Guide

Repair it, or replace it.

The questions that actually decide it, from someone with nothing to gain from either answer.

Updated October 3, 2026EPA 608 certified

Every page you will read on this question was written by somebody who sells air conditioners. This one was not, and cannot be.

No job here can go over one thousand dollars in labor and materials. That is a legal limit, not a positioning statement, and it means there is no version of this conversation that ends in me selling you equipment.

So here is the honest version.

The question that decides most of it

What actually failed?

Almost everything that commonly fails in a Valley summer is cheap and worth fixing regardless of the system’s age: capacitors, contactors, fan motors, float switches, drains, thermostats. A fifteen year old system with a dead capacitor needs a capacitor.

The expensive failures are the compressor, the evaporator coil, and a refrigerant leak that cannot be reached. Those are the ones where replacement enters the conversation legitimately.

If somebody is quoting replacement for anything on the first list, that is worth a second set of eyes.

The second question

Is the rest of it healthy?

A compressor failure on a system whose coils are clean, whose charge was right, and whose other components are sound is a different proposition from a compressor failure on a system that has been limping for three summers.

Readings tell you this. Amp draws, pressures, delta T, coil condition. Numbers you can keep and compare.

The things that genuinely point at replacement

  • R-22 and a real leak. The refrigerant is out of production, which makes the arithmetic stop working rather quickly.
  • A failed compressor on an older system. Often close enough to the cost of new equipment that it is no longer a repair decision.
  • Repeated failures of different components. One part failing is a part. Three different parts in two summers is a system telling you something.
  • An evaporator coil leak on equipment past its warranty.

The things that do not, on their own

  • Age. Age is a risk factor, not a diagnosis. Plenty of old systems run fine on cheap parts.
  • “They do not make parts for it.” Sometimes true, often said. Universal replacements exist for a great many components.
  • A high power bill. Usually a dirty coil, a bad charge, or leaking duct, and all three are cheaper to address than a new system.
  • Noise. See the noises guide.

What I would actually do

Get the fault identified with measurements, in writing. If it is on the cheap list, fix it and get on with your summer. If it is on the expensive list, that is when you get replacement quotes, and that is when a second opinion on those quotes is worth the small money.

If yours genuinely needs replacing, you will hear it here plainly, and you will be handed to a licensed Arizona contractor. No fee, no markup, no obligation. That is the whole offer, and it is the same whichever way the answer falls.

Straight answers

Questions that come with this one.

How long should an AC last in Phoenix?

Less than the national figures suggest, because a Valley system runs close to flat out for five months a year. Treat any national average as optimistic here. What matters more than age is what has already been replaced and what the current readings look like.

Is the 5000 rule real?

It is a rule of thumb, not arithmetic: multiply the repair cost by the age of the system and compare against 5000. It is better than nothing and it ignores everything that actually matters, like whether the failed part is the expensive one and whether the rest of the system is healthy.

Why should I trust this page?

Because no job here can exceed $1,000 in labor and materials, so selling you a system is not something that is allowed to happen. If yours needs replacing you get told so and handed to a licensed Arizona contractor, with no fee and no referral markup. There is nothing to gain from either answer.