What Pipe Pitch and Drainage Do in a Steam Main

Why this matters

Pitch is the cheapest thing in a steam system and the only one that cannot be added later without cutting the pipe down. A main that drains carries dry steam to the equipment and stays quiet; a main that does not carries water to the equipment, hammers, and delivers heat at a rate nobody can explain from the gauge. The mistake is thinking of pitch as a quantity, as in whether a run has enough of it. Pitch is a direction first. Until you know which way the condensate is meant to travel, no number is right, and the same fall that drains one main perfectly will flood the identical main next to it.

The one decision: which way does the water go

Steam moves toward the load. Condensate forms all along the way and has to get to a drip point. There are only two arrangements.

Parallel flow pitches the main down in the direction the steam is travelling, so gravity and steam drag both push the condensate the same way, toward drip points ahead of it.

Counterflow pitches the main back against the steam, so condensate runs downhill toward the source end while steam travels uphill past it. This is used on risers, on one-pipe heating systems, and any time a main has to be dripped back toward where it came from.

Everything else in this article follows from that choice, because the two arrangements have opposite sensitivities. In parallel flow, steam drag helps drainage, so a gentle fall works and higher velocity is tolerable up to the point where the drag tears the film into waves. In counterflow, steam drag fights drainage, so the fall must be steeper and the velocity must be held down, because the same drag that helps in one direction holds the film in place or drives it back in the other.

The conventional figures reflect that. Parallel-flow mains are commonly run at about 1 in of fall in 20 ft, roughly 0.4 percent. Counterflow runs are commonly pitched at about twice that, on the order of 1 in in 10 ft, with drip points spaced closer and the design velocity kept well below what a parallel-flow main of the same size would carry. Those are common practice, not law: the governing figures for a given job come from ASME B31.1 Power Piping in the edition your jurisdiction or your contract has adopted, which binds the installation, and from the design guidance the system was engineered to. What travels is the relationship, which is that counterflow costs you both pitch and velocity, and it costs both for the same reason.

Drip spacing along a straight run

On a straight run with nothing interrupting it, drip points on a parallel-flow main are commonly spaced in the range of 150 to 200 ft, with the closer end of that range used where the main is uninsulated, where it passes through unheated space, or where the run is counterflow. Between drips the condensate depth builds continuously, and longer spacing means a deeper film at the far end of each interval, which is the condition the sibling article on what water hammer actually is in a steam line identifies as the start of slug propulsion.

The spacing number is the least portable thing in this article. What is portable: if you find yourself reasoning that a run can go further between drips because the pipe is insulated, check whether it also runs through a space that is unheated at night, because insulation reduces the running load and does nothing about the load produced when the surrounding air drops twenty degrees.

The points that get a drip regardless of the interval

Interval spacing is the floor. These points get a drip whether or not the interval calls for one, because at each of them the condensate has stopped being able to continue:

• Ahead of every riser. Steam goes up; water does not.
• Ahead of every expansion loop, since the loop's own geometry contains a low point.
• Ahead of every control valve and isolation valve. A closed valve is a dam, and a control valve that throttles produces its own condensate on the downstream side.
• At the end of every main. The end of a main is where everything that was not caught arrives.
• At every reduction in pipe size, and at the low point of any pocket the routing could not avoid.
• Wherever a main leaves a heated space for an unheated one, because the condensing rate steps up at that boundary.

Every one of those drips is a full-size leg with a dirt pocket, not a tapped connection; the sibling article on why a drip leg exists covers the capture geometry and the depth.

Three things that build a dam

Pitch on the drawing is undone at the details, and these three do it quietly.

A concentric reducer in a horizontal main. A concentric reducer keeps the centrelines aligned, so when a main steps down in size the bottom of the bore steps up, and the step is a dam that holds a pool on its upstream side for the life of the system. The correct fitting on a horizontal steam main is an eccentric reducer installed flat side down, so the bottom of the bore stays continuous and only the crown steps.

A branch taken off the side or the bottom of a main. Any branch that leaves below the crown collects the condensate running along the main and delivers it to the equipment, which is how a perfectly good coil ends up wet and hammering while the main it hangs from is fine. Branches come off the top of the main and turn down after they have left it.

A sag between hangers. This is the one that is not a fitting at all, and it is the most common. A pipe deflects under its own weight plus the insulation plus the water it carries, and every sag contains an uphill half. That means a main which is parallel-flow on the drawing contains a short counterflow section at every sag, at whatever the design velocity happens to be, which is the one combination the counterflow rules say to avoid. Hanger spacing comes from the adopted piping code and from the support standard the job specifies, such as MSS SP-58 where the specification calls for it and which binds through that specification; do not extrapolate a spacing from what looks fine on a cold water line, because water lines do not have to keep a continuous fall.

The case: sixty feet of level pipe

A shop extended a 4 in main 60 ft to feed a new piece of equipment. The ceiling was tight, the run looked short, and it went in dead level with the branch tapped off the side of the main near the end.

Within a month: intermittent hammer at the new equipment on start-up, output at the new coil below what the pressure said it should deliver, and every existing trap on the original main testing fine.

Why the existing traps were a dead end. They were fine. The extension had no drip point at all, because at 60 ft it was inside the interval spacing and nobody applied the rule that the end of a main gets a drip regardless of interval.

What level pipe actually did. A dead-level run has no defined drainage direction, so the condensate goes wherever the local deflection sends it. With hangers at the spacing used for the rest of the run, the extension held a pool in the middle and another at the far end, and the branch took off from the side into the second pool. The coil was receiving water and steam together, which is why its output was low: the coil's rating assumes dry steam, and water arriving in the tubes both occupies surface and shows up as the wet fraction that a pressure gauge cannot see.

The arithmetic that had been avoided. At the parallel-flow figure of 1 in in 20 ft, 60 ft needs 3.0 in of fall. The installer had about 2 in of ceiling clearance at the far end and concluded it could not be done. That conclusion contains the error: fall is a difference, not an absolute. The extension was re-hung 3 in higher at the tee end, which cost nothing overhead because the tee was in a deeper part of the ceiling, and the far end stayed exactly where it had been. Had the run been counterflow instead, at roughly 1 in in 10 ft, the same 60 ft would have needed 6 in, and 6 in was genuinely not available at either end, which would have forced a drip point mid-run and a shorter counterflow section rather than a re-hang.

The rest of the repair. A full-size drip leg with a dirt pocket at the new low point, the branch moved to the top of the main, and the eccentric reducer at the size change turned flat side down after it was found installed as a concentric. Three of those four are geometry that costs nothing at installation and a shutdown to correct afterwards.

Failure mode if it had been left. Not a burst. The coil would have continued underperforming, someone would eventually have raised the system pressure to compensate, and raising pressure raises the saturation temperature and the condensing rate together, which makes more condensate for the same run to carry. That is the loop worth recognising: a drainage fault presents as a capacity complaint, and the intuitive response to a capacity complaint makes the drainage fault worse.

Measuring pitch so the number means something

A 0.4 percent grade over 4 ft is 0.19 in, about three sixteenths, which is at the edge of what a spirit level resolves and well inside the error of a level that has been dropped.

Cancel the instrument's own offset. A spirit level's calibration error is a fixed systematic offset, and a fixed offset reverses sign when you turn the tool end for end. Read the pitch, rotate the level 180 degrees on the same spot, read again, and average the two; the offset cancels in that average. That is a different operation from averaging to beat noise: an independent random reading spread does not cancel, it only shrinks by the square root of two when you average two readings, so the two errors need different treatment even though the field action looks identical.

Do not step a short level down a long run. Fifteen placements of a 4 ft level to cover 60 ft, each carrying an independent spread, accumulate when you add them, and the total spread grows as the square root of the number of placements, about 3.9 times the single-placement spread. Measure the total fall end to end with a laser or a water level instead, then divide, and the accumulation problem disappears because there is only one measurement.

Measure on the pipe, not the insulation. Insulation thickness varies enough along a run to invent a grade that is not there. Read at bare sections, at fittings and at existing gaps in the lagging rather than opening it, because thermal system insulation installed before 1981 is presumed asbestos-containing under 29 CFR 1926.1101 for construction work and is regulated under 29 CFR 1910.1001 in general industry, which makes it an inhalation exposure to be sampled or removed by qualified people rather than pulled open for a level reading.

Do the measuring cold where you can. A main at 250 F burns on contact, and a main that has grown thermally is not sitting where it sits cold. Where a live main has to be reached from a ladder, the ladder rules that apply are 29 CFR 1910.23 in general industry and 29 CFR 1926.1053 on a construction site, and a field-service shop can be under either depending on the job. If a run has to be re-pitched, that is not live work: isolate the main, lock out the isolation under 29 CFR 1910.147, confirm zero pressure on a gauge known to be good, drain it, and let the metal fall below 120 F verified by a non-contact reading taken at arm's length before a hanger is loosened, because a hot main released at one support moves.

References

• ASME B31.1 Power Piping, in the edition your jurisdiction or contract has adopted, which binds the installation, for pitch, drip point and support requirements
• MSS SP-58, where the project specification adopts it, for pipe hanger and support spacing
• 29 CFR 1910.147 for isolation before re-pitching, 29 CFR 1910.23 and 29 CFR 1926.1053 for ladder use in general industry and construction respectively, and 29 CFR 1926.1101 with 29 CFR 1910.1001 for asbestos in thermal system insulation
• See related: articles in this library on why a drip leg exists and what happens without one, and on what water hammer actually is in a steam line