing. If the casing has been re-bored during restoration without restoring the original port-edge geometry, or if the abutment slot has been moved by even 1-2°, you get a brief moment where the inlet port is open to both sides of the abutment simultaneously. Steam blasts past the abutment, pressure on the working face spikes, then settles to the steady admission line.
Fix is to check the port-edge-to-abutment-slot angle against original drawings. If the engine has no surviving drawings, set the trailing edge of the inlet port to coincide with the leading face of the abutment within 0.5° of rotation.
Strictly no, but practically yes for most heritage installations. The torque output does have a ripple — peak torque happens when the piston face is fully exposed to inlet pressure, minimum when the abutment is mid-crossing. The ripple is typically ±15% of mean torque on a single-abutment Forrester, which is small compared to a single-cylinder reciprocating engine but still enough to put audible pulsation into a centrifugal fan.
A modest flywheel — say 30-50 kg·m² of inertia on a 5 hp engine → knocks the ripple down to ±3% and makes the drive feel electric. Many surviving Forresters were built with integral flywheel pulleys for exactly this reason.
References & Further Reading
- Wikipedia contributors. Rotary engine. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
