Why a Drip Leg Exists and What Happens Without One

Why this matters

A drip leg looks like a stub of pipe hanging off a main, and on a lot of jobs it gets value-engineered into a threaded tap and a trap because the trap is the part that costs something. The trap is not the part doing the capture. A trap can only discharge what reaches it, and condensate travelling along the floor of a steam main will pass a small connection almost entirely, so the same trap on the same main either drains it or does nothing depending on what it is hung from. Every failure downstream of that decision looks like a trap problem, gets a trap replacement, and comes back.

This article runs one gate against two ways of connecting the same trap to the same main.

The gate

Does the opening give the condensate film enough time to leave the pipe wall before the steam carries it past?

That is the whole question, and it is a timing question rather than a gravity question. Condensate does not fall into a drip connection; it is dragged along the bottom of the main by the steam moving above it, as a film that thickens toward every low point. To be captured, that film has to detach from the wall while it is over the opening.

The time available is the length of the opening in the direction of flow divided by the film velocity. The distance the film can fall in that time is the ordinary free-fall figure, one half of 32.2 ft per second squared multiplied by the time squared. Because the drop goes with the square of the time and the time goes with the length of the opening, the capture distance scales with the square of the opening length. Doubling a connection size does not double the capture, it roughly quadruples it.

Two conditions attach to that calculation and both matter. It is a free-fall estimate that ignores surface tension holding the film to the wall and ignores the steam drag that keeps acting on the film the whole time, so it understates the difficulty rather than overstating it. And it uses a film velocity, not the steam velocity, which is considerably higher; the film is slower than the steam above it, which is the only reason capture is possible at all.

Case one: the full-size leg

   steam main, pitched down in the flow direction
   ==============================================>

                 full-size tee, not a reducer
                        |
                    +---+---+
                    |       |  drip leg, same size as
                    |       |  the main on smaller
                    |       |  mains
                    |       +---> trap take-off from
                    |       |     the side, above the
                    |       |     bottom
                    +---+---+
                        |
                  dirt pocket and blowdown
                  valve below the take-off

Take a 6 in main with a full-size leg dropping out of a full-size tee. The opening in the direction of flow is the full 6 in bore, so at a film velocity of 10 ft per second the film is over the opening for half a foot divided by 10, which is 0.05 seconds, and it falls one half times 32.2 times 0.05 squared, or about 0.040 ft, which is roughly 0.48 in.

A condensate film in a loaded main is commonly a few hundredths of an inch thick, thickening toward a low point. Half an inch of fall detaches the entire film and then some, and just as importantly the floor of the pipe has been removed for that half foot, so there is no surface for the film to run on and the sudden expansion into the leg collapses its forward momentum. The gate is satisfied.

Case two: the reducing tee off the bottom

Same main, same steam, same trap, connected through a 1 in tapped connection in the bottom of the pipe.

The opening is 1 in long, so the film is over it for one twelfth of a foot divided by 10, or 0.0083 seconds, and falls one half times 32.2 times 0.0083 squared, which is about 0.0011 ft, or 0.013 in.

Thirteen thousandths of an inch. Against a film a few hundredths of an inch thick, on a wall it is stuck to by surface tension, with steam drag still pushing it forward. Most of the film bridges straight over the hole and continues down the main to the next elbow, the next riser, or the equipment.

The two cases resolve opposite ways on the same gate, and the ratio is exactly the square of the opening ratio: six times the length gives thirty-six times the fall. Nothing about the trap entered the comparison. Put a trap ten times the capacity on case two and it still drains a fraction of the condensate, because capacity is a discharge rating and the problem is upstream of the trap inlet.

This is why the sizing convention exists in the shape it does: full-size legs on smaller mains, and on large mains a leg not smaller than about half the main diameter. Take the specific dimension from the trap manufacturer's drip-leg table and from the piping code the job is built under, which is ASME B31.1 Power Piping in the edition your jurisdiction or your contract has adopted, which binds the installation. Those two own the number; the reasoning above only tells you why a number exists.

The second job: holding the surge

Capture is the leg's first job and buffering is its second, and the second is the one that decides the depth rather than the diameter.

Condensate does not arrive at a steady rate. A cold main produces condensate at many times its running rate while the metal comes up to temperature, and the sibling article on why a cold start is the most dangerous moment derives that load from the pipe mass. The leg is what stands between that surge and a flooded main while the trap works through it.

Take that article's case, which is the same 6 in main: 200 ft of Schedule 40, warmed from a cold building to 100 psig, produces on the order of 144 lb of condensate as warm-up load. A 6 in Schedule 40 leg has a bore of about 6.065 in, so a flow area of about 0.20 square feet, and a leg 28 in deep holds about 0.47 cubic feet, which at roughly 59 lb per cubic foot is about 28 lb of water.

That is about a fifth of the total warm-up load held at any one instant. The leg is not storage for the whole event and was never meant to be. It is a buffer sized so that the arrival rate and the discharge rate never have to match instant by instant, and so that when they do diverge the water backs up into a stub of pipe rather than into the main where it becomes a slug.

Why the depth is a pressure, not a preference

Here is the part that gets missed, and it explains the deep legs on automatic warm-up systems that look like overkill.

A trap is differential-pressure driven. At the moment a cold main is admitted steam, the pressure in it is near atmospheric, so the differential across the drip trap is near zero and the trap can barely pass anything, which is precisely when the condensate load is at its highest. The static head of water standing in the leg is the differential available.

Water column converts at about 27.7 in per psi. A 28 in leg therefore supplies about 1 psi of differential across the trap at the moment the main has none of its own. That is the entire reason a leg on a supervised warm-up can be shallower than a leg on a system that starts itself unattended: the unattended one has to drain during the period when nobody is standing there cracking a manual drain, so the leg has to buy the differential the main is not yet providing.

Read a trap's capacity chart at the differential you will actually have at start-up, not at operating differential. A trap rated to pass several hundred pounds per hour at 100 psi differential passes a small fraction of that at 1 psi, and the drip trap on a main is sized on the warm-up load at the start-up differential. Sizing it on the running load at operating differential is the most common way a correctly-specified trap turns out to be undersized on the only morning it matters.

The third job: keeping dirt out of the trap

The take-off to the trap comes off the side of the leg above the bottom, leaving a pocket below it, with a blowdown valve at the base. Scale, weld slag and pipe rust settle into that pocket instead of into the trap seat, and a trap that fails open because a piece of scale is holding its valve off the seat wastes steam continuously and silently.

Blowing that pocket down is a routine task with a real hazard. The discharge is condensate at saturation temperature plus the flash that forms as it leaves, so route it to a point where nobody stands or walks, open the valve slowly while standing to the side of the outlet and never in front of it, and wear hearing protection where the discharge is above the action level in 29 CFR 1910.95. Do not blow down a leg whose discharge is an open pipe pointing across a walkway, at any pressure, until the discharge has been redirected.

Verifying a leg you did not build

Four checks, all of which can be made without opening anything on a live line.

Look at the tee, not the trap. A reducer at the main is the finding. You are looking for whether the branch leaves the main at the main's own diameter, or at least at the fraction your code and the trap manufacturer allow for that size.

Measure the depth with a tape from the outside. Take-off centreline down to the base. Compare it against the differential available at start-up, using the 27.7 in per psi conversion above, and against whether this main warms up under supervision or on a timeclock.

Confirm the take-off is above the base. A trap connected to the very bottom of the leg has no dirt pocket, and that shows up later as a trap that fails open repeatedly for no apparent reason.

Check for a strainer and whether it has ever been cleaned. A blocked strainer in front of a drip trap produces exactly the symptom of a failed-closed trap, and it is the cheaper thing to eliminate first.

If a leg has to be modified, none of that is live work. Isolate the main, lock out the isolation under 29 CFR 1910.147, relieve the pressure and confirm zero on a gauge that is known good, drain the section, and let the metal cool below 120 F verified with a non-contact reading from arm's length before any cutting starts. Where insulation has to come off, thermal system insulation installed before 1981 is presumed asbestos-containing under 29 CFR 1926.1101 for construction work and regulated under 29 CFR 1910.1001 in general industry, so it is sampled or removed by qualified people rather than pulled by the tech, and where a respirator is required it is under a 29 CFR 1910.134 program because this is an inhalation route that gloves do not touch. Cutting or welding galvanized pipe releases zinc oxide fume, which needs local exhaust or supplied-air rather than a face shield, and the hot work itself runs under the site's permit process, which is NFPA 51B in the edition your authority having jurisdiction or your insurer has adopted and which binds through that adoption.

References

• ASME B31.1 Power Piping, in the edition your jurisdiction or contract has adopted, which binds the installation, for drip point and drainage requirements on steam mains
• Trap manufacturer drip-leg sizing tables and capacity charts read at the differential available at start-up
• 29 CFR 1910.147 for isolation of the main, 29 CFR 1910.95 for hearing protection at open discharges, and 29 CFR 1926.1101 with 29 CFR 1910.1001 for asbestos in thermal system insulation
• NFPA 51B, in the edition your authority having jurisdiction or insurer has adopted, which binds the hot work permit on that site
• See related: articles in this library on why a cold start is the most dangerous moment, and on what pipe pitch and drainage do in a steam main