What Latent Heat Explains That Sensible Heat Cannot

Why this matters

A thermometer can only see one of the two ways a system absorbs heat. Everywhere a phase change is in progress, large amounts of energy move with the temperature standing perfectly still, and a tech reading that stalled number concludes the heat source has failed. The condemned part gets replaced, the same stall happens on the next cycle, and the customer has paid for a part that was never wrong. The same blind spot runs the other way on the air side, where a system removing a great deal of moisture shows almost nothing for it on a dry-bulb reading.

The two kinds of heat, and which one your instrument sees

Sensible heat changes temperature. It is what a thermometer, a thermistor and a thermal imager all respond to. Latent heat changes state at constant temperature: ice to water, water to vapor, refrigerant liquid to gas, and each of those in reverse.

Three water values anchor everything below, and they are published constants rather than field estimates:

• Warming liquid water: 1.0 Btu per lb per F
• Melting ice at 32 F: 144 Btu per lb, no temperature change at all
• Boiling water at atmospheric pressure: about 970 Btu per lb, no temperature change at all

Those numbers are the whole lesson in three lines. Melting one pound of ice costs the same energy as warming one pound of liquid water by 144 F. Boiling it off costs nearly seven times more again. A process moving that much energy with a flat temperature trace is not a broken process. It is the normal appearance of a phase change, and the flat trace is the evidence rather than the problem.

The stall that looked like a dead heater

A tech is sent to a unit that ices up and comes back iced. He wants to watch the defrost cycle himself rather than trust the history. The heater energizes, current draw holds steady at nameplate, and the coil surface temperature climbs to 32 F and then simply stops. Five minutes at 32 F. Ten minutes at 32 F. His read: the heater is weak, or the coil sensor is stuck.

Before he can measure anything: the circuit he wants to clamp is a live heater circuit inside a cabinet with wet metal in it. Water plus energized conductors is the specific hazard, so the clamp goes over the insulated conductor from outside the enclosure where the geometry allows it, and any work with hands inside the enclosure happens after the circuit is de-energized, locked out under 29 CFR 1910.333(b)(2), and proved dead with the live-dead-live sequence in NFPA 70E-2021, 120.5. The coil itself reaches burn temperature at the element and stays cold two inches away, so the surface gets a non-contact reading before a gloved hand goes near it, with hand protection selected under the hazard assessment at 29 CFR 1910.132(d)(1). Nobody opens a refrigerant-bearing path to look at the coil.

Now the arithmetic that kills his hypothesis. He scrapes and weighs the frost he can reach and estimates the coil is carrying roughly 2 lb of ice, entering at about 10 F. The heater nameplate, for this walkthrough, reads 400 W.

• Warming 2 lb of ice from 10 F to 32 F, using about 0.5 Btu per lb per F for ice: 2 x 0.5 x 22 = 22 Btu
• Melting that 2 lb at 32 F: 2 x 144 = 288 Btu
• Heater output: 400 W x 3.412 = about 1,365 Btu per hr

The warming step, at perfect coupling, takes 22 / 1,365 = 0.016 hr, about 1 minute. That is the climb he watched and accepted. The melting step takes 288 / 1,365 = 0.211 hr, about 12.7 minutes, and that is with every Btu landing in the ice. In a real cabinet a large share of the element's output goes into the coil metal, the surrounding air and the cabinet structure instead, so the observed stall runs longer than 12.7 minutes rather than shorter. A ten-minute plateau at 32 F is not evidence of a weak heater. It is roughly what the phase change costs.

What was actually wrong. The stall was normal and the icing was not. The finding that mattered came from the other end of the cycle: the coil was accumulating that much ice again within a day, which points at a moisture source, a door seal, a drain that is not clearing, or a defrost interval that is too long for the actual load, none of which are the heater. Replacing the heater would have restored the identical behaviour at the identical rate.

The failure mode in one line: treating a flat temperature as proof that no heat is flowing. During any phase change, flat is exactly what full heat flow looks like.

Why the temperature then moves so fast

Once the last of the ice is gone, that same heater is pushing into liquid water and metal with nothing left to melt. Warming the 2 lb of melt water from 32 F to 40 F takes 2 x 1.0 x 8 = 16 Btu, which at 1,365 Btu per hr is 0.012 hr, about 0.7 minutes.

Set the two beside each other. The phase change cost 288 Btu and held the reading still for the better part of a quarter hour. The 8 F rise that followed cost 16 Btu, an eighteenth as much, and happened in under a minute. Nothing about the heat source changed between those two observations. Only what the heat was being spent on changed.

That ratio is the single most useful thing to carry out of this article, because it inverts the instinct. A temperature that suddenly starts moving fast usually means a phase change just finished, not that something just got stronger. Techs read the fast climb as the heater "finally kicking in" and write it up that way, which cements the wrong story in the service history for the next person.

The same blindness on the air side

Air carries latent heat as water vapor, and a dry-bulb thermometer is blind to all of it. Two constants make the split measurable with instruments most trucks already carry, both for standard air near sea level:

• Sensible only: 1.08 x CFM x dry-bulb difference = Btu per hr
• Total, sensible plus latent: 4.5 x CFM x enthalpy difference in Btu per lb = Btu per hr

Enthalpy tracks wet-bulb temperature closely, which is why the wet-bulb reading is the one that sees moisture work.

Work a cooling coil at 1,200 CFM. Return air reads 78 F dry bulb, 65 F wet bulb; supply air reads 60 F dry bulb, 57 F wet bulb. From a psychrometric chart, 65 F wet bulb is about 30.1 Btu per lb of dry air and 57 F wet bulb is about 24.5 Btu per lb.

• Dry-bulb split: 78 - 60 = 18 F
• Sensible: 1.08 x 1,200 x 18 = about 23,300 Btu per hr
• Total: 4.5 x 1,200 x (30.1 - 24.5) = 4.5 x 1,200 x 5.6 = about 30,200 Btu per hr
• Latent, by subtraction: 30,200 - 23,300 = about 6,900 Btu per hr
• Sensible heat ratio: 23,300 / 30,200 = about 0.77

A tech carrying only a dry-bulb reading sees the 18 F split and books 23,300 Btu per hr. He is missing about 6,900 Btu per hr, roughly 23 percent of what the coil is actually doing, all of it wringing water out of the air. Both figures above come from the same coil at the same moment; the sensible figure is not wrong, it is incomplete, and it is incomplete in the direction that makes equipment look worse than it is.

Which way that matters depends on the complaint. On a "not cold enough" call, the missing latent capacity explains why the equipment is working harder than the dry-bulb split suggests. On a "clammy house" call, the same split is what tells you whether the coil is doing enough moisture work or has been pushed toward all-sensible operation by excess airflow. The commonly cited residential design band for sensible heat ratio runs roughly 0.70 to 0.80, and where a specific coil should sit is on its own performance data at the airflow and entering conditions you actually measured, not on that band.

When a temperature reading is really a pressure reading

There is a second consequence of a flat temperature, and it is the one that carries across to pressurized systems. A fluid sitting at its boiling point is at a temperature set by its pressure, not by its load. Raise the pressure and the boiling temperature rises with it; drop the pressure and it falls.

So inside any saturated region, a temperature reading has stopped being a load measurement and become a pressure measurement in disguise. Two practical consequences:

• Watching a saturated temperature and reasoning about capacity from it is reasoning about pressure. The load information is in the mass flow and the phase-change enthalpy, not in the number on the screen.
• Anywhere a fluid enters saturation earlier than expected, the temperature stops responding to the thing you were trying to observe. Steam and refrigerant systems both do this, and it is why pressure and temperature get read as a pair rather than one standing in for the other.

Nothing here justifies opening a pressurized or refrigerant-bearing path to investigate. Those readings come from installed ports or from surface temperature plus a gauge, and any isolation of a pressurized boundary happens under 29 CFR 1910.147 with the stored energy relieved and verified at the point of opening.

What to write down so the next tech is not fooled

The notes that make a phase-change stall readable later are not temperature notes:

1. Time-stamp the plateau, do not just record its temperature. "32 F for 11 minutes, heater energized throughout, current steady" is diagnostic. "Coil 32 F" is not, and it is what most tickets contain.
2. Record a mass or a quantity whenever a phase change is involved. Pounds of ice, cups of condensate in a measured interval, inches of accumulation. Latent energy is proportional to mass, so without a mass there is no way to check whether the observed time was reasonable.
3. Pair every dry-bulb reading on a wet coil with a wet-bulb reading. One without the other is a partial measurement of a process that has two parts, and the missing part is the one nobody can reconstruct afterward.
4. Note whether the reading was taken inside a saturated region. If it was, label it as a saturation temperature so the next reader does not treat it as a load number.

References

• ASHRAE Handbook Fundamentals, psychrometrics, enthalpy of moist air and phase-change properties of water
• 29 CFR 1910.333(b)(2), de-energized electrical work; NFPA 70E-2021, 120.5, live-dead-live verification
• 29 CFR 1910.147, control of hazardous energy including stored pressure and thermal fluid
• 29 CFR 1910.132(d)(1), PPE hazard assessment for hot-surface contact
• See related: How to Read a Pressure-Temperature Chart; How Heat Actually Moves and Why It Matters on a Call