INTRO (0:00 to 0:30)
ON CAMERA, Darrel at the truck, tool bag in hand, training house behind him
This call came in as "blowing but not cold." That phrase is a clue, not a diagnosis, and the difference between those two words is this entire video. I am going to run the whole IB diagnostic flow on this system, out loud, every decision narrated, door to driveway. Watch what I do, but more important, listen to what I refuse to do until I have a number.
MAIN (0:30 to 10:15)
Beat 1: Intake at the door (0:30 to 1:45)
ON CAMERA, at the front door with the homeowner; Darrel listens first, tablet in hand
First sixty seconds belong to her. Listen to the exact words: "It runs all day but the house keeps getting warmer, and yesterday afternoon it just seemed to give up." I write that down word for word, because I will have to answer it word for word at the end of this call.
CLOSEUP: tablet, the six intake questions being filled in
Now my six questions. What is it doing, she just told me. When did it start: about a week ago, gradually, then worse yesterday. What changed: there was a dust storm last weekend. Does it ever work: mornings are better. Sounds, smells, ice, water, trips: she heard a hum and a click outside yesterday. Anyone worked on it: not in two years.
ON CAMERA, walking toward the backyard
Before I touch a single panel, listen to the picture those answers paint. Gradual decline says something is fouling or fading, not snapped. Worse in afternoon heat says heat-sensitive, and that usually means electrical. Hum and click outside is the classic capacitor presentation. And a dust storm last weekend is a suspect with a motive. I have a starting bias: outdoor electrical, with the storm involved. A bias, not a verdict. Now I survey the whole machine anyway, because the system loves to hide the cause one domain over from the symptom.
Beat 2: The survey, no tools yet (1:45 to 3:15)
AT THE THERMOSTAT, closeup of display
Survey starts inside. Thermostat: cooling call active, set 76, room reading 84. The brain is asking; something downstream is not answering.
CLOSEUP: filter pulled from the return grille
Filter: moderately loaded, not collapsed. Goes back in for now, noted; if I were taking performance readings today they would need a fresh filter first, and that note matters later.
AT THE AIR HANDLER, hand at a supply register, then flashlight on the coil access
Blower is running, good airflow at the register but the air is room temperature. Coil access: no ice, pan dry, no oil staining. So indoor side is alive and the refrigerant circuit is not doing any work. That moves my suspicion outside.
AT THE CONDENSER, wide shot, Darrel circling the unit slowly
Outside, and I have not opened anything yet. Eyes and ears first. The fan is not spinning. Listen, there it is: hum, a few seconds, then a click. Something is trying to start and giving up. And look at this coil, walk around with me: the house-facing side, the side nobody ever looks at, is matted with storm dust. Filthy. Remember that. It is about to matter more than anything else on this call.
ON CAMERA, crouched by the condenser
Survey verdict: indoor side healthy, outdoor unit failing to start, hum and click, dirty coil, after a dust storm, worse in heat. Every arrow points at the start circuit, and the number one resident of that neighborhood is the run capacitor, one in five of all service calls. That is my hypothesis forming. But a hum and click is also what a locked rotor sounds like, and I do not condemn parts, or compressors, on sounds. Time for numbers.
Beat 3: Measurement (3:15 to 5:00)
CLOSEUP: disconnect pulled, meter proving dead
Disconnect pulled, meter proves dead before my hands go in. Every time, no exceptions.
CLOSEUP: bleed resistor across the capacitor, then wiring photographed
Capacitor: discharge through the bleed resistor, photograph the wiring before a single wire moves, wires off.
CLOSEUP: meter in capacitance mode, reading legible on screen
Here is the moment the whole call turns on. Rated 45 microfarads on the HERM section. The minus 6 percent rule puts the floor at 42.3. Meter says 33.6. That is not marginal, that is condemned, and now I have what this track calls a condemning reading: a measured value, against a published spec, failed. The photo of this meter on this part goes in ServiceTitan before the new part goes in the system. No reading, no replacement. Today there is a reading.
ON CAMERA, holding the failed capacitor
Notice what I did not do. The gauges are sitting on the truck shelf, and they are staying there. "Blowing but not cold" is the phrase that launches a thousand top-offs, and on this call the refrigerant circuit was never the patient. If I had connected gauges to a system whose condenser fan cannot even start, I would have read garbage pressures and maybe "found" a charge problem that does not exist. That is failure pattern one, misreading the charge, and the way you avoid it is exactly this: survey and electrical first, refrigerant judged only on a system that can actually run. D24 goes deep on that.
Beat 4: Hypothesis, written, and root cause (5:00 to 6:30)
CLOSEUP: tablet, Darrel typing the hypothesis line
The hypothesis goes in the record, in a full sentence: failed HERM section run capacitor, suspected root cause heat stress from fouled condenser coil plus storm exposure. Why the second half of that sentence? Because a capacitor this far gone, this young, usually had help, and I am looking at the help: a coil matted with dust, driving head pressure and cabinet heat up all summer, plus a thunderstorm on the right date.
ON CAMERA, hand on the dirty coil face
This is the difference between a parts swapper and a diagnostician, and it is failure pattern two from your article. The swap fixes the part. The diagnosis fixes the killer. If I replace this capacitor and leave this coil, the new part bakes behind the same blanket, and I have scheduled my own callback for August. So the repair on this call is two items, both documented, both with the customer's approval: the condemned capacitor, and the coil cleaning that removes the thing that killed it.
CLOSEUP: contactor points, then wire lugs
While the panel is open and dead: contactor points, light wear, passes. Lugs tight, insulation sound, one photo each. I check the neighborhood whenever a part dies, because whatever killed one component had access to all of them.
Beat 5: Repair (6:30 to 7:45)
CLOSEUP: new capacitor installed, wiring matched to the earlier photo
New capacitor, rated match for the original, wired exactly per the photo I took before teardown. Bench-checked before install: 44.8 against 45, healthy.
WIDE: coil rinse, water visibly flowing inside-out, then cleaner applied
The root cause gets fixed C21-style: inside-out rinse, opposite the airflow that packed this dust in, non-acid foam on the matted face, full rinse until the water runs clean. Before and after photos, both.
ON CAMERA, wiping hands
Notice the order of operations on this whole call. Customer's words aimed the survey. Survey aimed the measurement. The measurement, not the hum, not the click, not my twenty years of pattern matching, condemned the part. And the condemned part pointed back at its killer. Symptom in the electrical domain, root cause in the airflow domain. That is system-first thinking doing its job.
Beat 6: Verification (7:45 to 9:15)
CLOSEUP: disconnect restored; the condenser fan spins up; Darrel lets it run
Power restored. Fan spins, compressor pulls in, and now the system gets what every system gets before final numbers: ten to fifteen minutes of stabilized runtime. "It started" is not verification. Numbers are verification, and they have to be the same kind of numbers that defined the fault.
CLOSEUP: clamp meter on compressor common
Compressor amps: 13.8 against a nameplate RLA of 18.9. Healthy margin, recorded.
CLOSEUP: probe thermometers in return and supply
Temperature split after stabilization: return 84.2, supply 64.6, that is 19.6, inside the 18 to 22 target, and remember the loaded filter from the survey: it gets noted and replaced so this split stays honest.
CLOSEUP: probe on suction line
Suction line at the condenser: 54 degrees, cold-drink cold. Liquid line warm, not hot, now that the coil can breathe. The refrigerant circuit just proved itself with temperatures, and the gauges never left the truck. A healthy sealed system keeps its charge and its Schrader cores.
ON CAMERA
And verification answers her words, not just my checklist. She said it runs all day and gives up in the afternoon heat. The part that quit in heat is replaced, the thing that was cooking it is clean, and the readings on this hot afternoon are inside spec. Her symptom, answered on her terms.
Beat 7: Documentation and the doorway (9:15 to 10:15)
CLOSEUP: tablet, readings entered, photo checklist filling in
Nothing on this call exists until it is in ServiceTitan: her words at intake, the survey notes, the written hypothesis, the condemning reading with the meter photo, 33.6 against a 42.3 floor, the root cause finding with the dirty coil photo, the repair, and the verification numbers. Photos run to the IB 8-photo close-out standard; D30 walks that standard photo by photo, so here I will just say: if it was measured, it is in the record, and if it was condemned, there is a picture of why.
ON CAMERA, at the door with the homeowner
Doorway summary, findings first, photos doing the talking: here is the part that failed and the reading that proves it, here is what was killing it, here is both of them fixed, and here are the numbers showing the system healthy right now. No jargon, no fear, just the evidence trail.
OUTRO (10:15 to 11:00)
ON CAMERA, Darrel at the truck, holding the dead capacitor
One call, eight steps, zero guesses. The intake gave me a bias, the survey gave me a direction, and only the meter gave me a verdict. This part was guilty, and I can prove it; the coil was the accomplice, and I can prove that too. That is the whole Diagnostics track in one driveway. Next up is D23, where the staged faults get sneakier and the electrical tests get deeper. Bring this process with you. It is the only tool that works on every call.