Why a Failed-Closed Trap Is Found First and Fixed First

Why this matters

A trap that fails closed loses no steam at all. It still outranks every leaking trap on your list on the morning it happens, because the process itself is the sensor: the equipment stops heating and somebody calls within the hour. What that call is really reporting is a vessel filling with water, and the damage clock that starts when it does has nothing to do with energy. It is a split coil, a hammer event that moves piping, or a main slowly filling with the slug that will do both.

Stop the water before you diagnose anything

Never admit steam to a coil, main or line you believe is holding condensate. A steam bubble meeting subcooled water collapses in milliseconds and the surrounding water accelerates into the void it leaves. That is condensation-induced hammer, and it has broken flanges and moved hangers.

So the first action on a suspected flooded coil is to close the steam supply valve, then drain the low point before anything else. Drain it through the drip or blowdown connection with the discharge piped to a floor drain or a secured hose directed away from every walkway, cracking the valve slowly from the side of the stem rather than standing in front of the outlet, in heat-resistant gloves and face protection under your PPE hazard assessment per 29 CFR 1910.132 and 1910.133. Saturated condensate flashes to steam the instant it reaches atmosphere, so the discharge is a steam jet, not a water stream.

If the trap has to come off, close the inlet and outlet isolation valves, lock and tag them under your written energy control procedure per 29 CFR 1910.147, relieve through the test valve, confirm zero at a gauge on the trap body rather than assuming it, and let the body cool before breaking the joint. Where the trap sits in a below-grade condensate pit or a tunnel, that space is a permit-required confined space until evaluated under 29 CFR 1910.146, and the steam and hot condensate in it are part of why.

What a flooded coil actually looks like

The mechanism is worth walking, because the symptom is easy to misread as a supply problem.

Condensate cannot leave, so the coil fills from the bottom up. Steam still enters at the top and still condenses on whatever tube surface remains dry, so the upper portion stays at saturation temperature while the lower portion cools toward entering-air temperature. Discharge air temperature falls, and the fall is progressive rather than sudden, because the flooded fraction grows over minutes to hours.

The trap body itself sits below the coil, full of condensate that has been standing and giving up heat, so it reads well below the saturation temperature of the supply. Cool trap, hot top of coil, cold bottom of coil, falling output.

Now name the other end of the range, because a symptom stated at one end is not established until you can say what the opposite case looks like. A trap on the same coil failed open gives you a coil that is uniformly hot top to bottom, discharge air at design, and a downstream pipe climbing toward inlet saturation temperature. Normal output, evidence only at the return. That is exactly why one gets called in and the other does not.

The three damages, and what sets each clock

Freeze split. A coil on outside air below freezing, holding stagnant water, with no steam flow to keep it warm, is a coil that is going to split, and the time available is a function of the water volume, the face velocity and how far below 32 F the entering air is. That is not a number to carry in your head, and the equipment manufacturer's freeze-protection requirements own it. The operating rule is what matters: a suspected flooded coil on sub-freezing entering air is drained immediately, not scheduled.

Condensation-induced hammer. Highest risk on re-admission of steam, which is why the safety section leads with it. A flooded steam main carries the other kind as well, a slug of standing water picked up and accelerated by steam velocity until it hits a fitting or a closed valve. Both are called water hammer and they are different events: one is a collapse driven by heat transfer, the other is momentum driven by velocity. The flooded-line case sets up both.

Standing corrosion. Condensate that sits rather than flows is condensate holding whatever dissolved gases came with it, against tube wall, at temperature, with no flow to carry it away. The chemistry of why returning condensate is aggressive belongs to the water treatment articles and to the article on how return lines fail from the inside; the point here is only that flooding converts a flowing problem into a standing one, and that its clock is simply the repair delay, because it runs for exactly as long as the water sits.

Why these almost never appear on a survey list

Read this next to any trap survey you have run. Failed-closed traps are underrepresented on survey lists, and the reason is not that the mode is rare.

The process removes them. A failed-closed trap is reported by an operator and repaired within hours, so it has already left the population by the time the surveyor walks the plant. A failed-open trap produces no complaint and stays in the population until someone tests it. The survey therefore samples a population that has been filtered, and the open-heavy split it reports is partly that filter.

This matters when you use survey data to plan. The failed-open share tells you what your survey interval is costing you in energy, which is the subject of its own article. It does not tell you the true relative frequency of the two failure directions, and it should not be quoted as though it does.

Worked example: a makeup air unit on a 10 F morning

The call is "the makeup air unit is blowing cold," logged at 07:10. Outside air is 10 F. The unit has a 15 psig steam preheat coil with a float and thermostatic trap, and design is a 60 F rise.

First action, before any measurement. Close the steam supply valve to the coil, then drain the low point through the drip connection into a piped drain or a secured hose aimed away from every walkway, cracking that valve slowly from the side of the stem in heat-resistant gloves and face protection and never standing in front of the outlet, because what comes out flashes to steam on contact with atmosphere. Do this because entering air is below freezing and the freeze clock is running while you diagnose. This is not a step you take after confirming the fault. It is what makes it safe to look for the fault.

Then the readings. Discharge air is 34 F against entering air at 10 F, so the achieved rise is 24 F against a 60 F design rise, which is 40 percent of design.

A scan of the coil face shows the top third hot and the bottom two thirds falling toward entering-air temperature. Use that pattern, not the absolute face temperatures, because a finned face reads the fin, and fin temperature sits between tube temperature and air temperature rather than at either. The pattern is the evidence; the numbers on the fins are not.

Now check whether the numbers hang together. Two thirds of the tube surface flooded, and 40 percent of design rise delivered rather than the 33 percent a straight area proportion would suggest. That gap is real and it has a reason: the dry upper section is working against air that is no longer being preheated by the lower rows, so it runs a larger temperature difference than it does at design and delivers more per unit of surface than its share. Area proportionality is a first approximation for a coil operating at design conditions, and this coil is not at design conditions, so do not use the 40 percent to back-calculate a flooded fraction. Use it to confirm the story is consistent, which it is.

The trap. Body reads well below the roughly 250 F saturation temperature for 15 psig, and cool to a contact probe. Ultrasonic gives nothing at all: no rhythm, no continuous flow, silence. On a float and thermostatic trap, which discharges continuously under load, silence with the coil live is a failed-closed reading.

Second finding, do not skip it. This unit has a low-temperature limit that should have closed the outside air damper and stopped the fan long before the discharge fell to 34 F on 10 F entering air. It did not trip. Before treating the trap as the whole fault, establish why the limit did not act. A limit that failed to protect is a separate defect, and replacing the trap leaves it in place for the next event. A limit found jumpered or bypassed is a finding in its own right and gets reported as one.

Timeline. Call at 07:10, on site 08:40, supply closed and coil drained by 09:05, trap replaced and coil back in service at 10:30. Total exposure from call to repair is 3 hours 20 minutes, or about 3.3 hours.

Set that against a trap on the same system that fails open instead. Nothing changes for anyone, and the clock runs to the next survey. The exposure arithmetic for that case, and what shortening the survey interval does to it, is worked in the article on why failed-open traps cost more; it is not re-derived here.

How to verify you caught the whole fault

• Confirm the coil actually drained before you re-admit steam. A drain that stops flowing may be plugged rather than empty. Verify at the drain, and open the steam valve slowly on the first re-admission with nobody standing at a flange or a fitting.
• Confirm the trap failed closed rather than being starved. A trap that never received steam because a supply valve is shut or a strainer ahead of it is plugged reads cold and silent exactly like a failed-closed trap. Check the strainer and the supply valve position before you condemn the body.
• Confirm the protective device that should have acted. Where a freeze limit, a low-temperature cutout or a high-level alarm exists and did not operate, that is a second work order, not a footnote.
• Check the coil for a split before returning it to service, particularly on any coil that stood flooded on sub-freezing entering air. A hairline split leaks into the airstream and shows up later as a wet filter bank rather than as a puddle.
• Log the failure direction on the trap tag. A population where the same station repeatedly fails closed is telling you about debris carryover or a strainer that is not being blown down, and that pattern is only visible if the direction was recorded.

References

• 29 CFR 1910.147, control of hazardous energy, for isolation and verification before removing a trap or breaking a joint on steam or condensate piping
• 29 CFR 1910.132 and 29 CFR 1910.133, PPE hazard assessment and eye and face protection, covering hot surface and flashing condensate exposure during a drain
• 29 CFR 1910.146, permit-required confined spaces, where the trap station is in a below-grade pit, vault or tunnel
• Equipment manufacturer documentation for coil freeze-protection requirements and low-temperature limit setpoints
• See related: Why a Failed-Open Trap Costs More Than a Failed-Closed One; Why Condensate Return Lines Fail From the Inside