Oil in the main header of a lubricated rotary screw plant is not an anomaly to be discovered — it is a known output of the machine that the filtration train is supposed to remove. Three years of visible drips means that train is undersized, saturated, misplaced, bypassed, or draining back into the pipe instead of out of it. Work the fix in order: restore the filters and their drains first, have the airend's oil separation circuit checked second, and add a vapour-phase stage only if the process still fails the purity target after the liquid and aerosol stages are proven. Replacing the oil-wetted hose downstream of the first UFMT-1 is part of the repair, not cosmetics — elastomer that has absorbed oil for three years will keep releasing it.
Reading the Oil Trail
Three markers are worth separating before anyone touches a wrench. Drips in the main line are liquid-phase oil that has already coalesced — it arrived as aerosol and dropped out where velocity fell, or it passed through a filter as liquid. The grease-like marks on the hose at the outlet of the first UFMT-1 are oil emulsified with condensate; free water and oil aerosol combine downstream of any cooling step and leave a sticky film rather than a clean oil sheen. And the fact that the trail begins at a filter outlet is the single most useful datum in the whole system: a filter housing that is clean on the inlet side and wetted on the outlet side is re-entraining, not filtering.
Re-entrainment has three causes and only three: the element is blinded and the pressure drop is now high enough to strip coalesced liquid off the drain layer; the bowl has filled because the automatic drain has failed and the element is sitting in liquid; or actual flow through the housing exceeds the element's rated flow, so face velocity is above the coalescing regime. All three get worse with time, which matches a problem that grew over two to three years and survived a January element change.
Note also what did not fix it. Elements were replaced in January under a calendar schedule and the drips continued. That eliminates "dirty filter" as the sole cause and points either upstream of the filters — the compressor itself — or at the architecture: a correctly sized element in the wrong position removes nothing.
How Oil Leaves a Lubricated Screw Compressor
An oil-flooded screw injects lubricant into the compression chamber for sealing, cooling and rotor lubrication. Downstream of the airend, the oil-air mixture enters the receiver/separator vessel, where bulk oil drops out by gravity and impingement, and the remaining aerosol is captured by a coalescing separator element. Coalesced oil pools at the base of that element and is returned to the airend inlet through a scavenge line containing an orifice, a strainer and usually a check valve. Carryover in a healthy new machine runs on the order of 2-3 ppm by weight. That is not zero, and it is not small in annual terms.
Quantify it before deciding how much filtration to buy:
Take and loaded hours from the CompAir data plate and the controller's load-hour counter, not from the total hour meter. Whatever the exact number, the conclusion holds: an unfiltered lubricated compressor pushes litres of oil per year into the distribution system, and it keeps doing so until a filter stops it.
Carryover rises above the nominal figure when any of the following drift: separator element differential pressure climbs as the element blinds; the scavenge orifice or strainer plugs so oil pools and is swept out with the air; sump level is overfilled and floods the element; discharge temperature rises, increasing the fraction of oil present as true vapour rather than aerosol; the minimum pressure valve fails to hold sump pressure during unloaded or low-demand running, raising velocity through the element; or the lubricant is degraded, wrong grade, or foaming.
Oil Phase Versus the Device That Removes It
Filters are phase-specific. Choosing by micron rating alone is how plants end up with three filters and still get drips.
| Phase in the air stream | Removal device | Correct position | Failure mode if misplaced |
|---|---|---|---|
| Bulk liquid oil and water slugs | Cyclonic / centrifugal water separator, inline oil-water separator | Immediately after the aftercooler, before any coalescer | None — but if omitted, the next coalescer floods |
| Coarse aerosol and solid particulate | General-purpose coalescing filter (~1 micron class) | After bulk separation | Loads with liquid, dP climbs, re-entrains |
| Sub-micron oil aerosol | High-efficiency coalescer (0.01 micron class) | After the general-purpose stage, on cooled air | Blinds within weeks if fed unprotected |
| Condensed water | Refrigerated or desiccant dryer | After coalescing stages | Water re-emulsifies oil residue downstream |
| Oil vapour and hydrocarbon odour | Activated carbon adsorber / oil vapour eliminator | Last stage, on dry air only | Instantly consumed by liquid or aerosol carry-in |
The minimum for a lubricated machine feeding a clean-air process is two coalescing stages in series — one to take the bulk liquid and water, one to take the sub-micron fraction. Add the vapour stage only where the process is sensitive to hydrocarbon vapour, and only behind proven aerosol removal.
Symptoms Versus Causes
| Observation | Most probable cause | Check that decides it |
|---|---|---|
| Oil film on the outlet hose of the first filter | Re-entrainment: blinded element, flooded bowl, or over-flow | Filter dP at load; open the bowl and look for standing liquid |
| Drips persist after a fresh element change | Source is upstream (separator/scavenge) or the stage is wrongly positioned | Separator element dP; scavenge line flow and orifice condition |
| Emulsified, grease-like deposit rather than clear oil | Free water co-present; no dryer or drains not discharging | Pressure dew point; manual drain override test |
| Carryover worse at low plant demand | Minimum pressure valve not holding sump pressure when unloaded | Sump pressure during unloaded running against the manual's figure |
| Carryover worse on hot days or after long full-load runs | High discharge temperature increasing vapour fraction; fouled aftercooler or oil cooler | Discharge temperature trend vs. thermostatic valve setpoint |
| Oil appears only at certain drops | Legacy contamination in that branch, or a downstream lubricator feeding back | Isolate the branch, blow to a clean white card |
Restoration Procedure
- Log the baseline before changing anything: total and loaded hours on both machines, discharge temperature, separator element dP from the panel, differential pressure across each of the three UFMT-1 housings at normal flow, sump oil level, and top-up quantities from the maintenance log.
- Service all three filter housings with the correct elements — not two of three. A blinded element passes more oil than an empty housing does; running two degraded units while spares are on order sustains the problem.
- On each housing, verify element orientation and seating. Coalescing elements flow inside-out; a reversed or unseated element bypasses entirely past the O-ring. Confirm every bypass valve around the filter station is closed and tagged.
- Test and, where the test fails, replace each automatic drain. Route drain discharge to an oil-water separator, not to floor or to a common header that can develop back pressure.
- Correct the architecture. The stage at the compressor outlet handles hot, wet, unseparated air — it belongs there only as a bulk separator. Sub-micron coalescing belongs downstream of the aftercooler and water separation, where the two units at the facility entrance already sit. Commission the idle clean-air filter station in that position rather than leaving it as an unused asset.
- Deal with legacy contamination. Piping and hoses that have carried oil for three years hold a film that keeps bleeding. Replace oil-wetted flexible hose, blow down low points and drip legs, and drain every filter and drop that has been accumulating.
- Set service intervals on differential pressure and elapsed element life, with the December-January outage as the scheduled replacement window rather than the only trigger.
Verification Checks
- Filter differential pressure. Read each housing at normal plant flow with fresh elements. Expect the clean-element value from the Donaldson datasheet; anything materially above it on a new element means the housing is flowing above its rating. Replace at the manufacturer's maximum dP, whichever comes first with the time-based interval.
- Separator element dP. Read from the compressor panel at full load. Expect a low, stable value on a new element; a value at or near the manual's replacement limit is a direct explanation for the carryover.
- Scavenge line function. With the machine at load, feel the return tube — expect it warm along its length. A cold scavenge line means no flow: plugged orifice or strainer.
- Drain test. Manually override each automatic drain. Expect a short discharge of condensate followed by air, then reseat. Continuous air blow means a failed seat; nothing at all means a blocked drain and a flooded bowl.
- Discharge temperature. Expect a stable reading inside the thermostatic valve's control band and well clear of the high-temperature trip. A slow upward trend across a shift points at cooler fouling.
- Oil consumption. After 500-1000 loaded hours, compare sump top-up volume against the figure in the CompAir manual. Expect it at or below that figure; above it, oil is still leaving through the air side.
- Pressure dew point, once a dryer is in the train. Expect the dryer's rated PDP at rated inlet conditions; a high reading means water is still reaching the header and will keep producing the emulsified deposit.
Recurring Pitfalls on Filter Trains
A coalescer sized on pipe connection rather than actual flow is the most common installed error. Coalescing efficiency collapses above rated flow because captured droplets are stripped from the drain layer faster than gravity can carry them down; the filter then behaves as an aerosol generator. Size on the compressor's FAD plus a margin for both machines running, at the lowest expected line pressure — efficiency and capacity are both stated at a reference pressure and fall with it.
Placing a high-efficiency coalescer immediately at the compressor outlet looks intuitive and does not work. That air is hot, at or above its dew point, and carries bulk liquid; the element floods, dP climbs within days, and the housing passes oil while appearing to be in service. Bulk separation there, fine coalescing after cooling.
Activated carbon gives no warning of exhaustion. There is no dP rise when the bed is saturated — it simply stops adsorbing and the first sign is odour at the point of use. Change it on time-in-service and mass loading, and protect it absolutely with the coalescing stages, or it is money spent for weeks of benefit.
Two more that recur: drains piped into a shared header that develops back pressure and stops every drain in the group from opening; and downstream lubricators or in-line oilers on tool drops that push oil back toward the header during pressure transients. Where a machine is genuinely oil-free and oil still appears, one of those two external sources is almost always responsible.
Frequently Asked Questions
How do I tell whether oil in the line is coming from the compressor or from a failed filter?
Compare inlet and outlet of each filter housing. A housing clean on the inlet side and wetted on the outlet side is re-entraining — check its dP, its bowl level and its drain. If every housing is clean on both sides and oil still reaches the header, read the compressor's separator element dP and check the scavenge line for flow; a cold scavenge tube at full load means the orifice or strainer is plugged.
How do I size a coalescing filter for oil removal?
Size on actual volumetric flow at the lowest expected line pressure, not on pipe connection size, and include both compressors running if they can run together. Filter ratings are published at a reference pressure and capacity falls as pressure falls, so a unit that is nominally adequate at 7 bar can be over-flowed at 5 bar. Take the flow figure from the compressor data plate and have the filter supplier confirm the selection against your duty.
How do I clear oil out of piping and hoses that have been contaminated for years?
Replace oil-wetted flexible hose outright — elastomer absorbs oil and releases it for months after the source is fixed. For rigid pipe, drain every low point and drip leg, blow down each branch to a waste collection point, and drain all filter bowls. Then verify with the white-card test at the most remote drop, which is where residual oil shows up last.