AC issue

[Bushwick]

Have an issue with the AC in a 95' Legacy that I can't quite figure out. It's been refilled with 134a and is within the correct range when charging for outside temps.

 

Problem is compressor is randomly engaging and disengaging at idle, and when hot out it might not even engage at all though that's rarer. Both fans are kicking on with the AC. If you pull the AC relay in far left of the row of 4 AC labeled relays, driver side fan shuts off and the odds of AC staying on just increased to almost 90%. I tried pulling each relay when AC is charging just to see if any of them were faulty and they appear OK. I ruled out the sensor on top of the aluminum can that's up against passenger corner of firewall by pulling connector off and jumping it. 

 

To be clear, AC and both fans kick on, might run 30 seconds, then both AC and fans kick off. Sometimes it'll kick back on immediately, after 5-10 seconds, or maybe minutes later. HOT weather like in upper 80's, it might not even come on. Last time that happened drove several miles and AC never came on. When AC is on, it's very cold. Also, revving at idle above a certain point will kick unit off correctly.

 

Any other sensors directly connected to AC that might be worth inspecting? I can't find any other than the one threaded into top of the can mentioned above. Maybe a faulty temp sensor somewhere?

[Fairtax4me]

95 would have had R134 anyway. 1995 was the year the EPA mandated the dis-use of R-12 in new vehicles.

 

How many wires go to the compressor? If there are 3 wires the compressor has a thermal shutoff switch. If just 1 wire then no switch.

 

There is a thermal amplifier on the evaporator core which will shut off the compressor if the core gets too cold.

 

The pressure switch on the reciever drier is a dual pressure switch. It turns off the compressor if the high side pressure is too high, or if overall system pressure is too low.

 

How do you know the charge is within the correct range? Do you have manifold guages on it to see the system pressures?

 

Sounds to me like its possibly overcharged, or the system has air or moisture in it. Moisture in the system will cause the expansion valve to freeze and cause the high side pressure to rise and shut off the compressor.

[Bushwick]

Compressor just has 1 wire. Pressure switch on the receiver drier only has one wire in/out. I pulled the connector off that and jumped it with a short wire section. AC still cut on/off like it had been. Evaporator core is the radiator thing up front, right? Where's that sensor at? I don't see any wires feeding around it.

 

I have a gauge that reads the low side pressure with ambient temps listed on face and adjustable dial. When AC is on and charging, it's right at the temp and within range. Revving the engine to near the cut-off, pressure drops out of that range. When AC kicks off, pressure goes higher than that range. But when AC compressor is actually on, it's within that range. 

[Fairtax4me]

Those can guages are hardly accurate. How much refrigerant have you added, if any?

 

If jumping the pressure switch doesn't change the operation of the compressor its possible the thermal amplifier is bad, or the evap core is getting too cold. With the compressor running check the temp of the low pressure line coming out of the firewall. Should be above 35°. If the line is frosted or below 35° the core is getting too cold.

 

Evap core is inside behind the glove box. Very much a pain to get to.

 

If you search around here there's a thread from earlier this year where I posted diagrams for the thermal amplifier that should give you an idea of how to test it.

[Bushwick]

Low pressure line was very cold, but no frost. It was in the mid/upper 70's today. That wouldn't account for it not kicking on prior when it was in the upper 80's. Car had been sitting a day or 2 when that happened.

 

I tried decreasing amount and the compressor kicking on got worse, so added a small can until it matched with temps.

 

 

Any idea what you typed in your thread's header so I can find it easier? Also, does the thermal amplifier wiring exit to engine compartment? Or does it stay under the dash and connect directly to ECM? What's it purpose? Some sort of line heater?

[Fairtax4me]

Its a fancy thermal switch. But it's got a tiny PCB and a couple transistors that turn a signal to the ECU on/off depending on the resistance of a temperature bulb that's stuck to the fins of the evap core.

The amp is on the outside of the blower box with 2 small wires that to inside the box through the seam. You can get to the amp itself to check voltages, but getting the temp bulb out of the evaporator box is tricky. The amp is mounted on the outside of the evaporator box near the plug for the blower motor resistor. Its a black plug with three wires.

 

I'll Search for the thread and post a link in a short

[Fairtax4me]

http://www.ultimatesubaru.org/forum/topic/152408-ac-clutch-wont-engage-96-legacy/

 

I called it a thermoswitch before, probably why you couldn't find it. FSM calls it a thermal control amp.

Wire colors at the connector should be the same for yours.

 

Your 95 ECU pinout is different from what's referenced in that thread for a 96. I don't know the which pin it is on the 95 ECU.

If you have the FSM for 95 it will say in the ECU I/O chart.

[Bushwick]

Thank you. Going to check the thread now.

[Bushwick]

I'll check the unit tomorrow. Wondering if the ground he didn't have might be intermittently weak on mine (leading up to his scenario).

 

Since everything is either on/off at the same time i.e. fans and compressor kicking on/off at once along with idle, it sounds like it might be ground or the 12v+ referenced below:

 

"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"

 

One more thing Fairtax, you mentioned "tricky". Is it really fragile or do you need be a contortionist to get at it w/o removing the entire dash? Just pop glove box door and reach for it? Or does something have to come out?

[Fairtax4me]

FSM says the blower box has to come out.

 

I kinda think if you can get the screws out of the front part of the box and pull the bottom half down some it MAY be possible to get the thermal bulb out of the core without pulling the whole box out, but it's very tight under the dash. I've not had to replace one of those, so this "theory" is just that.

 

My other thought is to cut a hole in the front of the air box to get access to the bulb and then tape the cut-out back in place, but that's risky because you can damage the core.

[Fairtax4me]

Your drier should have a sight glass on top by the pressure switch. Take a look in the sight glass with the compressor running. If it looks frosted/totally white the charge is low. Jump the compressor relay if the compressor won't run.

 

With R134 the sight will appear kind of milky clear with some bubbles going by when the system is full. If there are no bubbles the system is overcharged. If it just looks like frost add some more refrigerant.

Edited June 28, 2015 by Fairtax4me

[Bushwick]

I didn't see one when jumping the connector but wasn't really looking. Will check for it.

 

 

Which relay are you talking about and why/how would I jump it? One of the 4 in a line in the underhood fuse box? Looking at the 4 relays from driver-side fender, far left being 1, it was getting hotter than the other 3. Pulling it disengages the driver side radiator fan. But all 4 have same part number so swapped it with others and no change.

[Fairtax4me]

I want to say there are 5 total relays for the AC, but I don't remember off-hand.

IIRC there are Two for each fan. One hi speed, one low speed.

Then a separate one for the compressor clutch.

 

If you can't figure which it is, you can also jump 12v straight from the battery to the clutch, just be sure to use a fuse on your jumper wire.

Edited June 28, 2015 by Fairtax4me

[Bushwick]

OK, with glove box door removed and looking at the blower motor assembly, where exactly is the sensor behind the motor? Up top? or near back. I can feel the condenser core if I force hand up there.

 

Also, what's best way to remove the blower assembly? Anything to watch out for? Like little itty bitty springs that pop out if you aren't careful? Also, if blower motor is out, how can I check operation of AC engagement?

[Bushwick]

Went back and did more testing while waiting on response. In the row of 4 relays for AC, far right (no. 4) is for compressor. If I pull it out and wait 10 seconds, fans kick on and idle kicks up. If I reinsert the compressor relay, within 5 seconds everything kicks off. Remove compressor relay again and like before, everything kicks on. If I leave compressor relay OUT, everything will stay on. I've even tried jumping sensor on the can again and same deal. Tried jumping compressor directly with 14v (running voltage) and same deal. Everything will run briefly, then kick off minus compressor which'll nearly stall engine.

 

Still sound like this thermal switch? Or does this sound like the ECM? How the hell does it know it's even running if I'm bypassing everything? Anyone?

[Bushwick]

OK, cleaned the sight glass really good. When unit kicks on, don't really see anything. Fluid seems lower than glass. Then it kicks off, which a bunch of little bubbles surface to the glass briefly, but bubbles only appear after unit shuts off. It's cycling a LOT more with the ambient temp drop and can't even get it to stay running now.

[Bushwick]

Ended up having the line evacuated to a low psi. Basically evacuate until compressor remains on. Was blowing mildly warm air at this point. Then got a small can and slowly added until 1. cold was coming out vents 2. compressor was still staying on. Seems there's a very small window here before this thermal sensor goes nuts. Had to ignore the gauge and just fill enough until cold was blowing. Sight glass shows a bunch of fluid turbulence with occasional bubbles. Even at this, compressor occasionally cut out then couldn't make up it's mind if it wanted to come back on or not. Really low 60's out and air was cold but not as cold as it can be. Will have to wait until it gets hot out and see if it's enough to blow cold or not. At least for now it was staying on fairly consistently, but have a feeling that thermal sensor is too sensitive.

[Fairtax4me]

Have you tried adding more refrigerant?

 

My computer with all of my service manuals is down at the moment, so I'm actually having to refer to my haynes manual.

 

The relay furthest towards the firewall should be the compressor clutch relay. With that removed the clutch will not engage, but the ECU will have input from the AC switch, the pressure switch, and the thermal switch that the system is working, so the ECU will engage the fans, the idle will bump up, and it will then ground the compressor relay circuit to try to engage the compressor clutch.

Without the compressor running, the ECU will still get the 12v signal it needs to see to know that the system is working properly, because neither of the sensors will react. There's no change in pressure, and no change in core temperature.

 

When you jumped the clutch relay, now the compressor is running, and one of those sensors is telling the ECU to turn off the compressor. The ECU will turn off the fans, and decrease idle speed. But since the relay is jumped, the ECU can't turn off the compressor, which continues to run.

 

 

It sounds like the system may just be undercharged. An undercharged system will freeze the evap core quickly, and the computer will turn the compressor off until the core temp comes back above 35°.

[Fairtax4me]

Cross post.

How much refrigerant did you add? When refilling from empty I've always put in two 12oz cans. System holds 21-25oz.

 

If you're still seeing a lot of bubbles through the sight glass the system is still undercharged.

 

Also when its cool out the core will freeze up much quicker. Try running the heat on full blast on recirculate for 10-15 minutes to heat up the cabin. Then run the AC with the windows up, on recirculate.

Edited June 29, 2015 by Fairtax4me

[Bushwick]

I only added a small can. Seems the more you add, the more it wants to shut off.

 

 

Since originally posting the other day where it was randomly shutting off, only ambient temps changed and today it wouldn't even stay on. Started removing glove box and decided to try messing with fluid level instead leaving that sensor as last stop since not really looking forward to dropping the blower motor assembly.

 

 

With AC OFF, pressure was around 55 psi assuming gauge is even accurate. When charging, it was in the high 20's. This is lower than before, where it was in the mid/upper 30's psi when charging and I think high 60's-low 70's when AC was off, but again compressor just wouldn't stay on. Sight glass had nothing on it, but when compressor shut off, it'd bubble then. Which I think it was overfilled. Now when running, it's white and right on sight glass, but it's more like a turbulent rapids appearance and not frost. But it's at least staying on for lengths of time.

 

 

The low side line coming straight from firewall was cold, but not that cold. Wondering if that thermal sensor might of drifted as it's the only sensor left and is apparently overly sensitive. Before refilling, pulling compressor relay or disconnecting the single wire fans would come on and idle would kick up within 5 seconds but as soon as I reconnected the compressor it'd run maybe 5 seconds top before kicking everything off again. Beyond frustrating. There was a little more left in the can and every time I tried adding more it'd start kicking everything off. But when it was too low (after evacuating) it ran non stop. Seems like everything is pointing at that stupid thermal sensor. Should have a better idea once it warms up into the upper 70's hopefully.

[Silent Bob]

I think you need to put a proper set of gauges on it to see what the system is actually doing.   Just because the gauge on your can tap is in the green, does not mean the system is correctly charged.  I think at this point you  should have the system properly evacuated, vacuumed and charged by weight.  That way you have a good baseline to start diagnosing the problem.

[Bushwick]

I think you need to put a proper set of gauges on it to see what the system is actually doing.   Just because the gauge on your can tap is in the green, does not mean the system is correctly charged.  I think at this point you  should have the system properly evacuated, vacuumed and charged by weight.  That way you have a good baseline to start diagnosing the problem.

 

If this was a newer car, I'd agree. But as it stands, it was originally purchased as a winter only vehicle, which was then rotated into daily driver status due to being an auto, easier to drive + get in/out, and 10x cheaper to keep on the road vs. my Saab which might as well be a Porsche given how much more expensive parts are for that (Injured back some time ago and had several surgeries which limit life in too many ways as of late). Just spent $550 for new tires + alignment + inner/outers which would have been a fraction of that if I'd been able to do everything except mount tires. Unless absolutely necessary like tires/alignment, trying to watch how much goes into it. 

 

As it stands, AC works. But the random on/off was annoying. AC worked fine the last 2 summers it's been driven after initially being charged after acquiring. Wasn't until it started blowing warm that it caught my attention, which started all of this. Had been using a line with a gauge that was purchased separately a few years back and been accurate with other cars and this one prior.

Edited June 30, 2015 by Bushwick

[Silent Bob]

 trying to watch how much goes into it. 

Totally get not wanting to spend $ on a beater car. 

 

Point is, the gauge you are using is only telling you part of the story- the low side pressure.  Unless the system is almost totally empty, the low side is going to be  +-70psi with everything off, and +-30psi with the A/C running.  The high side is a bit more telling.  It should be around 225psi, A/C running.  Excessive pressure will shut off the compressor, and can indicate a sticking/frozen/ obstructed expansion valve.

 

One thing you can check without a gauge- Run the A/C until it acts up and shuts off.  When the compressor quits, unplug it, and with a meter or test light see if there is still voltage at the connector.  If so, the clutch may not be engaging.  Ran into that a few weeks ago,  cooled fine until the compressor cycled off- then the clutch would not re-engage.

[Bushwick]

As far as checking for voltage, it seemed to drop immediately while trying to see if voltage was still going to compressor or not when everything was kicking down. But a few times compressor engaged/disengaged very quickly when not probing, so it might be going out anyways, especially with all the random starting and stopping. I was reading elsewhere on a DIY clutch pull post, general consensus seemed to be to remove a spacer and reattach clutch assembly. Is that what you are referring to? Or in other words the way to fix it if that's the case? 

 

 

 

Where is the expansion valve? Does AZ or anyone else rent out a high pressure gauge? Didn't realize it's reading could lead to a diagnosis!

[Silent Bob]

. I was reading elsewhere on a DIY clutch pull post, general consensus seemed to be to remove a spacer and reattach clutch assembly. Is that what you are referring to? Or in other words the way to fix it if that's the case? 

That's what I was getting at.

 

 

The expansion valve is attached to the evaporator core.  Inside the HVAC box under the dash on a 95 I think.

 

Not sure if Autozone, etc have loaner gauges, but Harbor Freight sells a set that's around $50 or so. along with a $13 vacuum pump that's pretty slick.