What Condensate Tells You About the System That Made It

Why this matters

Condensate is the only water in any of these systems that arrives pure. Steam condensing in a return and moisture condensing on a coil both start as distilled water, which means every measurable property it has, including how much of it there is, was put there by the equipment. That makes it the most sensitive sample on the job and the one almost nobody takes. A conductivity reading on a condensate return costs a minute and finds a leak that a year of routine boiler water tests will never see, because boiler water is already full of things and a change of a few percent hides in it.

Before you take a condensate sample

• Steam condensate is hot and often above atmospheric boiling. Sample through a cooler, never into an open container. Water above 212 F flashes to steam the moment it reaches atmospheric pressure.
• Rooftop and above-ceiling work carries the hazard that actually injures people on this equipment. Fall protection under 29 CFR 1910 Subpart D for general industry work or 29 CFR 1926 Subpart M where the job is construction, before anything about water.
• Lock and tag the fan motor under 29 CFR 1910.147 before reaching into a unit to inspect a pan or a trap; panel work is 29 CFR 1910.333(b)(2), with live-dead-live per NFPA 70E-2021, 120.5.
• Cooling coil condensate carries whatever the coil and the drain grew. Wear gloves, and do not blow a drain line clear with compressed air or nitrogen into an occupied space: that aerosolises biofilm. Cap and direct the discharge outdoors or into a contained receiver, and stand out of the plume.
• Opening a ceiling that has water standing above it puts that water, and the tile, on whoever is below. Clear the area, and treat wet ceiling work near lighting or a fire alarm device as an electrical hazard until the circuit is proven dead.

The gate

One question, asked of any condensate stream: it started as pure water, so what is in it that pure water could not have carried, and how much of it should there be? Both halves are diagnostic, and the two cases below resolve on opposite halves.

What "pure" means, and the one thing that is never zero

Freshly condensed water is not chemically neutral, and knowing that stops a false finding. Pure water in equilibrium with atmospheric carbon dioxide sits near pH 5.6, because dissolved carbon dioxide forms carbonic acid. Slightly acidic condensate is therefore the expected state and not, by itself, evidence of anything.

The numbers that mean something are the departures. Clean steam condensate from a well-treated system runs at a very low conductivity, in the single digits to low tens of microsiemens. Condensate from a condensing gas appliance runs far more acidic, commonly in the 3 to 5 pH range, because combustion products dissolve into it; take the figure and the neutraliser requirement from that appliance's own manual, since it is the manufacturer that specifies both. Cooling coil condensate is essentially air-washed distilled water with whatever the coil surface contributed.

Case one: something in it that pure water could not carry

A steam system whose condensate return had been sampled at commissioning at about 10 microsiemens. A routine check two years later read 240 microsiemens. Nothing else had changed on the panel, and the boiler water tests were in band.

Pure water does not become conductive on its own, so something ionic entered the return. Estimate how much with a mixing calculation. Site makeup water measures about 600 microsiemens. If raw makeup is what is entering, the fraction of the return that is not condensate is the excess over baseline divided by the difference between the contaminant and the baseline: 240 minus 10, over 600 minus 10, which is 230 over 590, or about 39 percent.

Confirm with a second species before acting, because a single number is a hypothesis. Return hardness measured 46 mg/L as calcium carbonate against makeup at 120 mg/L, which is 38 percent. Two independent estimates at 39 and 38 percent is a finding: roughly two-fifths of what is coming back is not condensate at all.

That points at a heat exchanger. Before naming the direction, establish the pressure differential, because that is what sets which way the leak runs. On a steam-to-domestic-water exchanger with the condensate return near atmospheric and the domestic side at street pressure, water flows from the domestic side into the condensate, which is what this sample shows. Reverse the differential, which happens on a pressurised return or when the domestic side is drawn down, and boiler-treated water enters the potable system. That is a cross-connection and a public-health event, not a maintenance item: isolate the exchanger, notify the building, and involve the authority having jurisdiction rather than repairing it quietly.

The failure mode on this case is the one that made it a two-year problem. Everyone was watching boiler water, and 39 percent of the return being raw water raised boiler feedwater solids by an amount that blowdown absorbed without anyone noticing a step change. Boiler water is a poor detector because it is already loaded. Condensate is a good detector because it starts at zero.

Case two: nothing in it, because there is none

A rooftop unit, roughly five tons, running in humid weather, with the space reading 62 percent relative humidity and the tenant complaining it feels damp. The condensate drain is dry.

Here the gate resolves on its other half: there is no sample, so the question is what volume there should be. Work it from the load. Five tons is 60,000 Btu/h of total capacity. Latent heat of vaporisation at coil conditions is near 1,060 Btu per pound, and water is 8.34 pounds per gallon, so condensing a gallon of water absorbs roughly 8,840 Btu.

The latent share of total capacity is the variable, and it is genuinely a range rather than a constant, so run the ends. At a 25 percent latent share, 15,000 Btu/h of latent capacity divided by 8,840 Btu per gallon is about 1.7 gallons an hour. At a 15 percent share it is about 1.0, and at 35 percent about 2.4. So a five-ton unit doing its job in humid weather should be putting somewhere between about one and two and a half gallons an hour down the drain, and a dry drain during runtime is a real finding rather than an impression.

Two families of explanation, and they need separating before anyone touches a trap.

The water is being made and is going somewhere else. A blocked trap with the pan overflowing into the ceiling, a cracked or corroded pan, a pan draining backwards because the unit is not level, or a missing or incorrectly sized trap on a draw-through coil letting the fan pull air in through the drain line so water stands in the pan instead of leaving it. The tell is water somewhere: a stained tile, a wet insulation face, a full secondary pan.

The water is not being made. The coil is not below the air's dew point, so nothing condenses. And this is where direction has to be established rather than assumed, because both ends of the airflow range end at a dry drain by opposite mechanisms. Too much airflow raises coil surface temperature, cuts the latent share, and reduces condensate. Too little airflow lowers coil temperature, raises the latent share and produces MORE condensate, until the surface goes below freezing and the coil ices, at which point the drain goes dry while water accumulates as frost. A warm coil and a frosted coil both give you a dry drain, and touching the coil separates them in one second.

What the two cases share

The gate is the same and the instrument is different, and that is the portable lesson. In case one the sample existed and its composition was impossible for pure water. In case two the composition question could not be asked, so the volume question carried it. Both were answered from the physics of what condensate is, before anyone opened anything.

It generalises past these two systems. Condensate from a condensing appliance that has stopped being acidic is telling you the neutraliser media is doing its job or that the appliance is not condensing at all, and which one it is depends on whether you sampled upstream or downstream of the neutraliser. Defrost water from a refrigeration coil that carries a taste or an odour is carrying something off the coil surface. A dehumidifier that produces nothing in a damp room is the same case-two arithmetic at a smaller size.

How to verify a condensate reading before you act on it

• Establish a baseline before you need one. A commissioning conductivity reading on a return is what makes a later number mean something; without it, 240 microsiemens is a number with no comparison.
• Sample from a fixed point and say which point. A return header, a receiver and a pump discharge do not read the same, and a sample taken from a receiver open to atmosphere has been picking up carbon dioxide from the room air.
• Use two species before naming a contaminant. Conductivity plus hardness, or conductivity plus a species unique to the suspected source, so the mixing fraction is confirmed rather than assumed.
• Measure volume by timed catch, not by eye. Catch the drain into a container of known volume for a measured period during steady runtime, and compare against the load arithmetic rather than against a memory of what a drain usually does.
• Check the trap before the coil on a dry drain, because a trap fault is common, cheap and visible, and because a coil temperature reading taken while a plugged trap is holding water in the pan can be misread as normal operation.

References

• Appliance manufacturer documentation for condensing appliance condensate pH and required neutralisation
• Water treatment supplier documentation for condensate return conductivity and iron limits appropriate to the system
• 29 CFR 1910 Subpart D or 29 CFR 1926 Subpart M for the fall hazard; 29 CFR 1910.147 for fan isolation; 29 CFR 1910.333(b)(2) and NFPA 70E-2021, 120.5 for electrical work
• See related: Why a Condensate Line Is a Chemistry Problem Too; What a Boiler Does to the Water in It