AC compressor not engaging

You have not said anything about the ground circuit. There are actually two places you could have a problem, the ground wire or the twelve volt feed wire you think is okay. Most of the time checking for twelve volts is sufficient to move on, but not always. You have two fuse link wires feeding the AC controller, and a voltage spike suppression diode electrically across the compressor coil, but in a different part of the wiring harness. That diode could be shorted, but regardless, when a fuse link wire burns open, it leaves a carbon track behind inside the wire's insulation. When you unplugged the connector to take the reading, there would be no current flow, so no voltage drop across the carbon in the fuse link wire. Any digital voltmeter will see the full twelve volts, but you will never get enough current through that carbon to do any work. The best way to take these readings is to back-probe the wires in the connector while it is connected. If you find zero volts that way instead of twelve volts, the zero volts is correct. That is on the black/yellow wire.

If you do have that twelve volts, back-probe the black wire. You should find zero volts there. If you find twelve volts, it is because it is coming through the clutch coil, and there is a break in the wire going to ground.

Measuring resistance is not the best way to test the ground either. All you need is one tiny strand of wire that is not corroded off yet to read zero ohms, but you will not get enough current through it to turn the clutch on. While the resistance is too small to measure, it is much easier to measure the results of that resistance, which is the voltage drop. Typically if you find twelve volts on the ground wire, it is cut or rusted off. If you find more than zero volts but less than twelve volts, expect to find a corroded terminal in a connector, or a badly-rusted connection where the wire bolts to the body or engine.

I never measured the resistance of a clutch coil, but since you found a valid reading, check it again with the connector plugged in. If it drops to zero ohms, you will need to find that spike-suppression diode and replace it.

You were on the right track. Do not go off on wild goose chase. Did you verify that twelve volts is still there with the connector plugged in? If it is, the problem is in the ground wire.

The wiring diagram does not show where the diode is, just that it is in the circuit. Most commonly I have seen them soldered right between the two terminals in the plug.

That relay supplies the 12 volts to the clutch coil, and you said you have that. I found a better wiring diagram. It shows the diode inside the relay box. If it was shorted, you'd be blowing a fuse when the relay turned on.

The coil's ground wire also goes through the relay box. If the ground wire for the entire box was open, the engine wouldn't run because the PCM relay wouldn't turn on. That just leaves the ground wire between the coil and the relay box. Check again for 12 volts on both wires at the clutch coil when it's plugged in. You must find 12 volts on the blue / white wire and 0 volts on the red wire. If you find 12 volts on the red wire, that proves there's a break in it.

If you have the chance to wire something, you have the chance to take a couple of voltage readings first. We need to know for sure if you have 12 volts on one wire, both wires, or neither wire when the clutch is supposed to be engaged. The engine needs to be running for the clutch to engage, so that is when the voltages need to be taken. Be sure to do that while the plug is connected.

As for running a new ground wire, if that is where the break is, that will get the system working, but I don't know why the engineers complicated the system unnecessarily by running it through the relay box. The ground returns for Engine Computer sensors run through the computer too, before they get to ground, because that ground circuit is monitored by the computer. That is not obvious simply from looking at the diagram. Your relay box diagram shows the ground path as a simple wire, but there might be more going that we don't know about.

Nope. This is actually making more sense now. Imagine if you had a garden hose with a pressure gauge on the end. With the faucet on the house open, and the nozzle on the hose closed, you find 12 pounds of pressure. Now stick your foot somewhere in the middle of the hose so it's 99 percent blocked. You'd still see 12 pounds of pressure. Now open the nozzle and try to get some water out of it that you can do some work with. All you'd get is a tiny dribble. The pressure would drop to 0 psi.

This is what commonly happens on cars when we use the more expensive digital voltmeter. You would have gotten different results if you had used a test light. Voltmeters measure electrical pressure, (voltage), just like the pressure gauge at the end of the hose. Test lights work by lighting their bulb's filament from current flowing through it. That means the circuit must be turned on and current is flowing. Your garden hose nozzle was turned on, but no current could flow because your foot created a restriction.

In your case, you measured 12 volts because you had the connector unplugged. No current was trying to flow, so the voltmeter picked up the voltage that made it through some huge restriction. Once you plugged in the connector, there was a path for current to flow through the clutch coil. That's the same as opening the nozzle on the hose. In the hose, the 12 pounds of pressure was dropped across the restriction caused by your foot. In your car, the 12 volts is being dropped across some restriction as yet to be located.

As a point of interest, this is a real common problem that confuses a lot of mechanics when working on dead radiator fans. Many are protected by fuse link wires which are smaller-diameter wires that are the weak link in the chain. Their insulation is designed to not burn or melt. The problem is when they burn open, that leaves a carbon track behind where the arcing occurred. You'll never get any current through that break to run a fan motor, but when it's unplugged, that carbon track lets enough current through for the voltmeter to "see" the full 12 volts. To avoid this incorrect test result, the test light is more accurate than the voltmeter, or just be sure to always take voltage readings with the circuit fully intact, meaning everything is plugged in and connected.

If you haven't already, switch the "WAC" relay with one of the other ones like it. Assuming that doesn't solve the problem, remove the relay, then test for voltage on the four terminals in the socket. I can't tell from the diagram if you should find 12 volts on one or two terminals. If you don't find 12 volts on any terminal, check the "AC fuse". That's a 15-amp. 12 volts should be there when the ignition switch is on.

Regardless if you find voltage on one or two terminals, remove the "PCM Power" relay, then see if any voltage went away in the "WAC" socket.

Another way to do this is to check for those voltages with the engine running. That will insure the "PCM Power" relay is on, and that means 12 volts will show up at one "WAC" terminal. At that point we know we're supposed to have voltage on two of those terminals. You'll need to remember which terminal gets 12 volts when, so we can figure out where to go next.

Rats. That is a good suspect. I've repaired some of those relays before, but only when they are not sealed. Three sealed relays were on a power window circuit board for a mid '90s Corvette. We found replacements on the Radio Shack web site. That board cost over $200.00 from GM, but we got it working with three new relays for a total of less than 20 bucks. Those were soldered in too.

If it looks like the relay is bad, I'd consider making a trip to a pick-your-own-parts salvage yard. You can just replace the module, or if you know how to solder and remove parts, you can transplant a relay.

Dandy. Happy it was an easy repair.