How Boiling and Condensing Differ From Heating and Cooling

Why this matters

Two exchangers can look identical on a drawing and behave nothing alike, because in one of them a fluid changes state and in the other it just gets warmer. That difference decides which side you should be cleaning, which side is worth improving, and which faults will take half your capacity overnight rather than degrading it over a season.

Techs who have not internalized it spend their time on the wrong surface. They rod out tubes on an exchanger whose problem is a blocked vent, or they chase a refrigerant-side film on a coil where the refrigerant side is a minority of the resistance. Be careful with the magnitude, because the clean 100-to-1 film-coefficient ratio is a BARE-SURFACE comparison and a coil is not bare: fins multiply the air-side conductance by the fin area ratio, so referred to a common area the refrigerant side of a finned coil is commonly 10 to 30 percent of the total, against 1 to 2 percent on bare tube. The conclusion holds either way, but spend the hour on the air side for the right reason.

The load-bearing point: boiling and condensing move heat so much more readily than single-phase flow that the two-phase side is almost never the bottleneck, which means routine diagnostic effort belongs on the single-phase side, and the only two-phase faults worth chasing are the ones that push that side back into single-phase behaviour.

Safety controls for the work below

• Venting a shell, a header, or a high point on a steam or hot-water system. The discharge is at saturation temperature and will scald anything in front of it. Pipe or direct the vent to a point where nobody stands, never toward yourself or into an occupied space, and open the valve from the side rather than over it. Steam blowing through a vent orifice is loud enough to matter; if you cannot hold a normal conversation at arm's length, use hearing protection under the occupational noise standard at 29 CFR 1910.95.
• Touching or probing a shell, a header, or a tube sheet. These surfaces sit at or near saturation temperature and cause a contact burn instantly at steam temperatures. Read with an infrared or strap-on instrument, not a hand, and do not remove insulation to reach bare metal on a live vessel.
• Opening any joint on a two-phase system. Isolate, drain to an open point, and confirm zero on a gauge on that section before loosening a fastener. A condensate line whose outer surface reads near ambient can still hold saturated condensate that flashes to steam the moment its pressure is dropped.
• Metering a control valve actuator or a level control. Electrical work sits at 29 CFR 1910.333(b)(2), separate from mechanical lockout. Prove the meter live, prove the circuit, prove the meter again, per NFPA 70E-2021, 120.5. Where you are isolating a pump or a fan rather than metering it, lock and tag under 29 CFR 1910.147 for mechanical and stored energy.

The two mechanisms, side by side

The temperature column is owned by the sibling card on what a phase change does to a reading, and the diagnostic consequences of that blindness live there. This card is about the other two rows.

The mass-flow consequence, which you can see on any job

Because latent heat per pound is so large, a two-phase side carries a given duty with a small fraction of the mass flow.

Take 400,000 Btu/h. On the condensing side, at around 960 Btu per pound of latent heat near low-pressure steam conditions, that is 400,000 / 960, about 417 lb/h. On a water side carrying the same duty with a 20 F rise, 500 x gpm x 20 equals 400,000 puts you at 40 gpm, which is 40 x 8.33 x 60, about 19,990 lb/h.

Roughly 48 times the mass flow on the single-phase side for the identical heat. That is why steam mains are small relative to their duty and hydronic mains are not, why a two-phase distribution system needs almost no pumping power, and why the pressure drop economics of the two are completely different.

One catch worth carrying: latent heat per pound falls as pressure rises. Pushing a steam system to a higher pressure to force capacity gets you a hotter fluid that delivers less heat per pound, so the mass flow has to rise more than the duty does. Take the actual value from steam tables at the operating pressure rather than reusing the atmospheric figure.

Coefficient magnitude decides where the bottleneck sits

Resistances add in series, so the largest one sets the total. Tabulated film coefficients put condensing and boiling roughly an order of magnitude above liquid forced convection and two orders above air, with the exact values depending strongly on geometry, velocity, and surface condition. Treat those as orders of magnitude and pull real values from design data when a number matters.

The consequence is blunt. On an air-to-refrigerant coil, the air side is essentially the whole resistance and the boiling refrigerant is a rounding error, which is why the coil has fins on one side and bare tube on the other, and why airflow problems and air-side fouling dominate the fault list. On a steam-to-water exchanger the water side is the limiting one, so tube velocity and tube-side fouling are the things that move the number.

That generalization has real limits and they are worth naming in the same breath, because each of them moves control back onto the two-phase side.

The four faults that turn a two-phase side single-phase

These are the only two-phase-side conditions worth chasing, and each has its own signature.

Flooding. Condensate that is not being removed sits on the surface and the exchanger loses that area to liquid convection instead of condensing. On a steam coil the usual cause is a trap that is undersized, blocked, or fighting back-pressure. On a refrigerant condenser it is excess charge or a restricted liquid path, which reads as high subcooling alongside a wide split.

Dryout. The opposite end of the same problem. If the boiling fluid runs out of liquid before the end of the surface, the remaining area is superheating vapor at a far lower coefficient. On a DX evaporator this shows as high superheat, and the cost is not mainly in UA (the air side still dominates) but in the fact that the dry portion of the surface runs too warm to condense moisture, so latent capacity collapses while sensible capacity looks acceptable.

An oil or liquid film on the wrong side. Oil that does not return properly coats the inside of the tube and adds a resistance in exactly the place where the fluid's own resistance was negligible. A small absolute resistance matters here precisely because it is being added to a very small number.

A non-condensable blanket. Air or another gas that will not condense collects at the surface where vapor is disappearing, and the condensing vapor has to diffuse through it to reach the metal. Even a small volume fraction can cut the condensing coefficient substantially, though the size of the effect depends on geometry and on where the venting is, so treat the direction as certain and the magnitude as system-specific. The tell is that a surface temperature reading sits well below the saturation temperature implied by the measured pressure, which is a symptom the phase-change card explains in full.

Worked case: half the UA, and the water side was fine

Steam-to-water shell and tube serving a heating loop. Complaint: loop supply temperature will not reach setpoint on cold days.

Design record on file: 400,000 Btu/h, steam at 5 psig (227 F saturation), water 40 gpm entering at 140 F and leaving at 160 F.

Design check: 500 x 40 x 20 = 400,000 Btu/h. The condensing side is at constant temperature, so the gaps are 227 - 160 = 67 F and 227 - 140 = 87 F. LMTD = (87 - 67) / ln(87 / 67) = 20 / 0.2612 = 76.6 F, and UA = 400,000 / 76.6, about 5,220 Btu/h per F.

Today, at the same measured 40 gpm: water 140 F in, 152 F out. Steam gauge reading corresponds to about 243 F saturation, with the control valve driven wide open.

Duty: 500 x 40 x 12 = 240,000 Btu/h. Gaps are 243 - 152 = 91 F and 243 - 140 = 103 F. LMTD = (103 - 91) / ln(103 / 91) = 12 / 0.1239 = 96.9 F, and UA = 240,000 / 96.9, about 2,480 Btu/h per F.

UA is at 47 percent of design while duty is at 60 percent, and the difference is the control system compensating: as transfer degraded, the valve opened, saturation temperature rose from 227 F to 243 F, and the larger driving difference recovered part of the capacity. A tech comparing only outlet temperatures sees a 12 F rise against a 20 F design and reads it as a moderate problem. The exchanger has actually lost more than half its conductance.

Which side? Water flow was measured at the equipment and matched design, so the tube-side velocity and its film are unchanged. That leaves tube-side fouling or something on the shell.

The deciding reading took 30 seconds. Shell surface, low on the vessel, read 228 F against a pressure implying 243 F. On a pure condensing vapor those two numbers agree closely. A 15 F gap says the vapor at the surface is not pure, which points at accumulated non-condensable gas rather than at anything on the water side. Confirmed at the shell's thermostatic vent, which was blocked.

Verifying, and reading the residual

After clearing the vent, re-measured with water flow re-confirmed at 40 gpm: steam settled back to 5 psig with the control valve modulating rather than pinned open, water 140 F in and 159 F out.

Duty: 500 x 40 x 19 = 380,000 Btu/h. Gaps 227 - 159 = 68 F and 227 - 140 = 87 F. LMTD = 19 / ln(87 / 68) = 19 / 0.2464 = 77.1 F, and UA = 380,000 / 77.1, about 4,930 Btu/h per F.

That is 94 percent of the design UA, and all three UA figures were built the same way, against a constant condensing temperature on the same vessel, so the comparison is like for like.

The remaining 6 percent is the number worth writing on the ticket. It is small, it is consistent with ordinary tube-side deposit after years in service, and it does not justify pulling a bundle today. What it does justify is recording it, so the next tech has a baseline that is honest about where the equipment actually stands rather than one that says "repaired, operating normally."

Where this changes how you spend the hour

Two practical conclusions follow that are not obvious from the physics alone.

Routine time goes to the single-phase side; the two-phase side gets trigger-based checks. The dominant resistance degrades gradually and predictably, so it earns scheduled attention: airflow, filters, fin condition, tube velocity, water treatment. The two-phase side does not degrade gradually. It steps, because its faults are events rather than accumulations: a trap fails, a vent blocks, a charge goes high, oil stops returning. Scheduling monthly inspections of the minority side is wasted labour, and having no trigger to check it after a capacity step is how a blocked vent lives in a building for two seasons.

Adding surface only helps on the dominant side. This is the design-side mirror of the same fact, and it explains equipment that looks strange until you see it. Fins exist on the air side because that is where the resistance is; putting the same fin area on the refrigerant side would change almost nothing. When somebody proposes an add-on to improve a two-phase exchanger, ask which resistance it reduces, and if the answer is the small one, it will not show up in the measurement.

References

• ASHRAE Handbook, Fundamentals volume, chapters on heat transfer, boiling and condensation, with tabulated film coefficient ranges
• Steam table data for saturation temperature and latent heat at the operating pressure, including how latent heat falls as pressure rises
• 29 CFR 1910.95 (occupational noise), 29 CFR 1910.333(b)(2) (electrical work), 29 CFR 1910.147 (mechanical isolation and stored energy); NFPA 70E-2021, 120.5 for live-dead-live proving
• See related: What a Phase Change Does to a Temperature Reading; What a Heat Exchanger Is Actually Doing; How Fouling Changes Heat Transfer