A steam trap is an automatic valve that discharges condensate, air and non-condensable gases from a steam line while holding back live steam. Properly sized traps shed condensate at rates from a few kg/h on tracing lines up to several thousand kg/h on large mains, with discharge temperatures sitting just below saturation. The purpose is simple — keep liquid water out of the steam path so cylinders, heat exchangers and turbines see dry steam. You see them on every preserved mill engine main, on the Flying Scotsman's superheater header drains, and on the steam-tracing of every refinery in Grangemouth.
How the Steam Trap Works
A steam trap sits at the low point of a steam line — the drip pocket — and waits for condensate to arrive. Steam carries latent heat. The moment that steam touches a cooler pipe wall, some of it gives up its latent heat and turns back into water. If you let that water sit in the pipe you get water hammer, eroded valve seats, wet steam at the cylinder, and in the worst case a slugged piston that bends a connecting rod. The trap's job is to dump the water and keep the steam, automatically, with no operator input.
Different trap types sense the difference between water and steam in different ways. A float and thermostatic trap uses a ball float that rises with the condensate level and opens a valve below the waterline — it discharges continuously and modulates with load. A thermodynamic trap (the disc trap) uses the velocity difference between flashing condensate and live steam to snap a single hardened disc open and shut, cycling roughly 4 to 10 times a minute. An inverted bucket trap uses buoyancy in reverse — the bucket floats when filled with steam and sinks when filled with water, opening a lever valve. A thermostatic trap uses a bellows or bimetallic element that closes when the temperature approaches saturation.
Get the sizing wrong and you have problems either way. Undersize and condensate backs up into the main, you get carryover, and the engine takes water. Oversize and the trap blows live steam straight to the condensate return, wasting fuel and overheating the return line until the next trap downstream fails. Tolerance on the seat is tight — a thermodynamic disc must seat on a lapped face flat to within roughly 0.5 µm or it will leak across the disc and refuse to snap shut. Bucket traps fail when the prime is lost on a sudden pressure drop. Float traps fail when the float corrodes through, sinks, and the trap locks closed.
Key Components
- Body and Strainer: The cast or forged body holds the working pressure — typically rated PN16 to PN40 for general saturated steam service, higher for superheated. An integral Y-strainer with a 0.8 mm or 1.0 mm perforated screen catches scale and pipe debris before it reaches the seat. Skip the strainer and you will be lapping the seat every 6 months.
- Sensing Element: This is what tells water from steam — a stainless float, an inverted bucket, a hardened disc, a stainless bellows charged with a water-alcohol mix, or a bimetal stack. Each has a characteristic discharge temperature; a balanced-pressure thermostatic element typically opens at roughly 12 °C below saturation.
- Valve Seat and Plug: Hardened stainless or Stellite-faced, lapped to a mirror finish. On a thermodynamic trap the disc and seat are the entire moving valve — surface flatness must be held to about 0.5 µm or the trap leaks. Replacement is normally a complete cartridge swap, not a field re-lap.
- Air Vent: A separate thermostatic element (or the main element on an F&T) that dumps air on start-up. Without it you get air-binding — the trap sees cool air, thinks it has no condensate, and the line never warms through. Heat-up time on a 200 mm main triples without a working air vent.
- Discharge Connection: Pipes condensate to the return main or to atmosphere. Sized for the flash steam that forms when high-pressure condensate drops to return-line pressure — typically 10 to 15 % of the condensate mass flashes at 7 bar to atmospheric, so the discharge pipe is one or two sizes larger than the inlet.
Who Uses the Steam Trap
Steam traps live wherever steam meets a cooler surface and condensate forms. That covers heritage engineering, process plant, building heating, marine engineering and food processing. The trap type follows the duty — modulating loads want F&T, fixed orifice loads want thermodynamic, freezing-risk outdoor lines want inverted bucket. Pick the wrong type and you either lose steam or take water at the cylinder.
- Heritage Steam: Drip-pocket traps on the saturated steam main feeding the 500 ihp Roberts cross-compound engine at Queen Street Mill, Burnley — typically a Spirax Sarco TD52 thermodynamic trap on each end-of-main pocket.
- Mainline Steam Locomotives: Superheater header drains and cylinder drain cocks on preserved A4-class locomotives such as Bittern, where a small thermodynamic trap clears the header during pre-heat before the regulator opens.
- Process Plant: Tracing-line traps on the bitumen lines at the Ineos Grangemouth refinery, where hundreds of small thermostatic traps maintain product temperature in winter.
- District Heating: End-of-main drips on the Pimlico District Heating Undertaking, London — F&T traps handling start-up condensate loads of several hundred kg/h on the larger sub-mains.
- Brewing and Food: Jacket drains on the wort copper at Hook Norton Brewery — F&T traps modulating with the boil load so the copper stays full of steam, not water.
- Marine: Bilge-line and galley-line drains on the SS Shieldhall preserved steamship at Southampton, where inverted bucket traps tolerate the heel and trim of a working hull.
The Formula Behind the Steam Trap
Sizing a steam trap means picking an orifice that passes the worst-case condensate load at the available differential pressure, with a safety factor for start-up. The condensate load itself depends on the heat lost from the pipe to ambient — a function of pipe surface area, insulation thickness and the temperature difference between steam and air. At the low end of a typical heritage main (say a 100 mm bore well-lagged pipe at 4 bar gauge in a covered engine house), running condensate is only a few kg/h per 30 m run. At the high end (a 250 mm bore poorly-lagged main outdoors at 10 bar gauge in winter), you can see 80 kg/h or more on the same 30 m. The sweet spot for trap selection is normally a 2× safety factor on running load, or 3× on a cold start where the entire pipe wall has to be heated from ambient to saturation.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| ṁc | Required condensate discharge capacity of the trap | kg/h | lb/h |
| Qloss | Heat loss from the pipe section served by the trap | kJ/h | Btu/h |
| SF | Safety factor — typically 2 for running load, 3 for cold start | dimensionless | dimensionless |
| hfg | Latent heat of vaporisation at line pressure | kJ/kg | Btu/lb |
Worked Example: Steam Trap in a heritage chocolate refiner steam jacket
You are sizing a float and thermostatic steam trap on the discharge of a recommissioned 1936 Carle Montanari longitudinal chocolate refiner being returned to demonstration running at a heritage confectionery workshop in York, where the refiner's cast-iron jacket runs on saturated steam at 3 bar gauge to hold conching temperature at 50 °C, the jacket surface area is 4.2 m², and the trustees want trap capacity verified at slow Sunday demonstration running, nominal conching duty, and a brisk start-up burst from cold before the public open day.
Given
- Pline = 3 bar gauge
- hfg = 2133 kJ/kg at 4 bar abs
- Ajacket = 4.2 m²
- Unom = 12 W/m²·K nominal heat transfer to chocolate
- ΔT = 94 K (steam 144 °C to chocolate 50 °C)
- SFrun = 2 running safety factor
- SFstart = 3 start-up safety factor
Solution
Step 1 — compute the nominal heat transfer rate from the jacket to the chocolate at full conching duty. This is what the trap has to clear as condensate once the refiner is in steady operation:
Step 2 — convert that heat load into a condensate mass flow at line pressure, applying the running safety factor of 2 to cover load swings as fresh cold cocoa mass is added to the refiner:
Step 3 — at the low end of the typical operating range, a slow Sunday demonstration with the refiner barely turning and chocolate already up to temperature, heat transfer drops to roughly 30 % of nominal because ΔT collapses as the chocolate sits near saturation of its own. Running load is then about 4.8 kg/h. A trap sized for 16 kg/h will short-cycle on this duty but will not back up — the F&T modulates cleanly down to roughly 5 % of rated capacity.
Step 4 — at the high end, cold start from ambient before the open day. The jacket and the cast-iron mass of the refiner (roughly 380 kg of iron) must be heated from 15 °C to 144 °C. Lump that heat in over a 20-minute warm-up and apply the 3× start-up factor:
ṁc,start = (73 530 × 3) / 2133 = 103 kg/h peak
That start-up peak is the sizing case, not the running case. A trap rated 16 kg/h will badly under-perform on cold start and the operator will see condensate gulping back into the main.
Result
Nominal running load is 16 kg/h; cold-start peak is 103 kg/h; lazy-Sunday running is roughly 4. 8 kg/h. Pick a Spirax Sarco FT14-4.5 or equivalent F&T trap rated around 110 kg/h at 3 bar differential — it covers the start-up peak with margin and still modulates cleanly at the Sunday demonstration load because F&T traps follow load down to a few percent of rated. If the measured discharge runs hot and continuous when it should be cyclic, suspect three failure modes: (1) the air vent thermostatic element has failed open and is passing live steam, identifiable by a constantly hot return line; (2) the float has corroded and partially flooded, sinking off-seat and locking the main valve open; or (3) the upstream Y-strainer is partially blocked with mill scale, starving the trap and giving the symptom of an under-sized unit when the trap itself is fine.
Steam Trap vs Alternatives
Trap selection is duty-driven. A modulating heat-exchanger load wants different behaviour from a fixed drip-pocket load, and an outdoor freezing-risk line wants different behaviour again. The four common trap types differ on capacity, response to load swings, tolerance to dirt and superheat, and what happens when they fail.
| Property | Float & Thermostatic (F&T) | Thermodynamic (Disc) | Inverted Bucket |
|---|---|---|---|
| Typical capacity range | Up to ~20 000 kg/h on large units | 5 to ~3000 kg/h | 5 to ~5000 kg/h |
| Response to varying load | Modulates continuously, excellent on heat exchangers | On/off cycling 4–10 times/min, poor on swinging loads | On/off, moderate on swinging loads |
| Suitability for superheated steam | Poor — float can be damaged by dry superheat | Excellent — disc handles superheat directly | Poor — loses prime under superheat |
| Failure mode | Usually fails closed (flooded float) | Usually fails open (worn disc, blows steam) | Usually fails open (lost prime) |
| Tolerance to freezing | Poor — body holds water | Excellent — small water content, self-draining | Poor — bucket holds water |
| Service life on clean steam | 8–12 years typical | 3–5 years before disc replacement | 10–15 years typical |
| Initial cost (DN25 nominal) | £££ | ££ | £££ |
Frequently Asked Questions About Steam Trap
Almost certainly yes, and the usual cause is wire-drawing of the seat. A thermodynamic trap snaps shut when flash steam builds pressure in the disc chamber and forces the disc onto the seat. If the seat face has been eroded by particulate or wet steam, the chamber leaks and the disc cannot hold — it lifts almost immediately, dumps a tiny slug, snaps shut, leaks, lifts again. You hear it as a rapid chatter rather than a clean tick-tick.
Replace the disc-and-seat cartridge. Field lapping rarely restores the 0.5 µm flatness needed and you will be back inside 6 months. If the new cartridge also chatters within weeks, fit a separator upstream — wet steam is your real problem.
The thermostatic air vent in the F&T is doing its job a bit too well. On cold start the line is full of air, the air vent opens, and as soon as the air clears the vent should snap shut on the rising temperature. If the bellows charge has weakened (typical after 10+ years) the vent stays partly open into the warm-up phase and you get live steam to return until the body fully heats through.
Quick check: feel the air-vent connection 5 minutes after start-up. If it is still hotter than the trap body, the bellows is gone. Replace the air-vent element — it is a separate cartridge from the main float valve.
Thermodynamic. A drip pocket sees a small, fairly steady running condensate load and a modest start-up surge. That is the textbook duty for a TD trap — small, cheap, tolerates superheat if the engine ever runs slightly dry, fails open so you see the failure as steam loss rather than as water in the cylinder, and has no float to corrode if the engine is laid up over winter.
F&T makes sense on a process load that modulates with demand — a wort copper, a calorifier, a chocolate jacket. On a drip pocket the F&T is overkill and the float adds a lay-up corrosion failure mode you do not need.
Differential pressure, almost every time. Trap capacity tables are quoted at a specific ΔP across the trap. If your condensate return line runs uphill, or if the return main is pressurised by other traps discharging into it, the actual ΔP at your trap can be a fraction of what you assumed. Capacity falls roughly with √ΔP — halve the differential and you lose about 30 % of capacity.
Measure the return-line pressure with a gauge. If it is non-zero, redo the sizing at the actual differential, not at line pressure. On long return runs you may need a pumping trap or a separate condensate pump rather than a bigger orifice.
Because high-pressure condensate at saturation cannot stay liquid when it drops to atmospheric pressure — some of it has to flash off as steam to shed the surplus enthalpy. From a 7 bar main discharging to atmosphere, roughly 13 % of the condensate mass flashes. That is a lot of visible vapour even when the trap is working perfectly.
The diagnostic is timing, not appearance. A healthy thermodynamic trap shows discrete puffs every 6 to 15 seconds. A continuous plume that never stops is live steam blow-through. An ultrasonic listener or a contact pyrometer on the discharge pipe gives a definitive answer — flash sits at 100 °C, blow-through sits at saturation.
No, and this is a common cost-saving mistake on heritage refits. Condensate forms continuously along a steam main and gravity does not move it sideways through a horizontal pipe — it sits as a film on the bottom of the bore until something pushes it. Steam velocity above roughly 25 m/s will carry that film as droplets, and you get water hammer at the next bend.
Rule of thumb: drip pockets every 30 to 50 m on a horizontal main, every change of direction, and immediately upstream of any control valve or engine stop valve. Each pocket gets its own trap, sized for the local condensate load. One big trap at the far end will not save the cylinder.
References & Further Reading
- Wikipedia contributors. Steam trap. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
