What a Steam Trap Is Actually Deciding

Why this matters

A steam trap has no way to detect steam. It detects one physical proxy that usually correlates with steam, and there are only three of them in the whole trade: density, temperature, and the behaviour of flashing condensate at a seat. Every trap on every job is running exactly one of those tests. That is why a trap that is perfectly healthy can be perfectly wrong for a load, why a batch of new traps can make a system worse, and why "the trap failed" is often a misdiagnosis of "the proxy was fooled." Ask which test the trap in front of you is running before you ask whether it is broken.

Before you test or change a trap

A trap body sits at line temperature and its discharge is live. Never test a trap by cracking a fitting, loosening a union, or feeling the body or the outlet with a hand. Where a trap has a proper test valve, operate it only if that valve is piped to a rated discharge point or a drain, stand beside the valve rather than in the plane of the outlet, and wear face protection under 29 CFR 1910.133 with sleeves down.

To remove or replace a trap: isolate the inlet and the outlet, lock and tag both under 29 CFR 1910.147, vent the section to atmosphere through a drain you operate from the side, confirm zero gauge pressure on a gauge you can read, and let the body cool to a temperature you have measured before you break a joint. Condensate trapped between two closed valves is above its atmospheric boiling point and flashes out of the first opening you make.

Diagnose with a non-contact or clamp method rather than by feel: an ultrasonic probe with its stinger held against the trap body through the instrument's own handle, or a contact temperature probe on a cleaned spot with your hand and the lead clear of the pipe.

The three tests, and what fools each one

Density. Saturated steam at atmospheric pressure occupies about 1,600 times the volume of the water it came from, so a float rides on condensate and sinks in steam, and an inverted bucket floats on steam and sinks in condensate. That is a physically enormous difference and it is why mechanical traps are the most decisive family. What fools it is anything that changes the buoyancy of the moving part rather than the fluid around it: a float that has been punctured and filled, a bucket that has lost its water seal, or a bucket vent hole plugged with scale so trapped air cannot escape and the bucket cannot sink.

Temperature. Condensate sitting at saturation is at exactly the same temperature as the steam above it, so a thermometer cannot tell them apart at all. A thermostatic trap gets round this by refusing to open until the condensate has cooled some way below saturation. That is the mechanism, not a defect, and it has a consequence you have to design around: a thermostatic trap necessarily holds condensate back until it subcools. What fools it is anything that shifts the element's temperature relative to the current saturation temperature - a change in system pressure, an unusually hot or cold space around the trap body, or somebody insulating a trap that was meant to lose heat to the room.

Flash velocity at a seat. A thermodynamic disc trap uses the fact that hot condensate flashes as it crosses the seat, so it accelerates, and the flash filling the small chamber above the disc holds the disc shut until that chamber's contents condense and the cycle repeats. It is a cycling device by design. What fools it is anything that changes the pressure relationship across the seat or the rate at which the chamber cools: too little differential pressure, too much back pressure, or wind and rain on an outdoor cap.

Say the three out loud on any trap complaint and half the diagnosis is done.

The case: twelve new traps and a bigger water bill

A low-pressure heating main, 5 psig, with twelve drip-leg traps that had been mixed and aged out over twenty years. The shop replaced all twelve in one shift with thermodynamic disc traps, on the reasoning that they are compact, have one moving part, tolerate freezing, and are sold as suitable for a wide pressure range. Six weeks later the complaint was a persistent rapid clicking along the main and a make-up water consumption that had risen to roughly 1.6 times its prior weekly baseline.

First hypothesis: a bad batch. Twelve traps from one carton could share a defect. But all twelve behaved identically, and identical behaviour across twelve units is far better evidence of a shared operating condition than of a shared manufacturing fault. Weak, and it stayed weak.

Second hypothesis: undersized. Easy to check by direction rather than by capacity table. An undersized trap passes less than the load makes, so the symptom would be condensate backing into the main and a cool far end. The observed symptoms were the opposite: more discharge, more noise, more make-up. Wrong direction, hypothesis dead.

Third hypothesis: back pressure. The return main measured 2 psig. Two psi of back pressure would not stop a mechanical trap on a 5 psig main, so a tech who stopped here concludes back pressure is not the problem and moves on. This is where the proxy question earns its keep.

What was true. A disc trap's test is flash velocity across a seat, so its ability to close depends on the ratio of back pressure to inlet pressure, in absolute terms, not on the gauge difference. Manufacturers publish a limit for this, commonly stated as back pressure not exceeding about 80 percent of the inlet absolute pressure, with some designs specified tighter, and the figure that governs belongs to that trap's own published data rather than to a rule of thumb, and the exact figure belongs to the specific trap's published data rather than to a rule of thumb. Run the numbers here: inlet 5 psig is 19.7 psia, return 2 psig is 16.7 psia, and 16.7 divided by 19.7 is 0.85. The back pressure was 85 percent of inlet absolute against a limit near 50 percent. The disc could not hold shut. It reopened almost as soon as it closed, and every reopening passed live steam.

Two of the twelve sat outdoors on an exposed section, and those two clicked noticeably faster than the other ten, which is the same mechanism from the other side: wind and rain cooling the cap condenses the chamber's charge sooner, so the disc reopens sooner.

How it was confirmed. An ultrasonic probe held against each trap body showed continuous flow rather than a discrete cycle with quiet between discharges. A contact probe on a cleaned spot on the return line downstream of the traps read close to the main's saturation temperature rather than well below it, which on a healthy drip trap it should not. And the make-up figure, which had already been established elsewhere as the cheapest single indicator of steam being thrown away, had moved by a factor a leak survey could not account for.

The fix. Ten of the twelve went back to float and thermostatic traps, whose test is density and therefore keeps working down to very low differentials, needing only enough pressure difference to push the water out, which the 3 psi here supplies comfortably. Every model still has a published minimum operating differential; take it from the manufacturer rather than assuming zero - it only needs enough pressure difference to push the water out, which 3 psi supplies. The other two, on the exposed run where freezing was the original argument for a disc trap, were also changed to float and thermostatic with a freeze-protected drip leg arrangement, because the freeze argument had been solving a problem the system did not have while creating one it now did.

The generalisation worth keeping

The disc traps were not defective and they were not the wrong quality. They were the wrong test for this load. That reframing changes what you ask on every trap job:

• On any low-differential application - low-pressure heating, a modulating coil, anything draining into a pressurised return - ask whether the trap's test survives at the lowest differential the system will present, not at its design differential.
• On any load with air in it - start-up, a coil that has been off overnight, a system with a poor vent - ask whether the trap's test can pass air at all. A density test cannot discriminate air from steam at all, because air is a gas and buoyancy cannot tell one gas from another. A float sinks and stays down; an inverted bucket floats and stays up. Both outcomes hold the valve shut, and that is what air-binding is. It is why mechanical traps are usually paired with a thermostatic air vent.
• On any load where the equipment must not flood - a heat exchanger, a process coil, anything with a temperature control on it - ask whether the trap's test deliberately holds condensate back. A temperature test does, by design.
• On any trap sitting somewhere unusual - outdoors, in a hot pit, inside an enclosure - ask whether the surroundings are part of the test. For a temperature test and a flash-velocity test they are. For a density test they are not.

What would change the reading

A higher system pressure would have rescued the disc traps. At 60 psig inlet, 74.7 psia, the same 2 psig return is 16.7 psia, about 22 percent of inlet absolute, comfortably inside the usual limit. Same trap, same return, opposite verdict, because the ratio is what the test depends on. That is the argument for disc traps on high-pressure drip service and against them on low-pressure heating.

A vented, atmospheric return would also have rescued them, for the same reason from the other direction.

Superheat changes which tests work at all. A density test needs condensate to arrive; a load running with superheat may present very little for long stretches, and a trap holding a water seal for its operation can lose it.

Verifying which test a trap is running, in the field

You will meet traps with no legible tag. Three observations, none of which require opening anything:

Listen to the pattern. A distinct discharge with silence between is a device that fills and dumps: a bucket or a float that has stopped modulating. A rapid regular click is a cycling device, which points to a disc. A continuous rush at a drip point that should be near-idle is a fault regardless of family.

Read the temperature relationship. Take an inlet and an outlet surface temperature on cleaned spots with a contact probe, hand and lead clear of the pipe. A trap whose inlet sits noticeably below the saturation temperature for the measured pressure while it is holding is running a temperature test and is subcooling on purpose. A trap whose inlet sits at saturation is running a density or flash test.

Before you call that difference real, note the character of the error. A single contact probe carries a systematic offset that is the same at both readings, so it largely cancels in the inlet-minus-outlet difference, leaving roughly that percentage of the difference rather than of either reading. Its random spread does not cancel, and two readings from one probe give a difference whose spread is about 1.4 times a single reading's. Use the same probe on both points for exactly this reason.

Look at what is beside it. A separate thermostatic air vent on the same drip leg is a strong tell that the trap itself cannot pass air, which narrows it to the density family immediately.

References

• Steam trap manufacturer published data for the back-pressure limit, minimum operating differential and capacity of any specific trap, which is the only authoritative source for those values
• ASME Steam Tables (IAPWS industrial formulation) for saturation pressure and temperature pairs used in the temperature checks above
• 29 CFR 1910.147 for isolation of pressure and thermal stored energy before removing a trap; 29 CFR 1910.133 for eye and face protection when operating a test valve
• See related: The Trap Families and What Each One Is Good At; How a Mechanical Trap Fails and What You See; Why Condensate Is the Expensive Half of a Steam System