Why a Manual Drain Left Cracked Open Is a Permanent Leak

Why this matters

Somewhere in most air rooms there is a drain valve a maintenance tech cracked open years ago so the tank would stay clear without anyone having to remember it. It hisses. Everybody knows it hisses. Nobody counts it, because it is not a fault, it is a decision, and a leak survey walks straight past anything with a reason attached to it. That valve is very often the single largest opening in the plant, and it is the one opening a leak report reliably omits.

Before you touch a drain

To operate a manual drain, stand to the side of the discharge path with it directed into a container or a floor drain, wear eye protection and hearing protection, and never look into the path or point it at a person: air driven into a break in the skin can enter the bloodstream. Air rooms commonly reach the 85 dBA eight-hour time-weighted average action level, at which point hearing protection and the rest of 29 CFR 1910.95 apply.

To replace a drain valve, close the upstream isolation, open the drain to vent the section to atmosphere, confirm zero on the receiver's own gauge with the drain still open, and lock and tag the isolation under 29 CFR 1910.147, which covers stored mechanical and pressure energy. If the work reaches the vessel's safety valve connection, 29 CFR 1910.169(b) forbids any valve between the receiver and its safety valve, so a shutoff added there to make the job easier is not permitted.

If a drain leg has frozen, isolate and confirm zero before doing anything about the ice, then thaw with heat trace listed for that piping or another controlled electric method used per its listing. Never put an open flame on a pressurised line, and never on a plastic line at any pressure, because heated plastic piping releases decomposition products and a flame on a frozen line raises pressure in a section whose vent may itself be blocked.

The rule, stated once

An opening venting to atmosphere passes air at a rate set by its throat area and the absolute pressure upstream of it. Intent is not in that relationship anywhere.

Above roughly 13 psig, any opening to atmosphere is choked: the air leaves the throat at the local speed of sound, and the mass flowing out becomes proportional to the absolute upstream pressure. That threshold comes from air's critical pressure ratio, which puts choking at an upstream absolute of about 1.9 times the downstream absolute, so every drain in a plant at normal working pressure is choked and this proportionality applies to all of them.

Two things follow, and the second is the one shops act on.

The flow does not depend on how far downstream the opening is, on what it is called, or on whether it is on a drawing. And because the flow tracks absolute upstream pressure, dropping a plant's setpoint from 105 psig to 95 psig takes the absolute from 119.7 psia to 109.7 psia, a ratio of 0.916, so every opening in the plant, deliberate and accidental alike, passes about 8 percent less air.

"Barely cracked" is a claim about area, and nobody can see an area

A valve a quarter turn off its seat looks like almost nothing. What matters is what fraction of its full bore the throat now presents, and that is not visible, not repeatable, and not what the person cracking it was estimating. They were estimating the sound.

For the comparison below, take a quarter-turn crack as presenting roughly a tenth of the valve's full bore area. That figure is illustrative and it is the weakest number in this article: the real effective area depends on the valve trim, and re-cracking the same valve after a service gives a different opening. Use it to see the shape of the result, not to quote a flow.

Two drains, one rule

Both drains below are the same quarter-inch manual ball valve on the same wet receiver at the same 105 psig. The rule above is applied to each without modification.

Drain A is cracked open and left. Effective throat is about a tenth of the full bore, and it is open 100 percent of the time the system is pressurised, which for most plants means every hour of every day including the ones nobody is in the building. Average effective open area, expressed as a fraction of the full bore: 0.10.

Drain B is the same valve on a timer, opening fully for 2 seconds every 5 minutes. Full bore, so ten times drain A's throat, but open for 2 seconds out of 300, a duty of 0.667 percent. Average effective open area as a fraction of full bore: 0.00667.

Divide: 0.10 / 0.00667 = 15.0. The drain everyone describes as barely open passes about fifteen times the air of the drain everyone describes as blowing wide open, because area multiplied by time is the quantity that matters and only one of the two drains has the time term working for it.

That result reverses the intuition the plant is running on, which is that a big opening is worse than a small one. It is not. A permanent small opening beats an intermittent large one whenever the duty fraction is small, and 2 seconds in 300 is a very small duty fraction.

What moves drain B. The open time is set by a person and it multiplies directly. Set the timer to 6 seconds every 5 minutes because a shorter cycle "did not look like it was clearing" and the duty goes to 2 percent, tripling drain B's loss and cutting the ratio from 15.0 to 5.0. Drain B is still the better of the two, and it is now spending three times what it needs to because nobody checked whether the shorter setting was actually failing to clear the condensate.

What removes the term entirely. A demand-actuated drain opens on the presence of liquid and closes when the liquid is gone, so the air loss approaches zero rather than being traded against a timer setting. Take the sizing and the settings from the manufacturer, because a demand drain that is undersized for the condensate load holds water in the vessel, which is the failure the timer was fitted to avoid.

Why the survey never finds drain A

An ultrasonic leak detector hears drain A immediately. It is one of the loudest things in the room. It gets walked past anyway, and the mechanism is social rather than acoustic: the surveyor asks what it is, gets told it is the tank drain and it is meant to be like that, and moves on. The leak report is then produced without it, and the plant's stated leak load understates the truth by whatever the deliberate openings are worth, which in a room with two or three cracked drains can be most of it.

So a leak survey needs a rule that overrides the explanation: an opening that passes air to atmosphere is counted, and the reason it is open goes in a separate column. Recording it does not commit anyone to closing it. It commits them to knowing the number.

Do the survey with every enclosure closed. If a leak is suspected inside an electrical enclosure, that is not a leak survey task; opening or working inside an energized enclosure is electrical work performed by a qualified person under 29 CFR 1910.333(b)(2), with 29 CFR 1926.417 as the construction counterpart, in the arc-rated protection the employer's electrical safety programme assigns under NFPA 70E-2021 in the edition adopted.

Somebody cracked it open for a real reason

This is the part that decides whether the fix holds. Drain A is not laziness, it is a rational answer to a maintenance system that failed.

The underlying duty is genuine: 29 CFR 1910.169(b) requires a receiver's drain to be operated frequently enough to prevent the accumulation of liquid, and a manual drain that depends on a person remembering it every shift will eventually not be operated. The consequence of not draining is not a wet floor, it is the largest term in the plant's water budget going downstream instead of out, arriving at the dryer as a load it was never sized for and at the tools as liquid.

So the person who cracked the valve was solving the right problem with the wrong device. Tell an owner to close it and nothing else, and within a month either it is open again or the receiver is full of water. The fix is a drain that does not depend on memory, which is a demand-actuated drain fitted, commissioned and then verified to be discharging.

The one condition that complicates it. A drain on an outdoor or unheated line can freeze, and a frozen demand drain stops draining without saying so, which is exactly the failure that gets a valve cracked open again the following winter. Heat trace and insulate that drain leg to its listing before removing the manual crack, or the fix will not survive the season it was installed in.

How to verify you got this right

Walk the plant with the system pressurised and no production running, and list every opening you can hear, with a column for whether it is deliberate. If the deliberate column is empty, you did not walk it honestly, because most plants have at least one.

Then confirm each replacement drain is actually discharging. A demand drain that has stopped is silent, and silence is what you were trying to achieve, so it looks like success. Watch it through several cycles under load and confirm it discharges condensate; on a cold day, confirm it still does.

Then re-take the plant's leak load with the deliberate openings closed and the drains working, and compare it against the pre-fix figure. The difference is the number the original survey should have carried, and it is usually the largest single line in the report that nobody wrote.

References

• 29 CFR 1910.169, air receivers, including the requirement at (b) that the drain be operated frequently enough to prevent liquid accumulation, and the prohibition on a valve between the receiver and its safety valve
• 29 CFR 1910.147, control of hazardous energy, for isolating and verifying depressurisation before replacing a drain valve
• 29 CFR 1910.95, occupational noise exposure, including the 85 dBA eight-hour time-weighted average action level, and 29 CFR 1910.333(b)(2) with 29 CFR 1926.417 as the construction counterpart for any work inside an energized enclosure, under NFPA 70E-2021 in the edition the employer has adopted
• Condensate drain manufacturer data for sizing, timer settings and demand-actuated commissioning
• See related: Where the Water Comes From and Where It Goes; How to Put a Number on What Leaks Are Costing; What a Receiver Is For and What Happens Without One