You get the call in July: condenser hums or sits dead, indoor blower runs fine, house is climbing. You swap the run cap because that’s the $15 part that fixes 70% of these, and… nothing. Or it runs for the rest of the day and you’re back on Thursday.

The trap is that a capacitor can measure perfectly fine on your meter and still not be the whole story — and even when the cap is genuinely good, there are half a dozen things downstream that produce the identical symptom. Here’s the order I actually work it, because skipping steps is how you end up condemning a compressor that was fine or replacing a cap twice.

First: is the static µF reading even the right test?

When you put a capacitor on a meter’s capacitance setting, you’re measuring it at rest, with no load. A cap that reads 43.5 µF on a 45 µF (±5%) label looks fine sitting on your tailgate. The problem is that capacitors fail under load and under heat — the values can drift once it’s sitting at 150°F on top of a running condenser and pushing start current.

So “the cap tests good” usually means “the cap tested good cold and static.” That’s necessary, not sufficient. If everything else checks out and you’re still chasing a no-start, a cap that’s marginal under load is back on the table — measure the actual run current and voltage on the start winding with the unit energized if you can do it safely, not just the bench µF.

But before you go deep on the cap, rule out the cheaper, more common culprits.

The order I work a no-start condenser

1. Confirm incoming power — 240V at the contactor line side. Sounds obvious. It catches more calls than anyone admits: a tripped breaker that looks set, a blown fuse in the disconnect, a chewed whip. Two minutes with the meter on L1-L2 before you touch anything else.

2. Is the contactor actually pulling in? Check for 24V on the contactor coil when there’s a call for cooling. No 24V means your problem is upstream — thermostat, transformer, a tripped float switch, a high/low pressure lockout. If you have 24V at the coil and the contactor isn’t closing, the coil’s bad. If it’s closing but the contacts are pitted and burned, you’ve got voltage on the line side and garbage on the load side — a weak contactor mimics a no-start perfectly, and people blame the cap for it constantly. Measure load-side voltage with it pulled in.

3. Now check the capacitor — and check it right. Static µF first; if it’s out of tolerance, done. If it’s in spec, don’t fully clear it yet — note it and keep it as a suspect if the rest checks out. Visual matters too: bulged top, leaked oil, rust streaks down the side all condemn it regardless of what the meter says.

4. Test the compressor windings. This is the step people skip because they don’t want the answer. Power off, leads off, ohm it out:

• Common-to-Run and Common-to-Start should read low resistance values consistent with the compressor.
• Start-to-Run should equal C-S plus C-R, roughly.
• Any winding to the case/ground should read open — infinite. A reading to ground means the compressor is grounded and it’s done.
• Open winding (infinite where you should read continuity) means an internal break — also done.

5. Internal overload. If the windings read open and the compressor is hot, you may be reading a tripped internal thermal overload, not a dead winding. Let it cool — sometimes an hour — and re-check. A compressor that’s been short-cycling or running low on charge can trip its overload and present as a dead no-start until it cools and resets. If it resets and runs, your real problem is whatever overheated it (low charge, airflow, a failing cap pushing high amps).

The calls that come back

Almost every “I replaced the cap and it died again” callback I’ve seen traces to one of these:

• The contactor was the actual problem. Pitted contacts dropping voltage to the compressor. The new cap masked it for a day.
• The compressor was already failing. Hard-starting on its way to a grounded or open winding. The cap swap bought a few cycles. A hard-start kit can buy a little more, but you’re managing a death, not curing it — tell the customer that plainly.
• Low charge / airflow cooking the compressor into repeated overload trips. Fix the cap all you want; if it’s tripping on heat, it’ll keep tripping.
• A genuinely load-marginal cap that passed static µF. Rare, but real — which is why it stays on the list instead of getting cleared at step three.

The point

A run capacitor reading in-spec is a data point, not a diagnosis. Work the condenser in order — power, contactor, capacitor under scrutiny, windings, overload — and the no-start tells you what it actually is instead of what the cheapest part on the truck wants it to be. The fifteen-dollar cap is the right first guess. It’s a bad place to stop.

— Cal