What a Steam Coil Does and Why It Freezes

Why this matters

A steam coil that splits does not fail because the steam was off. It fails because condensate could not get out while cold air was still crossing it, and in almost every case the drainage was already failing for most of the heating season before the split happened. The coil is the part that breaks. The return piping is the part that was wrong.

If a coil is suspected of having burst, treat it as a live steam release into occupied space. Shut the supply fan and close the steam supply valve from outside the air handler, keep everyone away from the supply diffusers that unit serves, and do not open an access door on a unit that still has a live steam connection, because the plenum behind that door is at supply temperature and the door will vent it into your face. Steam is invisible for the first stretch out of a split tube.

What the coil is actually doing

Steam enters the coil, condenses on the inside of the tubes, and gives up its latent heat through the tube wall to the air. The condensed water has to leave, continuously, or it takes up tube volume that was doing the heating.

Two things follow from that and both matter later.

The coil's temperature is its pressure. Condensing steam sits at the saturation temperature for whatever pressure is inside the coil, so a modulating steam valve is a temperature control wearing a pressure control's clothes. Throttle the valve and the pressure in the coil drops, the saturation temperature drops with it, and the coil delivers less heat. That is exactly how it is supposed to work.

The trap has to push the condensate somewhere. It has to overcome the pressure in the return main, plus any vertical rise between the trap and that main, plus the friction of the return line. Water column converts at roughly 2.31 feet per psi, so a lift is real back pressure even though no gauge in the room displays it.

Put those two sentences next to each other and the whole problem appears. The pressure the trap has available falls with load. The pressure it has to push against does not.

The call

Mid-January, six in the morning, no heat at a school. Hundred percent outdoor air unit, 12,000 cfm, preheat coil rated at 10 psig steam, design entering air 5 F, design leaving air 55 F. Techs arrived to water on the mechanical room floor and a split in the bottom two rows on the leaving-air face.

Two things were already known before anyone touched a tool. Three coils in six years. Every one of them split in the bottom rows.

That repetition is the finding. A coil that freezes because a valve failed shut on one cold night is a one-time event with a one-time cause. Three coils in six years, all in the same rows, is a condition, and a condition has arithmetic behind it.

The explanation that was wrong

The first two replacements had been written up as a control failure, and the freeze-stat had been replaced once along the way for tripping too often.

That is the part worth stopping on. A freeze-stat that trips repeatedly is evidence, and replacing it converts evidence into silence. A protective device that opens is doing its job until you establish why it opened; treating it as the fault reaches the burst coil one step slower and with a part number on the invoice. The correct response to a repeatedly tripping freeze-stat is to find out what is making the coil cold, and the second replacement went out with that question unasked.

The control valve, when it was finally checked, stroked correctly and held its setpoint. The steam supply pressure at the coil was available. Nothing in the control path was broken.

The number that closed it

Read the return main gauge: 3 psig. Measure the vertical rise from the trap outlet up to the return main with a tape: 5 feet, which is about 2.2 psi. Add roughly 0.5 psi for return line friction.

  outdoor air ->  [ steam coil ]
                        |
                   drop leg, full connection size
                        |
                      trap
                        |
                   5 ft rise
                        |
  ==================+============ return main, 3 psig

  the rise adds about 2.2 psi, so the coil must
  hold over about 5.7 psig for the trap to discharge

So the trap sees about 5.7 psig of back pressure, which from the steam tables is a saturation temperature of about 228 F. The coil has to be hotter than 228 F for one drop of condensate to leave it.

Now find out how often it is. With airflow constant and the surface unchanged, a coil holds a fixed fraction of the available temperature difference in reserve on the leaving side, and that fraction is what its design point tells you. At design, steam 239 F, air in at 5 F and out at 55 F: the coil captured 50 of the 234 degrees available and left 184, so it holds about 79 percent of the available difference in reserve. That fraction is valid as long as airflow and surface stay put, and on this unit both did.

The control holds leaving air at 55 F year round, so run the same fraction backwards at warmer entering air:

Read the last column top to bottom. This coil drains only when it is below about 8 F outside. Above that it is stalled, meaning the pressure inside it has fallen below what the trap has to push against, and condensate stops leaving. The coil then fills from the bottom up, which is why every split was in the bottom rows.

And the freeze exposure is the band between those two lines: entering air from about 8 F up to 32 F, where the coil is holding water and the air crossing that water is below freezing. In this climate that band is most of the winter. The coil was not unlucky three times. It was standing in a freezer holding a glass of water for six years.

Note the second row of that table as well. Once entering air is much above 10 F the coil is under vacuum, not just low pressure. A coil under vacuum without a vacuum breaker cannot drain even into an open drain, because atmospheric pressure is holding the water in.

What each fix actually solves

Four options, and they do not solve the same thing, so pick by the mechanism rather than by cost.

A vacuum breaker plus a gravity drain to a vented receiver below the coil, then a pump. The vacuum breaker admits atmospheric air when the coil goes below atmospheric, so the coil can drain by gravity. The vented receiver removes the 3 psig and the 5 foot lift from the trap's problem entirely, and the pump handles the return from there. This addresses both halves: the vacuum and the back pressure.

A pump-trap at the coil. A combination pressure-powered pump and trap takes condensate by gravity into its own body, then uses motive steam to push it against the return pressure. It works when the coil has no pressure at all, which is the defining condition of stall. This is the standard purpose-built answer for a modulated coil on a lifted return.

Two-position control with face-and-bypass dampers. The steam valve is either fully open or fully shut, so the coil is always at full pressure and never stalls, and capacity is modulated by moving air around the coil instead of through it. This is the traditional preheat arrangement and it is still the most robust one, at the cost of damper linkage that has to be maintained.

Nothing on the control side. Rebuilding the valve, retuning the loop, or resetting the leaving air setpoint changes when the coil stalls, not whether it does.

Two things that get bought and do not fix this are worth naming plainly. A larger trap does not help, because there is no differential for any trap to work with; the trap sizing article's gate exists to catch exactly this before a capacity is looked up. A steam-distributing tube coil, sometimes sold as a non-freeze coil, reduces the temperature stratification across the face that makes some coils freeze, and it does not drain a coil that cannot drain. A non-freeze coil on a stalled connection freezes too, later and in the same rows.

Confirming a coil is at risk before it splits

This is a walk that takes about twenty minutes per unit and it needs nothing opened.

1. Read the return main gauge. Not the drawing, not the schedule, the gauge. If there is no gauge on the return, that is the first finding and the second is that nobody has been able to check any trap on that system.
2. Measure the vertical rise from the trap outlet to the return main with a tape and convert at 2.31 feet per psi. Add the return main pressure and an allowance for friction. That sum is the pressure the coil must exceed to drain.
3. Read the coil's own steam gauge during mild weather with the unit running. If it reads at or below the number from step two, the coil is stalled at that moment. If it reads below zero, it is under vacuum. Both are conditions you can see on a day nobody is complaining, which is the point of doing this in the fall rather than in January.
4. Look for a vacuum breaker on the coil, and confirm its port is open to atmosphere rather than plugged with a fitting somebody added to stop a hiss.
5. Ask for the coil replacement history and where the splits were. Repeated bottom-row splits is the fingerprint; a single split anywhere else is a different story.
6. Verify the freeze-stat rather than assuming it. The capillary element responds to its coldest short length, so it has to be serpentined across the full coil face rather than run along one edge, and it has to be tested to the manufacturer's method with a cold pack on a section of the element, confirming both that the unit shuts down and that the alarm actually annunciates somewhere a person will see it. That test stops air to the space it serves, so it is scheduled with the building operator, announced to occupants, and not run on a unit serving a space that cannot lose conditioning without a plan in place.

The whole point of doing this in mild weather is that stall is visible on a 45 F day and a freeze is only visible on the morning after. A coil gauge reading below its own return back pressure in October is telling you exactly what January is going to cost.

References

• Coil manufacturer selection data, which is the authority for the coil's capacity at any entering air condition and steam pressure, and for whether a specific coil is built for modulated service
• Steam trap and pump-trap manufacturer literature, for stall charts and for the selection of a pressure-powered pump on a lifted return
• Saturated steam tables from any engineering handbook, for every pressure and saturation temperature pair used above
• 29 CFR 1910.147, the OSHA general industry energy control standard, for isolating stored pressure energy before a coil, trap, or return connection is opened
• See related: How to Size a Trap From the Load Rather Than the Pipe; Why an Oversized Trap Fails Differently Than an Undersized One; What Wet Steam Costs the Equipment Downstream