Supply and Return Temperature Split Verification

Purpose

This standing instruction produces a temperature split that carries its entering conditions and its measurement points, and it treats that split as a screen that routes the call rather than as a verdict that authorizes a repair.

Split is the most-quoted and least-qualified number in residential service. It is quoted with no wet bulb beside it, taken at two points a tech chose for convenience, and then used to justify adding refrigerant. Every one of those three habits can move the number further than a real fault would. A shop that fixes the measurement first finds that the split usually exonerates the equipment and points somewhere else, which is exactly what a screen is supposed to do.

Safety actions that gate this work

• Power off at the disconnect and under your control before drilling any test port. The cabinet interior carries line voltage, a refrigerant line and, on a furnace, a gas train, and a bit that breaks through with a half inch of clearance behind it opens a suction line into the room.
• Probes go in through ports from outside the cabinet, never by reaching past an open blower inlet with the wheel turning. The same applies to retrieving a dropped probe: the disconnect goes open first.
• Where a split is taken in heating on a fired appliance, a personal CO monitor is on your body before the burner lights and stays on until you leave, and plenum metal near the exchanger will burn a hand, so use a probe with a handle and work through a port rather than an open seam.

Scope

Covers cooling-mode and heating-mode temperature split measurement on residential and small light-commercial forced-air equipment: instrument verification, entering condition measurement, probe placement, the return-side contamination check, and the comparison against the manufacturer's target.

Does not cover refrigerant charge verification, which is its own SOP and which this procedure may route to but never replaces. Does not cover the airflow measurement itself, owned by the airflow measurement and CFM verification SOP, or return leakage correction, owned by the return air deficiency and duct sealing SOPs. Heating temperature rise as an airflow calculation belongs to the airflow SOP; here it is used only against the nameplate range.

What a split can and cannot tell you

A cooling split is return dry bulb minus supply dry bulb, and its correct value is not a constant. It moves with the entering wet bulb, because wet bulb carries the latent load: the more moisture in the entering air, the more of the coil's capacity goes to condensing water rather than to dropping temperature, so the sensible split falls. A system running a low split in a humid house at start-up may be doing exactly what it should. That is why the target comes from the manufacturer's or a published target-split table indexed by entering wet bulb and entering dry bulb, and why a split reported without those two figures cannot be checked by anybody.

Two error sources beat any real fault the split is being used to find:

• Instrument offset. A fixed offset on one instrument cancels in a difference, leaving only that error as a percentage of the difference itself. Two different meters, each with its own offset, do not cancel; their bounds add. Use one instrument with a matched pair of probes, or verify the pair against each other in the same air and carry the difference through.
• Return-side gain. If the return picks up attic or crawlspace air between the room and the equipment, the return plenum reads warmer than the room. That raises the measured split while the house gets less cooling, so the number improves as the system gets worse. It is the one error that runs in the flattering direction, which is why step 5 exists.

Roles and handoffs

Procedure

1. Verify the probe pair against each other before either one goes in a duct. Acceptance: both probes in the same still air, agreeing within 0.5 F, with the difference recorded and carried, and preferably a single instrument with a matched pair so a common offset cancels in the difference. Wrong looks like two different meters pulled from two pouches. Stop rule: a pair that will not agree within 0.5 F is not used for a measurement whose whole content is a difference. Hazard: none physical; do it at the truck before the ladder comes off.

2. Measure the entering conditions in the occupied space, not at the equipment. Acceptance: dry bulb and either wet bulb or relative humidity taken at the return grille in the room, at the same minute, recorded together. Wrong looks like a split reported with no wet bulb, which cannot be compared against any target table. Stop rule: no entering conditions means no target, so the measurement stops. Hazard: none physical; keep the instrument off exterior walls and out of sun.

3. Run to steady state before reading anything. Acceptance: at least 15 minutes of continuous operation in the mode being measured, with readings changing by less than 0.5 F over the last 2 minutes. Wrong looks like a split read three minutes after the compressor started, when the coil is still pulling down. Stop rule: a system short-cycling before it reaches steady state is a different fault and gets diagnosed first. Hazard: the equipment is running, so keep tubing and sleeves clear of the blower inlet.

4. Place the supply probe where the air is mixed, and prove it is. Acceptance: three readings across the supply plenum with a spread under 2 F, averaged, taken downstream far enough that coil stratification has evened out. Wrong looks like a single reading taken at the coil face, where the air off one circuit can differ from the next by several degrees. Stop rule: a spread over 2 F means move downstream past the first turn and re-read rather than averaging a stratified sample. Hazard: probes go in through ports from outside the cabinet.

5. Check the return for contamination before you believe the split. Acceptance: return plenum temperature compared against the return grille temperature in the room, taken within the same few minutes, with any difference recorded in degrees F. Stop rule: a return plenum warmer than the room in cooling season means the return is gaining unconditioned air, the split is flattered by that amount, and the call routes to the return air deficiency and duct sealing SOPs before any charge work is considered. Hazard: tracing that return usually means an attic or crawlspace, so walk on structure only and carry two lights.

6. Compare against the manufacturer's target at the measured entering conditions, and route rather than diagnose. Acceptance: the target split read from the table at the measured entering wet bulb and dry bulb, the measured split stated against it with a shop band of plus or minus 3 F, and a routing call. Wrong looks like refrigerant added because the split came in low. Stop rule: a split outside the band routes to airflow verification first and to charge verification second, in that order, because low airflow and low charge both lower it and airflow is the cheaper thing to rule out. Hazard: none physical; arithmetic done clear of the open cabinet.

7. Restore, verify and record. Acceptance: ports plugged, probes retrieved, filter in with the arrow correct, panels on, thermostat returned to the customer's settings, and the blower-door interlock proven by opening the door on a running blower. Stop rule: a probe that cannot be accounted for is found before the cabinet closes, because a loose probe in a blower housing is a service call of its own. Hazard: this step puts power and air back with the family present, so restore at the disconnect with the door closed and your body out of the blower door line.

The record this produces

• Probe pair verification: both readings in the same air and the difference carried
• Entering dry bulb and wet bulb or relative humidity, taken at the return grille in the room, with the time
• Run time before the readings and the stability observed
• Three supply plenum readings, their spread, and the average
• Return plenum temperature against the room return grille temperature, and the difference
• Split as measured at the equipment and, where they differ, split referenced to room air
• Target split from the table at those entering conditions, and the band
• The routing call: accept, route to airflow, or route to charge

The charge tech reads the entering conditions before deciding whether a split even mattered. The advisor reads the return-side gain in degrees F as the evidence behind a duct scope. Months later a reader can tell whether a split changed because the system did, or because somebody probed a different point.

Worked pass: cooling call on a 3-ton system, customer reports weak cooling

Step 1: both probes on one instrument, reading 74.2 F and 74.6 F in the same still air, a difference of 0.4 F, inside the 0.5 F gate. Because both come from the same instrument, a common offset cancels in the difference and only the residual matters.

Step 2: at the hall return grille, 78.0 F dry bulb and 50 percent relative humidity, which the instrument returns as about 64.5 F wet bulb.

Step 3: 18 minutes of continuous run time, readings moving less than 0.5 F over the last 2 minutes.

Step 4 FAILS. Three points across the supply plenum read 57.8 F, 58.4 F and 62.9 F, a spread of 62.9 minus 57.8, or 5.1 F, well over the 2 F gate. Stop rule taken: the probe moves downstream past the first turn rather than averaging a stratified sample. Re-read there: 59.6 F, 60.1 F and 60.3 F, a spread of 0.7 F, averaging 60.0 F.

Step 5: the return plenum at the equipment reads 81.5 F while the hall return grille reads 78.0 F, so the return is gaining 81.5 minus 78.0, which is 3.5 F, between the room and the cabinet.

Now the split, computed both ways, because the two answers are different jobs. At the equipment it is 81.5 minus 60.0, or 21.5 F. Referenced to the room air the house actually starts with, it is 78.0 minus 60.0, or 18.0 F. A tech who probes only at the cabinet reports the 21.5 and is flattered by 3.5 F.

Step 6: the target split at 64.5 F entering wet bulb and 78 F entering dry bulb reads 20 F off the table, and the shop band of plus or minus 3 F makes the acceptance 17 to 23 F. Both 21.5 and 18.0 fall inside it, so the split does not indicate a charge fault and no refrigerant is added. The screen has done its job by clearing the equipment.

The finding is the return. At the same measured airflow, the ratio of the two splits is the ratio of the sensible heat actually reaching the house to the sensible heat the coil is producing, because the airflow and the constant are identical on both sides of the comparison. That is 18.0 divided by 21.5, about 0.84, so roughly 16 percent of the coil's sensible output is being spent re-cooling air the return picked up on its way to the cabinet. The call routes to the return air deficiency assessment SOP with that 3.5 F on the ticket.

Checking the pass against the sections above: the split is reported with its entering wet bulb and dry bulb attached, as the target-table section requires, and never quoted bare; the probe pair was verified in the same air first and both probes came from one instrument, which is the offset condition that section states; the failing step is step 4 and its stop rule moves the probe rather than averaging a stratified sample; and the return-gain check in step 5 caught the one error the section names as running in the flattering direction, with the split at the cabinet reading 3.5 F better than the house experiences.

References

• The equipment manufacturer's target-split or performance data indexed by entering wet bulb and entering dry bulb, which is the only valid target for a cooling split
• Equipment nameplate for the heating temperature-rise range, which is the acceptance when a split is taken on a fired appliance
• Instrument manufacturer's specification for the accuracy basis of each probe, including whether the figure is a fixed offset or an independent spread
• See related: the airflow measurement and CFM verification SOP, the return air deficiency assessment SOP, and the duct sealing and mastic application standard SOP