Why an Oversized Trap Fails Differently Than an Undersized One

Why this matters

Both sizing errors get called "the trap is bad" and both get answered with the same part. They are not the same problem, they do not cost the same thing, and they announce themselves with opposite symptoms, which means the direction of the error is diagnosable from the outside before anything is opened.

Get the direction wrong and you replace a trap with another one of the same size, and the customer buys the identical failure on the same interval for as long as they keep calling you.

A trap body and every line touching it sit at supply temperature, and the condensate inside flashes to steam the instant it is released. Nothing in this article requires opening a live trap. Where a probe has to touch the body, it is an ultrasonic instrument with an insulated handle, held with a glove rated for the contact temperature under your shop's hazard assessment required by 29 CFR 1910.132. Where a trap has to come out, the section is isolated at both ends, drained to a safe discharge point, vented, confirmed at zero on the upstream gauge, and locked and tagged under the written energy control procedure required by 29 CFR 1910.147.

The rule, in two halves

A steam trap is a valve whose only job is to tell condensate apart from steam. Everything it must do reduces to one rule with two halves, and both halves have to hold across the whole load range, not just at design:

It has to pass condensate as fast as the equipment makes it, at the differential pressure that exists at that moment. And it has to spend most of its life sitting closed against steam rather than cycling.

Undersizing fails the first half. Oversizing fails the second. Every symptom below is one of those two sentences showing itself in the field.

Undersized: what backs up, and where it shows first

When capacity falls short of the condensate rate, the water has to go somewhere and the only direction available is backwards. The drop leg fills first, then the equipment's lowest passes. Flooded surface is surface removed from service, so output falls.

What happens next is the part that hides the fault. On any equipment with a modulating control, the controller responds to falling output by opening the valve. That raises pressure inside the equipment, which raises the trap's differential, which raises the trap's capacity. The system partially fixes itself by running hotter than it was designed to.

So the field fingerprint of an undersized trap at moderate to high load is not a cold coil. It is normal output at a steam pressure meaningfully above the design number. A coil selected for 15 psig that holds its leaving air temperature while its own gauge reads 25 psig has a drainage problem, not a control problem, and rebuilding the control valve will not touch it.

That trick has a limit and the limit is low load. At low load the controller is closing the valve, not opening it, so equipment pressure is falling and the differential is falling with it. The compensation runs backwards exactly when the trap needs help. An undersized trap shows first, and worst, at low load, which is the condition nobody commissions at and the condition a building spends most of the year in.

Two consequences ride along. Standing condensate in a horizontal run with steam moving over it is what condensate-induced water hammer is made of, so a system that has begun banging on startup is telling you water is being stored where it should not be, and the response is to slow the warm-up and keep people away from the line and its fittings until it is quiet. And on a coil handling outdoor air below freezing, backed-up condensate is a burst tube; a sibling article owns that mechanism in full.

Oversized: what wears, and why the waste comes later

An oversized trap works. That is the difficulty with it.

At a load far below its capacity, the trap opens, discharges the small amount of condensate waiting for it almost instantly, sees steam at the seat, and shuts. Then it does that again. The cycle rate goes up as the mismatch goes up, and every cycle drags the seat across the seat under full differential.

Two things follow. The seat wire-draws, which is a slow widening of a leak path once one exists. And on trap designs that carry a working prime, notably the inverted bucket, rapid cycling with very little liquid arriving is how the prime is lost, after which the bucket sits down and the trap passes steam continuously.

Note the direction carefully, because this is where the subject is usually stated backwards. An oversized trap is not wasting steam the day it is installed. It is discriminating correctly and it is wearing out early. The waste arrives when it finally fails open, at a fraction of the service life the customer was expecting, and it arrives invisibly, because a trap blowing live steam into a closed return system produces no puddle, no noise anyone notices, and no complaint. That is the whole reason a trap survey exists as a service: the failed-open population is the part of the loss nobody is calling about.

Flash is not a leak, and here is why it looks like one

Before any of this can be diagnosed, one thing has to be understood cold, because it is responsible for more misdiagnosed traps than any other single fact.

Condensate at pressure is liquid water sitting at its saturation temperature. Release it to a lower pressure and it is now above the saturation temperature for that new pressure, so part of it boils instantly. That is flash steam. The fraction is the difference in liquid enthalpy divided by the latent heat at the lower pressure.

Condensate at 100 psig carries about 309 Btu per pound of liquid enthalpy. At 0 psig, liquid enthalpy is about 180 and latent heat is about 970, all from the steam tables. So the flash fraction is 129 divided by 970, about 13 percent by mass.

Thirteen percent by mass is not a dramatic number. By volume it is, and volume is what a human eye is reporting. Flash steam at 0 psig occupies about 26.8 cubic feet per pound while the liquid it came from occupied about 0.018. Run that out and the discharge leaving that trap occupies roughly 200 times the volume of the condensate that entered it.

A working trap on a high pressure drip leg produces a large white plume on every discharge, and that plume is correct. Anyone who condemns a trap for it is condemning a trap for doing its job, and shops do it constantly.

Telling flash from blow-through without touching anything

Three separations, and the last one is the reliable one.

Timing. Flash is synchronized with the trap's discharge and stops in between. Live steam blow-through is continuous. This is the fastest read and it fails on a trap whose load is high enough to be discharging nearly continuously anyway, which is most modulated equipment at design load.

Appearance. Flash is a lazy white cloud that expands and dissipates. Live steam leaves a sharp, invisible core for the first stretch and only turns white further out. Useful, and not something to lean on, because both are affected by ambient humidity and both look identical from twenty feet.

Instruments. An ultrasonic listening instrument on the trap body gives the honest answer: a working trap has quiet intervals in its signature, a blown trap gives a continuous high-frequency flow signal. A temperature check adds a second read, and it carries an error character that has to be stated with it. A surface temperature reading, whether infrared on a prepared target patch or a contact probe, is biased low by the film, the wall, and the surface finish, and that bias runs in one direction only. So an outlet reading near supply temperature is strong evidence of live steam passing, because the bias could only have pushed it lower. A low outlet reading is consistent with a working trap and also with a trap failed closed, so it does not distinguish those two and must not be reported as if it did.

Case A: the drip leg that got the pipe size

Drip leg at the base of a riser on a 100 psig main, about 200 feet of insulated 4 in pipe upstream of it. Return system at 15 psig, so the differential across this trap is 85 psid, near the top of what the trap will ever see.

The running load on a drip leg is pipe heat loss, and the insulation manufacturer's heat loss table for that pipe size, insulation thickness, and ambient is the authority for it. On this run it came out near 30 pounds per hour.

The leg is 3/4 in, so a 3/4 in trap was fitted. Read that trap's own capacity curve at 85 psid and it comes back at roughly twenty times the 30 pounds per hour the leg makes.

Against the rule: the first half passes with enormous margin. The second half fails badly. At a twentieth of capacity, at the highest differential the trap ever sees, this trap is cycling constantly and doing it under maximum load on the seat.

What the customer experiences is a drip trap that needs replacing every year or two on a main that runs continuously, and, in between replacements, a period nobody notices where the worn trap is passing live steam into a 15 psig return.

The fix is not a smaller connection. It is a trap selected from the curve at 30 pounds per hour and 85 psid, with the drip leg staying full size so condensate still falls into it freely. The sizing article in this group carries that procedure.

Case B: the coil trap that was right on paper

Coil making about 847 pounds per hour of condensate at 15 psig. Trap was selected at 850 pounds per hour at 14 psid, which matches the load and looks correct on the submittal.

Then read the return main gauge: 8 psig, because three other pieces of equipment discharge into it. Add about 1 psi of return line friction. The actual differential at design is 15 minus 8 minus 1, so 6 psid, not the 14 the selection was made at.

Using the square root relationship as a check, which is derived for incompressible orifice flow and which the trap manufacturer's curve supersedes because flashing condensate chokes across the seat: capacity at 6 psid is about the square root of 6 over 14, roughly 0.65, so about 557 pounds per hour. Required is 847. The trap passes about two thirds of the load at design.

So the coil floods a little, output falls a little, and the control valve opens. Coil pressure climbs to about 25 psig, which makes the differential 16 psid and the capacity about 909 pounds per hour, and the coil drains again. The building never complains, and from the steam tables the coil is now running at about 267 F instead of the 250 F it was selected for, roughly 17 F hotter than design, for its whole life.

Now take the same coil to 40 percent load. Required condensate is about 339 pounds per hour. The coil's own capacity curve puts its pressure near 4 psig at that load, so the differential is 4 minus 8 minus 1, which is negative 5. The trap passes nothing at all, at any size.

That is the finding worth carrying: at design this trap was undersized and hidden by the control, and at part load it stopped being a sizing problem entirely. No larger trap fixes the second condition, because there is no differential to work with. The sizing article's gate exists precisely to catch this before a capacity is ever looked up, and the coil article covers what the flooded coil does next when the air crossing it is below freezing.

References

• Steam trap manufacturer capacity curves, which are the governing authority for capacity at a given differential and supersede the square root check used above
• Insulation manufacturer heat loss tables, which are the authority for the running condensate load on a drip leg
• Saturated steam tables from any engineering handbook, for the liquid enthalpy, latent heat, and specific volume figures in the flash calculation
• 29 CFR 1910.132 for the hazard assessment and protective equipment used when contacting a trap body, and 29 CFR 1910.147 for isolating stored pressure energy before a trap is removed
• See related: How to Size a Trap From the Load Rather Than the Pipe; What a Steam Coil Does and Why It Freezes