The Swape — also called the New Engine Sweep — is a small auxiliary hydraulic device that continuously bleeds a metered slipstream of coolant from the highest point of an engine cooling jacket back to the header tank. The concept traces to a 1962 patent assigned to Perkins Engines in Peterborough for their P-series diesel jackets. It works by using jacket pressure itself to drive a calibrated 2-4 L/min flow through a swept-orifice cartridge, dragging trapped air and combustion gas out with the coolant. The outcome is no airlock at startup and zero cavitation pitting on cylinder liners, even after 20,000 hours service.
Swape Interactive Calculator
Vary the swept-orifice bore and jacket pressure range to see the coolant bleed flow band through a Swape cartridge.
Equation Used
The calculator estimates the Swape bleed flow from the calibrated orifice bore and jacket gauge pressure. The coefficient is set so the article example range of about 1.8-2.4 mm bore at 0.7-1.0 bar produces the stated 2-4 L/min sweep band.
- Empirical swept-orifice coefficient is calibrated to the article example band.
- Coolant behaves like water and the header tank is near atmospheric pressure.
- Pressure is gauge pressure across the Swape return path.
- Return line losses are included in the empirical coefficient.
Inside the Swape or New Engine Sweep
The Swape solves one specific problem: air gets trapped at the top of an engine block when you fill it, when a head gasket weeps combustion gas, or when coolant flashes around hot exhaust valve seats. That trapped air collapses the local pressure, the water pump cavitates, and you get pitting on the wet liner OD within a few thousand hours. A Swape sits on the highest tapping of the jacket — usually the thermostat housing or a dedicated cast boss — and continuously sweeps a small flow of coolant plus any entrained gas back to the header tank where the gas can disengage and vent.
The sweep is not a vent. A vent opens when pressure exceeds a threshold. The Swape is always flowing. Inside the cartridge you have a calibrated orifice — typically 1.8 to 2.4 mm diameter — sized so that under nominal jacket pressure of 0.7 to 1.0 bar gauge, the bleed flow sits between 2 and 4 L/min. That's enough to scavenge entrained air without robbing the main circuit. If the orifice is undersized, gas accumulates faster than it bleeds and you get airlock at the head. If it's oversized, you starve the radiator circuit and overheat under load. The bore must be 2.0 mm ± 0.05 mm on a Perkins 1106 jacket — not 1.9, not 2.1.
Failure modes are predictable. Scale buildup over the orifice cuts effective bore and stalls the sweep — you'll notice it as a slow rise in header tank temperature swing on cold starts. A cracked cartridge body lets jacket pressure dump to atmosphere through the breather, and the system can't hold the 0.5 bar minimum needed to suppress nucleate boiling. Wrong cartridge for the engine spec — common on aftermarket replacements — and you've either choked the bleed or opened it wide.
Key Components
- Sweep Cartridge: Brass or stainless body holding the calibrated orifice. Threads into the jacket high point with a bonded seal washer torqued to 25 Nm on a typical M16 boss. The cartridge is matched to engine displacement — a 4-cylinder 4 L diesel takes a 2.0 mm bore, a 12 L six takes 2.8 mm.
- Calibrated Orifice: Drilled and reamed to ±0.05 mm. This single dimension sets the entire bleed flow. Anything tighter than design starves the sweep; anything looser dumps too much main-circuit flow and the radiator can't reject heat at full load.
- Return Line: Small bore hose, 8 to 10 mm ID, routed from the cartridge outlet to the header tank above the static coolant line. The line must rise continuously — any low spot traps air and defeats the scavenging effect.
- Header Tank Disengagement Volume: The tank needs at least 2 L of free vapour space above the static fill line for the swept gas to disengage before the coolant returns to the pump suction. Undersized tanks let microbubbles recirculate and you'll see foam at the filler cap.
- Anti-Siphon Check: A light spring check (cracking pressure 0.05 bar) on some installations prevents back-siphoning when the engine shuts down hot and pressure inverts. Not all engines need it — Caterpillar 3406 jackets do, Cummins 6BT typically don't.
Where the Swape or New Engine Sweep Is Used
You see the Swape concept anywhere an engine runs hot, hard, and continuously, and where airlock at the head would mean liner failure within days. It's not a passenger-car part — modern automotive cooling systems handle deaeration with the bottle and a continuous bleed line. The Swape lives in marine, off-highway, stationary genset, and heritage diesel territory where jacket geometry traps gas and you need active scavenging.
- Marine propulsion: Perkins Sabre M215C marine diesels in 12 m workboats — the Swape sits on the heat exchanger inlet to clear gas from the wet exhaust manifold jacket where coolant boils locally at full load.
- Stationary power generation: Cummins KTA19-G4 gensets at telecom backup sites — continuous sweep prevents nucleate boiling around exhaust valve bridges during 8-hour load tests.
- Off-highway equipment: Caterpillar D8T dozer engines in oil sands operations near Fort McMurray Alberta where ambient swings and steep grades cause coolant slosh and gas trapping at the rear cylinder bank.
- Heritage rail restoration: English Electric 8SVT Mk2 prime movers in restored Class 37 locomotives at the Severn Valley Railway — original Swape cartridges from the 1960s are still being remanufactured for these jackets.
- Industrial pumping: Detroit Diesel 12V-71 engines driving fire-water pumps on offshore platforms — the sweep keeps the head clear during weekly test starts where the jacket goes from 5°C to 90°C in 90 seconds.
- Agricultural prime movers: Massey Ferguson 8740S tractors operating at sustained 95% load during silage harvest — Swape installation on the AGCO Power 84 AWI engine eliminates the airlock that occurs when the cab climate system pulls coolant heat at idle.
The Formula Behind the Swape or New Engine Sweep
The bleed flow through a Swape orifice follows the standard incompressible orifice equation. What matters in practice is how the flow scales with jacket pressure across the operating range. At cold idle the jacket might only sit at 0.2 bar gauge — sweep flow drops to roughly 1 L/min, which is barely enough but acceptable because gas generation is also low. At nominal cruise of 0.8 bar you get the design 3 L/min — the sweet spot where scavenging matches gas evolution. Push to relief-valve setting of 1.4 bar during a full-load thermal transient and flow climbs above 4 L/min, which starts robbing the main circuit. The orifice bore is sized for the nominal point.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Q | Bleed flow through the sweep orifice | m³/s | GPM |
| Cd | Discharge coefficient (typically 0.62 for a sharp-edged orifice, 0.80 for a reamed cartridge) | dimensionless | dimensionless |
| A | Cross-sectional area of the calibrated orifice | m² | in² |
| ΔP | Pressure differential between jacket and header tank | Pa | psi |
| ρ | Coolant density (typically 1040 kg/m³ for 50/50 ethylene glycol) | kg/m³ | lb/ft³ |
Worked Example: Swape or New Engine Sweep in a Perkins 1106D-E66TA marine retrofit
You are sizing a replacement Swape cartridge for a Perkins 1106D-E66TA 6.6 L marine diesel on a fishing trawler refit at a yard in Peterhead Scotland. The original cartridge is corroded and the owner reports steam pockets at the head after long trolling runs. Header tank sits 0.5 m above the jacket high point, design jacket pressure is 0.8 bar gauge at cruise, coolant is 50/50 ethylene glycol at 1040 kg/m³. You need a sweep flow of 3 L/min at nominal to match the gas evolution rate quoted in the Perkins service bulletin.
Given
- ΔPnom = 80000 Pa (0.8 bar)
- ρ = 1040 kg/m³
- Cd = 0.80 dimensionless (reamed cartridge)
- Qtarget = 5.0 × 10⁻⁵ m³/s (3 L/min)
Solution
Step 1 — rearrange the orifice equation to solve for required area at nominal jacket pressure of 0.8 bar:
Step 2 — compute the velocity term first, then the area:
Step 3 — convert area to bore diameter:
Round to the nearest standard reamer size — call it 2.5 mm. Now check the operating range. At cold idle with jacket pressure of only 0.2 bar gauge, ΔP drops to 20000 Pa and the flow falls to Qlow = 0.80 × 4.91 × 10⁻⁶ × √(2 × 20000 / 1040) = 1.5 L/min. That is half the nominal sweep but it's adequate — gas evolution at idle is also low because the head isn't seeing combustion stress. At full-load thermal transient, jacket pressure can hit the relief-valve setting of 1.4 bar and ΔP climbs to 140000 Pa, giving Qhigh = 4.0 L/min. That's the upper bound — any larger orifice and you'd start pulling enough flow at high jacket pressure to noticeably reduce radiator throughput, with measurable temperature rise on the cylinder head outlet thermocouple.
Result
The required orifice diameter is 2. 5 mm, giving a nominal bleed of 3.0 L/min at 0.8 bar jacket pressure. In practice that's the rate where you can crack the header tank cap after a 30-minute run and hear a steady soft hiss but never see foam — that's what good scavenging sounds like. Across the operating range, flow swings from 1.5 L/min at cold idle to 4.0 L/min at relief pressure, with the design sweet spot sitting right at cruise where the engine spends 90% of its life. If your measured sweep flow comes in below 2 L/min at nominal pressure, the most common causes are: (1) silicate scale dropout from old coolant narrowing the orifice — pull the cartridge and check bore with a pin gauge, anything below 2.3 mm needs replacement; (2) a partially blocked return line from a kinked hose at the header tank fitting — symptom is jacket pressure reading correct but tank shows no return flow when warm; (3) wrong cartridge installed, often a 1106A-series part fitted to a 1106D jacket where the bore spec changed from 2.5 mm to 2.0 mm in the 2008 revision.
Choosing the Swape or New Engine Sweep: Pros and Cons
The Swape is one of three common ways to deaerate an engine cooling jacket. Each approach trades continuous bleed flow against system simplicity and against how aggressively it handles combustion gas ingress. The right choice depends on duty cycle, jacket geometry, and whether you expect occasional head gasket weep.
| Property | Swape / Engine Sweep | Passive Vent Line | Active Scavenge Pump |
|---|---|---|---|
| Bleed flow at nominal pressure | 2-4 L/min continuous | 0 L/min until vent opens | 5-15 L/min driven |
| Gas removal capacity | Handles up to 0.5 L/min combustion gas ingress | Only handles startup fill air | Handles severe gas ingress, including cracked-head conditions |
| Installed cost (USD, 2024) | $40-120 for cartridge plus fittings | $15-30 for tee and hose | $400-900 for pump, motor, controller |
| Reliability / MTBF | 20,000+ hours, no moving parts | 30,000+ hours, simplest of the three | 8,000-12,000 hours limited by pump seal |
| Main-circuit flow penalty | 3-5% of pump output | Zero except during venting | 10-20% diverted while pump runs |
| Best application fit | Marine, stationary diesel, off-highway under continuous load | Passenger cars, light commercial, intermittent duty | Heavy-duty engines with known gasket issues, mining haul trucks |
| Sensitivity to coolant condition | High — orifice scales over with poor coolant | Low — large flow path | Medium — pump seal sensitive to silicate |
Frequently Asked Questions About Swape or New Engine Sweep
Almost always silicate dropout. Fresh inhibited coolant carries dissolved silicate corrosion inhibitor that can precipitate on any surface where there's a velocity discontinuity — the orifice edge is exactly that. The deposit is soft and grey-white, and it doesn't shrink the bore so much as it changes the discharge coefficient by roughening the edge.
Diagnostic check: pull the cartridge after the first 100 hours on a new coolant fill and look at the orifice edge under a 10× loupe. If you see whitish haze, switch to an OAT (organic acid technology) coolant — they don't carry the silicate package and the problem disappears. Caterpillar ELC and Shell Rotella ELC are the common fixes.
You can, but you need a reamer not a drill, and you need to know the original spec exactly. Drill bits leave a tapered, rough bore that gives you maybe 0.55 discharge coefficient instead of 0.80, so a 2.5 mm drilled hole flows like a 2.0 mm reamed one. The flow won't match the original calibration and you'll either undersweep or oversweep depending on which way you missed.
Use a chucking reamer in a hand-held pin vise, work up in 0.1 mm steps from undersize, and verify final flow on a bench rig with a measured pressure head before installing. For an English Electric 8SVT or similar heritage application, the original drawings spec the bore to ±0.025 mm — that's machine-shop territory, not a workshop drill press.
The deciding factor is gas ingress rate, not engine size. A Swape handles up to about 0.5 L/min of combustion gas crossing into the jacket — that covers a healthy engine and a marginally weeping head gasket. Above that, you need active scavenging because the orifice can't pull gas faster than the jacket pressure differential allows.
Practical rule: if the engine has any history of head gasket weep, pinhole liner cracks, or runs above 0.5 bar mean effective combustion pressure leak-down on a cylinder test, go with an active pump. For a clean Cummins KTA19 or similar with good leak-down numbers, the Swape is simpler, cheaper, and lasts longer.
Foam at the filler means the disengagement volume in the tank is undersized for the swept flow rate. The coolant is returning to the pump suction before the entrained microbubbles have time to rise out. You need at least 2 L of free vapour space above the static fill line, and the return line must discharge above that line, not below it.
Quick check: measure the vertical distance from the static fill mark to the return line entry point. If the return enters below the coolant level, gas re-entrains immediately. Re-route the return so it discharges into the air space, even if it means adding a short standpipe inside the tank.
Classic oversweep symptom. At idle, jacket pressure is low so the bleed flow is small and the radiator gets nearly full pump output — engine runs cool. At full load, jacket pressure climbs and so does the bleed flow, but it climbs faster than the radiator can compensate because the orifice flow scales with √ΔP while pump output is roughly linear with RPM. You end up diverting 8-10% of flow to the header tank when you can least afford it.
Fix is to swap the cartridge to the correct bore. There's no clever workaround — the orifice equation is what it is.
For under 5 minutes at idle, yes. Beyond that, no. With the sweep blocked, any combustion gas that crosses the head gasket has nowhere to go, and it accumulates at the jacket high point. Within 10-15 minutes on a marginal engine you can build a gas pocket large enough to airlock the water pump suction, at which point local boiling at the exhaust valve bridges starts within seconds.
Better diagnostic approach: tee a clear hose into the return line and watch for bubbles directly. A healthy engine shows occasional fine bubbles, a weeping gasket shows a steady stream that increases with throttle.
References & Further Reading
- Wikipedia contributors. Internal combustion engine cooling. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
