Defrost Control Diagnosis on a Heat Pump

Purpose

This standing instruction guarantees that an iced outdoor coil is resolved to either a defrost fault or a refrigerant-side fault before any part is ordered, and that either way the shop keeps a record of what the coil looked like and what the numbers were.

An iced coil is not evidence that defrost is broken. A heat pump low on charge, or one whose outdoor coil is packed internally with debris, runs its coil far colder than the weather calls for and lays ice down faster than a healthy cycle can remove it. Boards get replaced on those units every winter and the ice returns the following week. Defrost also has two halves that fail in opposite directions: a system that never initiates ices solid, and one that never terminates runs nuisance cycles the customer reports as blowing cold.

Safety actions that gate the work

• Open and lock the outdoor disconnect and prove dead before any cover comes off, proved on a known live source before and after per NFPA 70E-2021, 120.5, with work practices at 29 CFR 1910.333(b)(2), written for qualified persons under 1910.332. If nobody on site is qualified for the energized measurements this procedure requires, that half goes to someone who is.
• The run capacitor holds a charge after the disconnect opens. Discharge it through a resistor before your hands are near the terminals.
• Do not chip, pry or hammer ice off a coil. A punctured tube turns a diagnostic into a refrigerant release and a fin-deep laceration. Melt water or warm air only, and clear the iced pad before you set up.
• Liquid refrigerant blinds on contact. Glasses and gloves on every hose connection, and 40 CFR Part 82 Subpart F governs every connection made and broken.

Scope

Covers defrost diagnosis on residential and light-commercial air-source heat pumps, split and packaged, with time-temperature or demand boards, from the as-found ice pattern to the decision to replace a component or not.

Does not cover charge correction, which the charge verification SOP owns and which this procedure only rules in or out, or the reversing valve as a component, which has its own SOP. On inverter-driven ductless equipment with no field-adjustable initiation setting, this procedure stops after the refrigerant-side rule-out and the fault code goes to the manufacturer's technical line.

Roles and handoffs

Procedure

1. Photograph the ice from four sides before you touch anything, and record outdoor dry bulb and the weather. The ice is the evidence and it melts while you work. Acceptance: four images with coil face and base pan both visible, plus a recorded ambient. Wrong looks like starting the unit to see what it does and destroying the pattern. Light uniform frost is normal between cycles; solid clear ice in the base pan rising into the coil means defrost runs but melt water is not leaving; frost on one band only is airflow or circuiting, not control. Hazard: the pad is iced, so set your feet before framing a shot.

2. Rule out the refrigerant side first, by comparing saturated suction temperature to outdoor ambient with the unit running in heating. On a clean coil with the fan running and no more than light frost, a correctly charged system holds the outdoor coil roughly 10 to 20 F below ambient; that is a shop rule of thumb under those stated conditions, and the manufacturer's charging chart is the authority. Acceptance: recorded ambient, recorded saturated suction, split inside that band. Wrong looks like a split wider than 20 F: the coil is colder than the weather explains, which is low charge or restricted outdoor airflow, and either lays ice down faster than a working cycle clears it. Stop rule: a wide split ends the defrost investigation here and sends the ticket to charge and airflow. Hazard: hoses onto a running system, so glasses and gloves before the first connection, and hands and hose ends outside the fan plane.

3. Identify the control type and have its literature in hand before you probe it. A time-temperature board initiates on accumulated compressor run time set by a jumper, gated by a coil sensor; a demand board initiates on coil-to-ambient difference. Acceptance: board part number and its published sensor open and close points, or its resistance curve, on the worksheet. Wrong looks like working from a remembered set point, which varies by part number and sends you to the wrong conclusion. Hazard: none at this step, it is reading done away from the equipment.

4. Verify the sensor's mounting before you verify the sensor. It belongs in its clip, in full metal contact, on the tube the diagram names. Acceptance: clip seated, tube confirmed against the diagram, and a resistance reading at a measured temperature matching the published curve. Wrong looks like a sensor hanging free: in heating the coil runs colder than the air around it, so a loose sensor reads warm, never closes, and the unit never initiates. Stop rule: reseat and re-observe before condemning anything. Hazard: reaching between fins cuts to the bone, so cut-resistant gloves, disconnect open and proved dead before your hand goes in.

5. Force a defrost and watch for four events, all of which must happen. Use the board's test position per its literature, never an improvised jumper. The four: the reversing valve shifts audibly and the line to the indoor coil goes hot, the outdoor fan stops, the compressor keeps running, and auxiliary heat energizes where the diagram calls for it. Acceptance: all four observed and recorded individually. Wrong looks like an outdoor fan that keeps turning, holding cold air on the coil so it never warms and never terminates on temperature. Hazard: energized work in an open cabinet with compressor and fan live, so stay outside the fan plane for the whole cycle. The command creates its own: forcing defrost puts the indoor coil into cooling with the blower running, so confirm auxiliary heat is available first in hard cold, and do not chain cycles on a house with plumbing in exterior walls.

6. Time the cycle and record what ended it. Acceptance: elapsed minutes from initiation to termination, the cause named as sensor or time limit, and a visibly clear coil face at the end. Wrong looks like a cycle that always runs to the published limit, meaning the sensor never satisfied and the coil never reached its open point: a sensor, a mounting, or a coil not receiving hot gas. Stop rule: a cycle ending with ice still on the face does not get a second forced cycle to help it along; go back to step 2. Hazard: an open energized cabinet for the whole cycle, so keep the leads in one work area and stay clear of the fan.

7. Put it back, and prove the sequence you disturbed still works. Close the panels, restore the disconnect, and watch one full return to heating: valve back to its heating position, outdoor fan restarted, auxiliary dropped out, supply air above return at the equipment. Acceptance: all four observed after restoration, with a supply and return pair recorded. Wrong looks like leaving at the moment power comes back; a valve left mid-shift is a fault you introduced and the customer finds it at 2 a.m. Hazard: this puts energy back with the customer in the house, so stand to the hinge side of the disconnect rather than square in front, and confirm nobody is at the indoor unit.

8. Write the finding and the disposition, including a negative one. Acceptance: worksheet complete, disposition stated as part replaced, charge or airflow referred, or no fault found with the conditions tested under. Wrong looks like closing a ticket on a unit that was not iced on arrival and writing nothing. Hazard: none at this step, it is desk work, but a thin record here buys the third truck roll.

The record this produces

A defrost worksheet filed to the equipment, not to the ticket:

• The four as-found photographs, outdoor dry bulb, and the weather
• Saturated suction, outdoor ambient, and the split between them
• Board part number and its published sensor open and close points
• Sensor mounting confirmed, and the resistance reading against the curve at a measured temperature
• The four forced-defrost events, each recorded as observed or not
• Elapsed cycle time and what terminated it
• The step 7 restoration observations and the supply and return pair
• Disposition, including negative findings and the conditions tested under

The next tech reads the split from the last visit before opening anything, which is what stops a shop replacing the same board twice. The manager reads the disposition column across a season: boards replaced on units whose splits were never recorded is a training problem, not a parts one.

Worked pass: 34 F, light rain, coil iced, two-year-old split heat pump

Step 1: four photographs. Solid clear ice in the base pan rising into the bottom of the coil, heavy frost above rather than light. Ambient 34 F, light rain.

Step 2 FAILS. Running in heating, saturated suction reads 8 F against the recorded 34 F ambient, so the split is 34 minus 8, which leaves 26 F. The acceptance is a split inside 10 to 20 F under the stated conditions, and 26 F sits outside it on the wide side. Stop rule taken: the defrost investigation stops here.

The rule-out is worked rather than assumed. Charge checks correct against the installed weight record, which leaves outdoor airflow. The face is clean from outside, but a light behind the coil shows the second row packed with cottonwood seed between the fin rows. Coil washed from the inside out. Re-measured on the same instrument at the same ports: saturated suction 20 F against the same 34 F ambient, so the split is 34 minus 20, which leaves 14 F, inside the band. Step 2 now passes.

Steps 3 and 4: board identified as time-temperature, literature pulled, published sensor points recorded; sensor seated in its clip on the tube the diagram names, resistance consistent with the published curve at the measured coil temperature.

Step 5: forced defrost from the board's test position. All four events observed: valve shifts with an audible clunk and the line to the indoor coil goes hot, outdoor fan stops, compressor keeps running, auxiliary energizes as the diagram calls for.

Steps 6 to 8: terminates on the sensor at 6 minutes against a published 10 minute limit, face clear. Panels closed, disconnect restored from the hinge side; valve returns to heating, fan restarts, auxiliary drops out, supply measures 96 F against a 68 F return at the equipment. Disposition recorded as outdoor airflow restriction, defrost verified functional, no parts replaced.

Checking the pass against the sections above: step 2 accepts a split inside 10 to 20 F, and both the failing 26 F and the passing 14 F print their inputs and the subtraction, taken at the same ports on the same instrument. Step 6 accepts an elapsed time, a termination cause and a clear face, and the run reports 6 minutes, sensor termination, face clear. That band is stated under the condition of a clean coil, and the failing reading was taken on a coil that was not clean, which is why the finding is the coil rather than the reading.

References

• The defrost board manufacturer's literature for that part number, the only authority on sensor open and close points, jumper-selected initiation intervals and the published time limit
• Equipment manufacturer's wiring diagram and sequence of operation, for whether auxiliary heat energizes during defrost on that model
• 29 CFR 1910.333(b)(2) and NFPA 70E-2021, 120.5, for de-energizing and proving dead
• 40 CFR Part 82 Subpart F, for refrigerant handling at the step 2 hose connections
• See related: charge verification by weight and superheat, reversing valve diagnosis and replacement, and the defrost board initiates too often in mild weather tree