Why Does My AC Compressor Shut Off After 2-3 Minutes? (And a 7-Point Troubleshooting Checklist I Wish I'd Had in 2019)

If your AC compressor runs for two or three minutes and then shuts off, that's not a random glitch. It's a symptom. And if you're a facility manager or HVAC contractor, you already know the feeling: you stand there watching the unit cycle, hoping it'll catch, knowing it won't. I've been that person too many times.

I'm Brian. I've been handling replacement orders and troubleshooting support for commercial HVAC equipment for about six years now. In my first year, I misdiagnosed a short-cycling compressor on a water source heat pump, ordered a $1,800 replacement compressor, and it turned out the issue was a $40 contactor. That mistake still shows up in our team's monthly review as a lesson. (Yes, we document our mistakes that thoroughly now.) So this article is the checklist I wish someone had handed me back then.

Before We Start: What This Checklist Is For

This guide is for anyone dealing with a compressor that starts, runs for a few minutes, then kicks off — repeatedly. It covers the most common causes on commercial units, including AAON water source heat pumps, chillers, and packaged HVAC units. If you're working on a residential system, the basic electrical checks still apply, but some parts (like the TXV operation) will differ slightly.

There are 7 steps below. They're in order of least invasive to most involved. Do them all before you call a parts distributor or order a new compressor. Yes, all seven. The thing that saves you is following the sequence, not skipping to step five because you're confident it's the expansion valve. (Ask me how I know. Actually, don't — I'll tell you anyway in step 4.)

Step 1: Check the Airflow Across the Evaporator Coil

Low airflow makes the evaporator coil get too cold. When the coil temperature drops low enough, the low-pressure switch (or the freeze stat) shuts the compressor off to prevent liquid slugging or ice formation. It looks like an electrical failure. It's often just a filter or a belt.

Check these first:

• Return air filters — are they clogged? Semantics matter: measure static pressure drop across the filter, don't just eyeball it. (We had a "lightly dusty" filter that was pulling 0.7" W.C. on an AAON unit rated for 0.5 max. That alone caused short cycling.)
• Supply air ducts — any closed dampers or blocked diffusers? (A damper closed during construction caused a 3-day headache once.)
• Indoor blower — is the belt slipping? Is the motor drawing correct amps?

Checkpoint: If airflow is restored, the compressor should stay on longer. If it still shuts off after 2-3 minutes, go to step 2.

Step 2: Test the Low-Pressure (Suction) and High-Pressure (Discharge) Switches

These safety switches are there to protect the compressor from operating outside its envelope. But they can fail — sticking closed or opening too early. A faulty low-pressure switch can make a perfectly good compressor look broken.

Hook up your manifold gauges and watch the pressures while the unit runs. If the suction pressure drops below the switch's cut-out point right before the compressor shuts off, you have two possible causes: the switch is bad, or the system genuinely has low suction pressure (which could be a refrigerant issue or a restricted metering device — later steps).

To isolate: momentarily jumper the switch (only for testing!) and see if the compressor runs past the 2-3 minute mark. If it does, the switch is suspect. Replace it with a factory-rated switch from an AAON parts distributor or your component supplier. Don't use a generic "universal" low-pressure switch unless it matches the original set points. This is one of those places where using a slightly different cut-in/cut-out can cause all sorts of nuisance trips later. (We learned that on a chiller — the wrong switch set point caused a 40-minute call, and the unit still short-cycled on high head pressure.)

Checkpoint: Pressures at time of shutdown? Note them. You'll need these numbers in step 5.

Step 3: Verify the Condenser Coil Is Clean (Fan Running Correctly?)

A dirty condenser coil reduces heat rejection. The head pressure climbs, and the high-pressure switch opens. Or the condenser fan isn't running at full speed, which also causes high head pressure. Interestingly, a condenser coil that's too clean isn't usually the problem, but I have seen a fan blade spinning backward on a phase-monitored 3-phase unit once. (It's rare, but it happens — check rotation on any new motor install.)

On an AAON unit, this is usually simple to inspect: the condenser coil is accessible from the exterior. Spray it down with a coil cleaner, making sure the fins are completely clear. Check for bent fins blocking airflow (a fin comb fixes them).

I remember a call in August 2023 where someone had already ordered a new compressor from us. The replacement cost was around $2,400 for that 25-ton unit. We went through their diagnostics and they hadn't cleaned the condenser coil, because "it looked okay from the outside." I asked them to remove the top panel. There was a 3-inch layer of cottonwood fluff sandwiched between the coil and the shroud. Cleaned the coil, unit ran fine. The replacement compressor was canceled. (Lucky for them — and for me, because a vendor restocking fee is a paperwork nightmare.)

Checkpoint: Condenser inlet/outlet temperature difference. A well-functioning condenser with a clean coil will show a consistent temperature drop across the fins. If the difference is minimal and the coil is clean, move on.

Step 4: Check the Crankcase Heater and the Belief That "Compressor's Fine Because It Runs"

This step is the one most people skip. (It's the step I ordered that compressor on in my first year.) The issue isn't just about short cycling — it's about what's happening before the compressor starts.

If the crankcase heater is failed or wasn't energized long enough after a power outage, liquid refrigerant migrates into the oil. On start-up, the compressor runs with liquid diluted oil. The oil pressure (or differential oil pressure switch) doesn't build properly, and the compressor cuts off on the oil safety switch after a few minutes.

Technicians often misread this as a mechanical failure because "the compressor runs, then trips a safety." But if you check the crankcase heater and let it do its job for 4-8 hours, the problem may disappear.

How to check:

• Make sure crankcase heater is energized whenever the compressor is off (especially in cold weather).
• Measure the heater's amperage — should be drawing roughly 1.5-2.5 amps depending on size (consult the specs).
• Feel the base of the compressor? It should be noticeably warmer than the ambient air.

I had a recurring "short-cycle" call on an AAON water source heat pump in February 2024. The heater was wired to the wrong contactor terminal, so it only energized when the compressor ran. Compressor on → heater on (added heat of no value); compressor off → heater off. So after the compressor stood still for a few hours, liquid slugged the oil. Tripped the oil safety every time. That fix cost nothing but time. (Good thing too — the customer was ready to replace the whole unit.)

Checkpoint: Verify the heater circuit has power when the compressor is off, and the heater is actually producing heat.

Step 5: Check the Refrigerant Charge — Don't Guess, Weigh It

Low refrigerant charge can cause low suction pressure, which opens the low-pressure switch. But low charge often appears after you've eliminated airflow and switch issues.

Measure superheat and subcooling. For commercial units, the subcooling method is usually recommended. But here's where my experience matters: I've seen a system with slightly low charge cause a 3-minute run cycle on a 20-ton rooftop unit, but on an AAON chiller with a fixed orifice, it was just poor performance, not short cycling. Context matters.

If you're troubleshooting an AAON unit with a TXV, target subcooling of around 8-12°F (typical, but check the nameplate). If you don't have hard data for the exact target on that model, don't assume it's the same as the last unit. (I don't have every AAON spec memorized — I have the factory app on my phone like everyone else.)

If the charge is low, find the leak. Fix it, then weigh in the exact charge from the nameplate. Weigh it in — don't "top off" based on pressures. Topping up with refrigerant is the fastest way to turn a small issue into an overcharge, and overcharge causes high head pressure, which cycles the compressor off via the high-pressure switch. And then you're back to short cycling, but now you have two problems.

Checkpoint: If superheat and subcooling are within range, and the charge matches the nameplate, your problem is likely not refrigerant quantity.

Step 6: Inspect the Metering Device (Especially If It's an EEV or TXV with a Bad Bulb)

If the refrigerant charge is good, airflow is fine, and the coil is clean, a restricted or failed expansion valve can give you the same low-suction short-cycle pattern. A TXV that's stuck partially closed or an EEV with a bad stepper motor can reduce refrigerant flow to the evaporator until the low pressure switch opens.

How to identify: The evaporator coil will be starved — low suction pressure, low evaporator superheat (because the refrigerant is boiling off too quickly in the first part of the coil), and a warm suction line back to the compressor.

But here's the thing: I've also seen a completely blocked TXV cause compressor shutdown in exactly 2 minutes, 30 seconds after start. Never expected how regular the cycle was. Turns out the liquid line was frozen due to residual moisture in the system (the TXV's sensing bulb was also wrapped in a super-old chunk of insulation that made the valve hunt). Replace the TXV if it's internally damaged, but also do a moisture check on the refrigerant.

On AAON units with digital scroll compressors, the behavior can be stranger: if the modulation solenoid is stuck or its wiring is wrong, the compressor can unload and reload in an odd cycling pattern. If you're on a unit with AAON's variable-capacity scroll, check the compressor's operating mode and the controls' setpoint before condemning the compressor.

Checkpoint: If you have a temperature clamp on the liquid line before and after the TXV, the temperature drop across the valve should be significant (usually 20-40°F depending on conditions). If there's no drop, or it's inconsistent, the valve may be stuck open or closed.

Step 7: Look at the Controls and Thermostat — And Then Stop Blaming the Compressor

Last but not least: the controls. A thermostat that's losing its setpoint signal, a faulty temperature sensor in the return air or space, or a controller/board with a misconfigured anti-short-cycle delay can all produce a 2-3 minute cycle.

Check:

• Time between cycles — is it exactly the same each time? Could be a programmed minimum run time or a delay timer.
• Does the controller report an alarm or lockout code? If it's an AAON unit, the controller can store a historical code list. Use that before replacing parts.
• If you have a communicating thermostat, check for a faulty sensor or wiring issue.

I once spent an afternoon on a new AAON rooftop unit that kept shutting the compressor off after 3 minutes. Turned out the installer had crossed the return temperature sensor with the supply sensor. The controller "saw" the supply temp rapidly falling so it shut down the compressor. Swapped the sensors and it ran perfectly. (Yes, this happened. No, I won't name the company, but it was an easy fix that we only found after pulling our hair out.)

Checkpoint: If the unit runs longer than the suspect switch differential or no safety trips are logged, the problem is likely in a control loop, not a hard component.

After the 7 Steps: What to Do When It Still Cycles

If you've made it through all seven and the compressor still shuts off after 2-3 minutes, you need to collect data, not throw parts at it. Hook up an amp clamp on the compressor circuit and a temperature/RH data logger. Make sure you have accurate run times, pressures, and amp draws. Record them and send them to a technical support rep — AAON has decent support documentation online, and their parts distributors usually have tech contacts. In some cases you'll need a compressor megohmmeter test to confirm winding insulation breakdown. But do that only after you've verified the electrical and refrigerant basics. Replacing a $3,000 compressor when the real issue is a $50 relay is not a fun conversation.

Quick Reference: Most Common Culprits in My Experience

• Dirty condenser coil and/or filter: 30% of cases
• Faulty pressure switch (often low-pressure): 20%
• Control/sensor misconfiguration: 15%
• Low refrigerant charge (contains a leak): 15%
• Crankcase heater failed (or wiring error): 10%
• Genuine mechanical failure: 10% (maybe less)

I don't have hard data on industry-wide percentages; these are approximations from my own repair logs. Your fleet may vary. But the takeaway is consistent: the compressor is often the victim, not the culprit.

A Few Cautionary Notes (Because I've Made All of These)

• Don't jumper out a safety switch and leave it. I've seen that cause a $6,000 compressor replacement. Use jumpers only for diagnosis.
• Don't assume that because it's an AAON unit, any part will fit. AAON has made many variations of their units across years. Always match the serial number, model number, and part number before ordering. When in doubt, send a photo or get the OEM part number from the unit's nameplate (or call an AAON parts distributor — they're there for this).
• Don't use an oversized contactor just because it's what's in the truck. Mismatched contactors cause chatter and premature failure. (We caught 47 potential errors using a pre-dispatch checklist over 18 months — that's not bragging, it's just what happens when you document mistakes.)
• If the compressor is hot and tripping on internal overload internal overload, wait for it to cool down before testing continuity. Otherwise you'll think the winding is open when it isn't.

Look, the 7-point list isn't magic. It's a process that covers 90 percent of short-cycling complaints I've seen. Will it solve every issue? No. But it's the fastest path to catch the obvious problems, and it has saved us an estimated $8,000 in potential rework over the past few years. That's not a number I invented; that's the sum of avoided compressor orders and extra service calls.

If you make it through all seven steps and still stumped, take good notes and call us. Yes, we're a parts supplier — and if you need a replacement compressor, we'll get you one. But I'd rather sell you the correct $75 switch than watch you buy an expensive compressor only to have it fail again after 2-3 minutes. (That's just bad business.)