Why Moisture Is the Central Problem in Compressed Air

Why this matters

Every other contaminant in a compressed air system enters at a rate the machine sets. Wear metal comes from the airend, oil carryover comes from the separator, scale comes from the pipe. Water is the one whose quantity is set by the weather outside the building, which is why a system that ran clean in February starts producing defects in July with nothing changed and nothing broken. It is also the only contaminant that changes phase inside the pipe, so removing it once does not remove it, and that single property defeats the whole mental model most shops apply to air quality.

The call

A collision shop was redoing roughly one panel in ten. The defect was small craters in the clear coat, scattered rather than patterned, appearing on some jobs and not others through the same day.

The obvious suspects were worked in order and all of them came back clean. Two different guns, one of them new out of the box, produced the same rate. The gun-mounted filters were changed to a fresh box and the rate did not move. Two painters with different technique produced the same rate, which is the reading that took technique off the table, because a technique defect follows the person.

Oil was the strongest candidate, since an oil-injected screw compressor was feeding the booth. It was eliminated on evidence rather than on the filter's service record: a clean white card held at arm's length in the drop's airstream, with skin kept out of the jet because compressed air driven into a break in the skin can enter the bloodstream, showed no film after a full minute in three separate tests, on days with and without defects.

What did correlate was the calendar. Nine of eleven defect days in the log were humid afternoons, and the two halves of that fit together rather than fighting: liquid that condenses on the cool morning run settles in the low points and gets swept forward as slugs once flow rises, while the humid afternoon adds a moisture load the dryer's fixed outlet dew point cannot follow. Both terms peak later in the day than the condensing does. The point-of-use desiccant cartridge at the booth was being changed weekly through July and had been lasting about a month through the winter.

What was going on

The dryer was fine. Measured at its outlet, at line pressure, under load, it was delivering air at its rated pressure dew point. The shop's whole air room was working as sold.

Two things were true at once and neither was a fault in a component.

The first is that the moisture load arriving at the intake in July was several times what it had been in January, because the intake air's capacity to hold water rises steeply with temperature. The point-of-use cartridge had been sized against a winter observation, so it was correct then and swamped later. Nothing degraded; the input changed.

The second is that the drop feeding the booth ran twenty feet along an uninsulated exterior wall. On cool mornings that pressurised metal fell below the dryer's outlet dew point, so liquid formed downstream of every device in the plant that removed liquid. The craters were rewetted air arriving at a gun that sat behind three separate filters, all of them working.

Why water is different from every other contaminant

Four properties. Two of them water shares with whatever else the intake happens to be drawing that day, because compression concentrates every incoming contaminant and the yard sets that input rather than the machine. The other two are water's alone, and they are the two that defeat the mental model.

Its quantity is an input, not an output. Every cubic foot of air the compressor draws brings in the water that was in it. A machine making twice the air brings in twice the water at the same weather, and the same machine on a humid afternoon brings in several times what it did on a cold morning. You cannot maintain your way out of an input.

Compression concentrates it rather than creating it. The compressor does not add a molecule of water. It takes a volume of air and hands back a much smaller volume containing the same water, so at a discharge of 100 psig from an atmospheric intake the absolute pressure ratio is 114.7 to 14.7, close to 7.8, and the air leaving the airend is carrying its original water load in about an eighth of the space. Hot air holds it; the aftercooler then takes the temperature back down and most of it drops out as liquid immediately, which is why the wet receiver's drain is the busiest fitting in the room. What a compressor is doing to the air thermodynamically belongs to its own article; the consequence here is that liquid water is normal and expected at that point, not evidence of a fault.

It changes phase, so removal is local and temporary. A particle filter removes a particle once and the particle is gone. A separator removes liquid water and the vapour walks straight past it, because vapour is not a particle and no mechanical filter of any rating touches it. That vapour re-condenses at the next place in the system colder than its dew point. So every water-removal device protects only what is downstream of it, and only while the downstream pipe stays warmer than the dew point of the air leaving that device. Move the pipe outdoors, or drop the shop temperature overnight, and the protection ends at that point in the run.

It carries the other contaminants. Liquid water in a steel main strips scale and delivers it downstream as particulate. It emulsifies compressor oil carryover into a milky film that a coalescing element passes differently from clean oil. It gives corrosion a cell to live in at every low point and threaded joint. This is why a moisture problem so rarely presents as a moisture problem: it arrives on the ticket as rust in the header, oil at the tool, or a filter that plugs too fast.

What this changes about how you diagnose

Once water is understood as an input that changes phase, three diagnostic habits follow, and they are the opposite of the habits particulate contamination teaches.

Stop comparing the dryer against itself. A dryer meeting its nameplate has proved nothing about your plant, because the nameplate says nothing about your pipe. The comparison that matters is the air's pressure dew point against the coldest pressurised metal in the system, and taking that reading is its own procedure.

Walk the pipe before you specify anything. The coldest pressurised metal decides the requirement, and it is almost never in the air room. It is a roof run, a dock header, a branch through an exterior wall, the line to a compactor outside. In this shop it was twenty feet of drop nobody had walked.

Read the season, not the service interval. A consumable at a point of use whose life swings by a factor of several between winter and summer is reporting the moisture load directly. That swing is data, and in this case it was the only instrument in the plant that had been watching the problem all along.

What was done and what would have changed it

The drop was insulated and rerouted inside the wall line, which raised the coldest pressurised metal above the dryer's outlet dew point year round and removed the condensing site. The point-of-use cartridge went back to a monthly interval in both seasons, which was the confirmation that the load reaching it had changed rather than the cartridge having improved.

What would have changed the answer is the coldest metal temperature. Had the booth's drop been inside conditioned space, insulating it would have fixed nothing, and the finding would have moved upstream to the dryer class itself, because at that point the air's dew point rather than the pipe's temperature would have been the term out of position. Two fixes, and the pipe walk is what tells you which plant you are in.

Changing that desiccant cartridge is not a bare-hands job either. Desiccant media generates respirable dust when a bed or cartridge is opened, dumped or refilled, so isolate the housing at its upstream valve, open the vent, confirm the housing gauge reads zero with the vent still open, and lock and tag that isolation under 29 CFR 1910.147 before the housing is cracked; then handle the media with a respirator selected and fit-tested under the employer's respiratory protection programme required by 29 CFR 1910.134, not a nuisance dust mask, because the route here is inhalation and gloves do nothing for it.

Where the stakes stop being about finish quality

One case moves this out of the quality column entirely. Where plant air feeds an atmosphere-supplying respirator, the air is no longer a product-quality question and becomes a respiratory protection requirement: 29 CFR 1910.134(i) requires compressed breathing air for supplied-air respirators to meet at least the Grade D requirements described in ANSI/CGA G-7.1, and sets the monitoring arrangements that apply where an oil-lubricated compressor is the source. A shop that has quietly teed a supplied-air hood off the same header that runs its impact wrenches has taken on that programme whether or not anyone wrote it down, and the moisture, carbon monoxide and hydrocarbon questions all belong to it rather than to the paint booth.

How to verify you have read this right on your own site

Three checks, none of which needs an instrument you do not already have.

Take the temperature of the coldest pressurised metal in the plant in the coldest week you have data for, and compare it against the dew point the dryer is rated to deliver at its own reference conditions. If the metal is colder, you have a condensing site downstream of your treatment regardless of how healthy the air room looks.

Look at the service life of every point-of-use consumable across two seasons. A life that halves in summer is a moisture load talking to you; a life that is stable across seasons and short in both is a different contaminant.

Open the wet receiver's drain path and see what the plant produces in a shift. Liquid there is normal and its absence is the finding worth chasing, because an aftercooler and separator that produce no condensate on a humid day are either bypassed or not cooling. Stand to the side of the discharge with it directed into a container or a floor drain, with eye and hearing protection on, and never look into the path or point it at a person. Never use a blow gun to clear a blocked drain line: 29 CFR 1910.242(b) permits compressed air for cleaning only when reduced to less than 30 psi and then only with effective chip guarding and personal protective equipment, and a plugged drain line is cleared by replacing it.

References

• 29 CFR 1910.134, respiratory protection, including the breathing air requirements at (i) that apply where plant air supplies an atmosphere-supplying respirator, and the programme requirements for respirator selection and fit testing
• 29 CFR 1910.147, control of hazardous energy, for isolating and verifying depressurisation of a filter or dryer housing before opening it
• 29 CFR 1910.242(b), compressed air used for cleaning
• ANSI/CGA G-7.1 Grade D breathing air, which binds through 29 CFR 1910.134(i) rather than on its own authority
• See related: Where the Water Comes From and Where It Goes; How to Tell Whether a Dryer Is Doing Its Job; What a Compressor Is Actually Doing to the Air