AC Clutch won't engage 96 Legacy

[Ceramiclover]

I have a 2.2 Legacy Wagon L series.

The AC does not work.

 

Light comes on at control head.  System was empty.  I Charged it up, and it did not appear to leak.  It held vacuum for 10 minutes.

Jumped AC Relay and compressor engaged.  Relay does not appear to be getting ground from ECM.

I checked continuity between terminals 2 and 3 of the sensor on the AC and it was OL.

I put a rheostat on this and went from 5k to 0 ohms.  Nothing.

As I understand it, this is a thermal protection switch on this model.  Therefore, a given resistance should mimic a proper temperature.

 

Jumpered pressure switch.  Nothing.

Checked ECM.  Wiring looks good.  There are splices at the wiring, but they appear to be factory splices that drop wire gauge going into ECM.

Wire diagram for 97 seemed to have the right colors.

 

Any ideas?  I have read what other people have done, and am stuck.

 

 

 

 

Any help is appreciated.

Edited March 13, 2015 by Ceramiclover

[Fairtax4me]

Have you checked the fuse for the AC in the underhood fuse panel?

 

The switch on top of the drier is a dual pressure switch. It does not react to temperature. This should be normally closed so if you have no continuity across it the system pressure may still be too low.

 

There is a thermal switch but it is inside on the evaporator core.

There MAY be a thermal switch on the compressor, which will have a 3 pin plug. If the compressor has only one wire in the plug, there is no thermal switch on the compressor.

 

Splices near the ECU are just the end of the sheilding around the signal wiring from the sensors on the engine.

[Ceramiclover]

All fuses in engine compartment are good.  Clutch relay terminals 85 (or 86) and 87 had 12v as well.  

0 ohms at pressure switch.

0 ohms at pins 2 and 3 of 3 pin connector on top of compressor.  I had OL before, but that may have been due to bad connection on my part, or doing it with the engine cold from sitting.  Today, I tested it after jumping (87 to 30) the compressor for 10 seconds or so.

[Ceramiclover]

Anybody?  Kinda stuck

[plogerjb]

Here's your diagnostic pinpoint procedure for a non engaging clutch... Should be helpful. You have two fuses, one relay, dual pressure switch. fan switch, AC switch, and an evaporator thermostat. You say the light at the switch comes on, which kinda rules that out as a problem for me. I assume your blower fan works (or else we'd be talking about that), but that doesn't rule out any possible harness issues. You say the clutch engages when jumping the relay, but is that leaving the relay in place and tripping it or are you bypassing it completely?

 

Forgive me, I'm kinda worn out from this weekend. Let me know if there's anything else I can do to help.

 

[Ceramiclover]

I have done both.  Just went out and grounded 85 (or 86) of relay.  Clutch engaged.  I might be able to grab an ECM and try that.  Otherwise, how do you test (and where) the evaporator thermostat switch?

 

Thanks for your reply.

[Fairtax4me]

Did you only leave it jumped for 10 seconds?

Relay to compressor clutch is a straight shot, so if the relay is jumped the compressor will run continuously.

 

Thermal switch on the compressor should be normally closed. I can't find a temperature diagram for that switch so I don't know at which temps it should open. If you have 0 ohms there AND at the pressure switch on the drier then you should get 12v to the coil side of the relay (pin 85) with the key On. Check for 12v at both of these by back probing the connectors.

 

The relay is grounded by the ECU, with the engine On and AC On you should have continuity to ground through terminal 86 at the relay holder.

 

If you have no ground at 86, find the evap thermo switch connector behind the glove box. This is a black 3 wire connector just above the resistor block (blue connector)

 

You should have 12v at pins 1&3 when back-probing the evap switch connector with AC on. The evap switch provides 12v signal to the ECU that commands the relay to ground.

 

The AC switch in the dash grounds the center pin, the thermal switch on the compressor sends 12v to pin 3, and pin 1 is 12v out to the ECU.

Should have 12v at pins 1 and 3. If you have 12v at pin 1, check pin 60 at the ECU for 12v.

 

Pin 31 at the ECU is ground for the relay coil. Check for continuity to ground there with AC on. Check for 12v there with AC off.

 

 

Edited for incorrect info and some bad spleling

Edited June 28, 2015 by Fairtax4me

[Ceramiclover]

Do you have a diagram for the ECU?  Wire color and visual position would be helpful.  Thanks.

[Fairtax4me]

www.main.experiencetherave.com/subaru_manual_scans/

 

Check here for the factory manual for your car. The wire diagram for the AC system has the pinout and wire colors for the connectors.

[Ceramiclover]

"If you have no ground at 86, find the evap thermo switch connector behind the glove box. This is a black 3 wire connector just above the resistor block (blue connector) 

Yuo should have 12v at all three wires when back-probing the evap switch connector with AC on. The evap switch provides 12v signal to the ECU that commands the relay to ground. 

The AC switch in the dash sends 12v to the center pin, the thermal switch on the compressor sends 12v to pin 3, and pin 1 is 12v out to the ECU. 
If you have 12v at all three, check pin 60 at the ECU for 12v. "

It seems like the center pin is ground side controlled from the control head.  It rests at 12v at goes to ground with AC switched on.  The sensor shows + - L on it.  The outside pins both had 12v.

 

I supplied a ground to it to see if it didn't have a good enough ground.  nothing.

Jumped 12v to it with switch on and it tripped my power probe circuit breaker.  seems ground-side controlled.

[Ceramiclover]

12V at Pin 60 and 31 of ECM.  Grounded 31 and AC got Cold.

 

Does this mean that all safety sensors (Pressure, thermal on compressor, and on evaporator) are good, but the ECM is not turning on relay?

Therefore faulty ECM?

 

If so, could I bypass ECM by wiring relay:

85 12v from Pin 60

86 (ground)

87 (ground)

30 (cut ECM out and connect Pin 31 wire)

 

Or am I going to screw everything up by not allowing the ECM to control duty cycle?

[Ceramiclover]

What years and models would have the same ECM if I went to the junk yard?

[Fairtax4me]

Hmmmm... Something isn't making sense here. The AC switch gets 12v direct from the fuse panel.

 

I had come to the conclusion that the wire diagram must have an error because I don't see how the circuit can work properly the way they've drawn it.

The diagram shows power from the AC fuse goes through the pressure switch, then the compressor thermal switch, then ties into the Blue/orange wire which is grounded by the ECU. But if the ECU supplies ground there it will just blow the fuse. I think the yellow wire from the compressor thermoswitch is supposed to tie into the yellow wire from the evap thermoswitch, thus supplying 12v to both the evap thermoswitch and the AC relay coil. The relay coil is then grounded by the Blue/orange wire via pin 31at the ECU. This is how it's done in the 95 diagram, which is the same minus the compressor thermoswitch.

 

It looks like the AC switch is supposed to supply 12v to the ECU via the evap thermoswitch, but if the mode control panel is grounding that circuit, then the ECU must be supplying 12v to the evap switch, and current flow in the circuit is being monitored to determine if the AC switch is On or Off.

 

I checked the 96 diagram against the 97, and 99 diagrams I have and they are all the same. When I go to school later I'll check Alldata and see if they have an updated diagram or better diagnostic checklist for the AC system.

[Ceramiclover]

I think they just have a typo on the yellow wire splice.  I think yellow from thermal switch on compressor splices onto yellow from evaporator thermal switch

[Fairtax4me]

Bypassing the relay isn't a good idea because the evaporator will ice over and can rupture. The pressure in the system is fairly well regulated by the expansion valve, so the compressor will continue to run indefinitely. The thermal switch on the evaporator is used to keep the core temperature in the proper range to prevent icing.

 

I suppose you can try it and see if the compressor cycles on/off as it's supposed to.

 

 

Still trying to determine proper operation of the evap thermoswitch.

 

Check if the ECU is supplying power to the evap thermoswitch, unplug the connector and see if you have 12v at pin 1.

[Ceramiclover]

I'm not going to bypass the relay.  I am going to use the 12v FROM the sensors to trip a relay and ground the relay under the hood.  I think I may need to use the ground from the AC switch.

This way, the compressor should cut out if any of the sensors opens.

Is this how the ECM normally does it?  Or does it duty cycle it more actively?

Edited March 18, 2015 by Ceramiclover

[Fairtax4me]

Ok, here's the scoop. Alldata had the goods on this. 

 

ECM is supposed to receive a 12v signal from the evap themoswitch. 

At Evap thermoswitch:

12V comes in through the yellow wire, pin3. 

The ground is supplied via the the AC switch/ mode control head. 

12V is applied to the Brown/yellow wire, pin 1. This wire goes to ECU pin 60.

 

12v applied to ECU pin 60 does three things

1. command the ac relay control to ground

2. commands AC sub-fans to ON

3. commands IACV to raise idle speed to compensate for compressor drag

 

 

If you apply 12v to pin 60 the ECU should make everything work (compressor engage, sub-fans turn on, idle speed raises). If so the evap thermoswitch is bad.

 

 

You've already checked ground through the control head and that works, and you verified 12v is present on the yellow wire at the evap thermoswitch, correct? 

Edited March 18, 2015 by Fairtax4me

[Fairtax4me]

Updated Diagram and schematics for your viewing pleasure

[Ceramiclover]

I have 12v at 60, but no ground at 31.

[Fairtax4me]

Have you unplugged the ECU and checked for any corrosion in the connector? 

Wiggle test the wire to make sure it isn't a loose connection at the ECU pin. 

 

Next step is replace the ECU. 

 

96-98 Legacy ECU is needed.

If your engine has EGR, you need an ECU from a car with EGR. Also try to match the Evap system. If yours has the charcoal canister under the hood, get an ECU from a car with the canister under the hood. 

If searching eBay or used parts sites try to match up the 10 digit part number on the ECU. Generally, EGR and Evap system differ between Manual and auto trans cars. So when searching online be sure to indicate auto or manual trans to match your car. 

[Bushwick]

Was this ever fixed? Hate seeing teased out threads with no results :'(

[Ceramiclover]

As anti-climactically as possible.  I hardwired the relay:

 

85 12v from Pin 60

86 (ground)

87 (ground)

30 (cut ECM out and connect Pin 31 wire)

 

This still utilized the sensors for the safety of the system.  It worked for awhile and then stopped.  I think there must have been a leak, despite the fact that it held vacuum when I tested it.

 

Then I gave up and sold it for $1000.

This was part of my Subaru rescue service.  Some girl had driven this into the ground and let it get impounded. I got it to start and to stop leaking gas from the filler neck and pressure regulator.  Now it is in a good home with a 16 year old.

[Ceramiclover]

Also, splitting with diodes, I connected 30 to the AC Fan relay.

[Fairtax4me]

Good work-around to get the ECU out of the circuit.

 

Probably had a leak at an o-ring. O-rings can seal with vacuum on them but leak when there's 150+ psi pushing on them with the system running.

[Bushwick]

Thanks for a quick reply. Never know if someone died or gave up. Though usually it means they fixed/replaced whatever the last thing they were going to check was and forgot to update it worked. Going through a similar issue where mine runs everything then shuts everything off randomly. 95' Wagon L. Thinking it might be leading up to what caused yours to fail. Oh well, it'll get straightened.