What a Condensate Receiver and Pump Set Is Doing

Why this matters

A condensate receiver looks like the least interesting object in a boiler room: a tank with a float switch and a pump bolted to it. It gets treated that way, which is why the same set gets a third pump in five years while nobody measures the two dimensions that were wrong from the day it was installed. The receiver is not storage. It is the only place in a steam system where an intermittent, two-phase, boiling arrival is converted into the continuous single-phase liquid a pump can actually move, and it does that with exactly two features: a vent and an elevation. When a set misbehaves, one of those two is almost always the reason, and neither of them is the pump.

What actually arrives at the inlet

Everything upstream of the receiver is a trap discharging into a pipe. A trap does not deliver water at the pressure it was holding. It delivers water that was at saturation temperature for the equipment pressure into a return line at a much lower pressure, and a fraction of that water immediately re-evaporates because it is carrying more heat than a pound of water can hold at the new pressure.

The fraction is arithmetic, not a guess. It is the sensible heat the water has to give up divided by the latent heat needed to boil a pound at the receiver pressure:

flash fraction = (h-f at the equipment pressure - h-f at the receiver pressure) / h-fg at the receiver pressure

For 100 psig condensate landing in a receiver open to atmosphere, h-f is about 309 Btu/lb at 100 psig, about 180 Btu/lb at 0 psig, and h-fg at 0 psig is about 970 Btu/lb. That is (309 - 180) / 970, or 13.3 percent by weight. Take the three enthalpies from steam tables for your actual pressures rather than reusing these; the method travels, the values do not.

Thirteen percent by weight sounds modest until you look at volume. A pound of saturated steam at 0 psig occupies about 26.8 cubic feet; a pound of water occupies about 0.0167. So the 0.133 lb of flash from that pound takes up about 3.56 cubic feet and the remaining 0.867 lb of water takes up about 0.014. Of what arrives at the receiver inlet, roughly 99.6 percent is vapour by volume. That is the whole reason the receiver exists and the reason its vent is not an afterthought.

The vent is the working part

The vent's job is to let that vapour leave at atmospheric pressure so the liquid can settle out and the tank can accept the next slug of return. Everything about a vented receiver depends on the vent being full size, unvalved, and terminated where a plume of live steam cannot reach a person or a walkway.

      vent to atmosphere, full size, no valve
                    |
  trap discharge    |
  ------------->+---+-------------------+
                |   flash leaving here  |
                |~~~~~~~~~~~~~~~~~~~~~~~|  pump-off level
                |                       |
                |  liquid sitting at    |
                |  its own boiling      |
                |  point, not below it  |
                +------+----------------+
                       |  suction
             submergence is measured from the
             pump-off level down to the pump
             suction centreline
                       |
                    [ pump ] ---> boiler feed

Throttle that vent, cap it "to stop the steam", or let it silt shut and the receiver pressurises slightly. A pressurised receiver does not just push back on the traps upstream, which is bad enough. It raises the boiling point of the liquid inside, and the liquid follows within seconds because flash is still arriving to heat it. You get a hotter tank and the same margin at the pump, which is covered in the sibling article on why a condensate pump cavitates when nothing is wrong with it. Do not diagnose a plume at the vent as a fault. On a receiver taking flash from medium-pressure traps, a visible plume is the vent doing its job, and the first thing to check is whether anyone can walk into it.

Elevation is the other working part

The pump takes suction from liquid that is at its own saturation temperature, so there is no subcooling to draw on. The only margin available to the suction is the vertical distance from the pump-off water level down to the suction centreline, less the friction in the suction piping and less the acceleration losses of an inflow that arrives in slugs. That is why receiver sets are floor-mounted with the pump under the tank and why a set relocated onto a housekeeping pad without moving the tank stops working for reasons nobody connects to the move.

Two elevations get measured at commissioning and neither is on most drawings: the pump-off level to suction centreline, and the inlet connection to the high water level. The inlet has to land above the water line. An inlet that ends up submerged turns the return main into an extension of the tank, so returning condensate has to displace liquid instead of dropping into vapour space, and the flash that would have left through the vent now has to travel back up the return line against incoming flow.

What the pump is being asked to do that a water pump is not

A general-service water pump moves cold, subcooled, continuously available liquid. A condensate pump moves hot liquid at saturation, arriving in bursts, with dissolved gases coming out of solution the whole time, into a discharge pressure set by the boiler rather than by a building loop.

Three consequences follow, and all three show up as complaints that get blamed on the pump:

• It is sized for surge, not for average. Condensate does not arrive at the average rate. It arrives at the running rate for most of the day and then at several times that during a warm-up. Common practice is to size the pump at two to three times the average return rate so it runs intermittently and can catch up after a start-up surge. Take the multiplier from the manufacturer's selection guidance for the set, which is the party that owns that number.
• The receiver is sized so the pump does not short-cycle. The usual floor is enough volume to hold at least one minute of pump capacity between the pump-on and pump-off levels. A tank that satisfies the surge but not the cycle produces a motor that starts every twenty seconds, and the failure that follows is a motor failure with a hydraulic cause.
• Seals live in hot water. Anything rated for general water service and installed here fails early and looks like a defect.

When a receiver set is the wrong answer

Gravity return does not need a receiver at all, and adding one because it feels like good practice adds an air-exposed hot surface, a motor, and a control to a system that had none. Two conditions genuinely call for the set: the return has to be lifted or pushed against a boiler feed pressure that gravity cannot supply, or several returns at different pressures need a common collection point before the boiler.

There is a third case a receiver set specifically does not solve, and it gets bought for it regularly. Where equipment stalls because a modulating control valve has dropped the coil pressure below the return system's back pressure, the flooding happens upstream of the receiver and a bigger pump does not reach it, which the sibling article on locating water hammer works through.

The artifact: one receiver set, commissioned

This is the sheet worth filling in on any set, new or inherited. The values below are one worked case; read the method, not the entries. Assume a return collecting from 100 psig distribution, an average return rate of 1,200 lb/hr taken from the boiler's own feedwater metering over a week rather than estimated, and an atmospheric receiver.

Two entries do the real work. The volumetric conversion uses about 8.0 lb/gal because condensate near 200 F is around 60 lb per cubic foot, not the 8.34 lb/gal you would use cold. And the submergence pair is the only line where the required value belongs to the pump manufacturer and the available value belongs to the installer. Compare those two before touching anything else.

Now run the surge. If the warm-up load on this distribution is three times the running load, the receiver sees roughly 7.5 gpm arriving against a 6.2 gpm pump. The set does not fail; the level rises toward the high mark and the pump runs continuously until the warm-up ends. What tells you the set is undersized is not the level rising, it is the level reaching the overflow, and the overflow is the field measurement nobody records. Mark the high water level with a paint pen at commissioning and read it after the coldest start of the season.

Verifying a set you did not commission

Do these in order, with the set running normally and nobody standing in the vent plume or in front of the overflow.

Read the sight glass through one full pump cycle and time it. A run time under about ten seconds means the working volume is short or the pump is oversized for the receiver, and both are fixable at the float switch before anyone quotes a tank.

Look at the vent from the floor, never with a hand. A steady low plume during normal operation is correct. A vent that pulses violently in time with a trap upstream says a single high-pressure trap is dumping into a receiver that cannot pass its flash, and a vent that is dry while the tank is hot says it is blocked. Do not put a hand, a rag or a mirror near a vent to test it; watch the plume against a dark background from outside its path.

Check the inlet elevation against the high water level with the set off. A tape on the outside of the tank plus the sight glass, with nothing opened.

If the tank itself has to be opened - for a float, a strainer, or scale - stop the return, isolate and lock out the pump circuit under 29 CFR 1910.333(b)(2) for the electrical side and 29 CFR 1910.147 for the stored energy in the return piping, prove the circuit dead with the live-dead-live sequence in NFPA 70E-2021, 120.5 as adopted by your employer's electrical safety program, confirm the vent is open to atmosphere, and let the contents fall below 120 F verified with a non-contact reading taken at arm's length before any cover comes off. Condensate sitting at 200 F flashes when the vessel is opened, and it does not need to be pressurised to scald.

References

• ASME Steam Tables, or an equivalent published saturated-steam table, for h-f, h-fg and specific volume at the pressures in your system
• 29 CFR 1910.147, control of hazardous energy, for isolation of the return piping and stored pressure
• 29 CFR 1910.333(b)(2) for electrical isolation of the pump circuit, and NFPA 70E-2021, 120.5, in the edition your employer's electrical safety program adopts, for the live-dead-live proving sequence
• Manufacturer selection and installation literature for the receiver set, which owns the required submergence and the pump sizing multiplier
• See related: articles in this library on why a condensate pump cavitates when nothing is wrong with it, and on what condensate tells you about the system that made it