Why Approach Temperature Matters

Why this matters

Two techs stand at the same unit on the same day. One reads a pressure or a leaving temperature, compares it to a number in their head, and condemns a coil. The other reads the gap between two streams, compares it to what the same machine did at commissioning, and finds a fan blade that somebody replaced with the wrong part. The second tech is not smarter. They are using a number that cancels out the weather, and the first one is not.

Approach temperature is that number. It is the one field reading that separates "this exchanger can no longer transfer" from "this exchanger is not being fed," which are the two faults that produce identical customer complaints and completely different quotes.

What approach actually is

Approach is the temperature gap between one stream leaving an exchanger and the other stream entering it, measured at the end where the two are closest. It is a measure of how nearly the exchanger managed to bring one fluid to the other fluid's temperature.

It only ever goes to zero on an exchanger with infinite surface. Real approach is set by UA against the load, so it is a direct read on the transfer surface in a way that a single leaving temperature never is. Every trade already uses it under a local name.

The mechanics are identical in all five rows. Only the pair of temperatures changes.

Why it survives a change in the weather

Capacity moves with everything. Ambient, load, entering water, indoor humidity, time of day. That is why "it was making 6 degrees of split last August and 11 today" is a useless comparison unless you know both days.

Approach is written as a difference against the other side's entering condition, so the entering condition is already subtracted out. A condenser split of 20 F is 20 F whether it is 85 F or 95 F outside. That built-in normalization is the entire reason the number is worth carrying.

It is not perfectly load-independent, and pretending otherwise is how techs get burned. Approach widens as load rises, because more heat has to cross the same surface. So the comparison you make is against the same machine at a similar load and similar entering conditions, not against a remembered rule of thumb from a different building.

Where the other side is single-phase, there is a second correction. The entering-to-entering span between the two streams changes with the day, and a fixed approach against a wider span means something different than the same approach against a narrow one. Divide the approach by that span before comparing days. Where the other side is boiling or condensing, its temperature holds nearly constant through the exchanger, so there is no span to divide by and raw approach is already comparable. That distinction traces back to what a phase change does to a temperature reading, which a sibling card owns.

Before you take these readings

Approach requires readings in places that will hurt you, and the control is specific to each.

• Reaching into a condenser or air handler: the fan can start on a call while your arm is inside the guard. Lock and tag the unit disconnect for the mechanical and stored-energy hazard under 29 CFR 1910.147, and wait for the wheel to stop before crossing the guard line, not just for the motor to be de-energized.
• Metering at an energized board to confirm a fan is commanded on: electrical work is excluded from 1910.147 and governed by 29 CFR 1910.333(b)(2). Prove the meter live, prove the circuit, prove the meter again, the live-dead-live sequence in NFPA 70E-2021, 120.5.
• Connecting gauges to a refrigerant circuit: liquid refrigerant contacting skin causes an immediate freezing injury, so wear chemical splash goggles and gloves rated for cryogenic contact and crack the fitting away from your body. Refrigerant vapor is heavier than air and will pool in a pit, a crawlspace, or an equipment room floor and displace oxygen, so ventilate a low or enclosed space before working and do not enter one that has held a large release.
• Reading water temperatures on a hot circuit: heating water above 140 F scalds on contact. Read on the outside of the pipe with a strap-on or infrared instrument rather than opening a port, and do not install a well until the section is isolated, drained, and reading zero on a gauge.

The gate, stated in full

Unit of analysis: one exchanger, in a steady operating window of at least ten minutes, with both flows measured at the equipment rather than taken from a schedule.

Gate: the surface is the limit when today's approach exceeds the same machine's commissioning approach by more than 25 percent AND both measured flows are within 10 percent of their commissioning values. Both halves must hold. If either flow is outside 10 percent, the approach reading is not interpretable, so correct flow first and re-measure before applying the gate at all.

Normalizing step: where the opposite stream is single-phase, divide approach by the entering-to-entering span before comparing to commissioning. Where the opposite stream is saturated, compare raw approach.

After a correction: re-record approach as the new baseline only if it lands within 10 percent of commissioning. If it lands outside that, record the residual as a known deficit on the equipment record rather than adopting a degraded number as the new normal, which is how a shop loses its baseline one visit at a time.

Case one: the gate says surface

Chilled water coil in an air handler, complaint of warm air on hot afternoons. Commissioning data on file. Readings taken at matched entering wet-bulb (63 F both visits), which matters because on a wet coil the leaving dry-bulb shifts with humidity for reasons that have nothing to do with the coil.

Approach then: 55 - 44 = 11 F. Approach now: 62 - 44 = 18 F. Entering-to-entering span is 78 - 44 = 34 F on both days, so the normalized approach went from 0.32 to 0.53, a growth of about 64 percent. That clears the 25 percent half of the gate with room.

Flows: water at 95 percent of commissioning, air at 95 percent. Both inside the 10 percent band, so the second half holds too. Both halves true, so the gate fires: the surface is the limit.

Duty confirms it. Water side then, 500 x 10.0 x (56 - 44) = 60,000 Btu/h. Water side now, 500 x 9.5 x (52 - 44) = 38,000 Btu/h, which is 63 percent of commissioning. The air-side sensible check today, 1.08 x 1,900 x (78 - 62) = 32,832 Btu/h, comes in below the water-side total as it should on a wet coil, because the difference is the latent heat of the moisture condensing out. The two visits work out to sensible fractions of about 0.86 today against 0.83 at commissioning, close enough that nothing strange is happening on the moisture side.

Pull the coil face. On this job the finding was a gap where a filter rack had been modified, letting a share of the air past the coil face entirely. Bypass reads exactly like fouling on every number above, because both reduce effective surface.

Case two: same gate, opposite answer

Air-cooled condenser, complaint of high head pressure and short cycling. Commissioning record lists a 20 F split at 4,000 cfm of condenser air.

Today: entering air 95 F, saturated condensing temperature 130 F, so a split of 35 F. Against the recorded 20 F that is 75 percent growth, well past the first half of the gate. A tech who stops there quotes a condenser coil.

The second half is what saves the job. Condenser airflow measured 2,800 cfm, which is 30 percent below commissioning and far outside the 10 percent band. The gate does not fire. Flow gets corrected first and the approach is re-read afterward, because an approach taken at 70 percent of design airflow tells you nothing about the surface.

Fan rpm was at nameplate and motor amps were normal, which ruled out a slow-turning motor. The blade had been replaced during an earlier repair with one of a different pitch. After fitting the correct blade, airflow measured 3,900 cfm, within 2.5 percent of commissioning, and at the same 95 F entering air the saturated condensing temperature settled at 118 F, a split of 23 F. That is 15 percent above commissioning, under the 25 percent gate, so the coil is not the limit and no coil was sold.

Note what the approach did for you here that a raw pressure reading could not. The commissioning day was 85 F ambient and today was 95 F. Head pressure was always going to be higher today. The split subtracts the ambient out, which is why the same 35 F reading means the same thing on both days.

What the gate cannot decide

Approach narrows the question. It does not close it, and three specific conditions imitate a surface fault closely enough to fool the gate.

Refrigerant charge and non-condensables. Excess charge backs liquid up into the condenser and takes that surface out of service, which raises split exactly as fouling does. Non-condensable gas in the high side does the same and adds partial pressure on top. So read subcooling in the same breath as split and compare it to the equipment's own data: a wide split with normal subcooling points at the surface or the air, while a wide split with high subcooling points at charge or at a restricted liquid path. Neither reading alone is diagnostic.

A saturated-side temperature that is not actually saturated. On the refrigerant rows in the table above, the number you need is the saturation temperature corresponding to the measured pressure, not a tube surface temperature. Using a suction line surface reading in place of saturated evaporating temperature builds the superheat into your approach and inflates it. Read the pressure and convert.

Air-side approach on a wet coil compared against a dry day. Leaving air dry-bulb on a coil that is condensing moisture is a mixed number, part sensible and part the coil's surface condition. Compare at similar entering wet-bulb or the comparison is between two different machines.

There is a fourth limit that is not a fault at all. An approach can be perfectly healthy on an exchanger that is too small for the load it was given, because approach reports how well the surface is performing and not whether there is enough of it. That question belongs to sizing, and to the separate difference between rated capacity and delivered output.

References

• ASHRAE Handbook, Fundamentals volume, chapters on heat transfer and on psychrometrics for wet-coil analysis
• 29 CFR 1910.147 (mechanical isolation and stored energy) and 29 CFR 1910.333(b)(2) (electrical work); NFPA 70E-2021, 120.5 for live-dead-live proving
• Equipment manufacturer performance data for the design approach and expected subcooling at the rating point
• See related: What a Heat Exchanger Is Actually Doing; How Fouling Changes Heat Transfer; What a Phase Change Does to a Temperature Reading