Testing Subaru Pressure Switch

[Gnarsinski]

Hi, first post here. Seems like a very knowledgeable place. My wife's 2010 Subaru Impreza is having an issue with the AC. System seems charged, although I need to invest in a set of gauges to check. Compressor runs when jumped (and air blows cold, although I did not want to to run it too long) but does not run otherwise. I think its electrical. I'm not getting power to relays. It leads me to believe it is the pressure switch. Is there a definitive test for the pressure switch? Checking resistance across the terminals? It looks like I will have to get the system evacuated to change the switch. So I'd like to confirm it before I go further.

Also, I've had zero luck locating the switch. NAPA has no mention of it, even the Subaru parts dealers online gloss over it in their parts diagrams. No listing on Rock Auto, any suggestions on where to find one would me much appreciated.

Thanks!

-Kevin

[Cornbinder89]

I'm sorry but the new stuff is not my area. I believe that the A/C may be controlled by the ECM so jumper and test things with great caution. It is easy to do damage. I would agree, if it gets cold when the compressor runs, that it is likely an electrical issue. I would not jump to the conclusion that the switch or sensor (and we have yet to determine which it is) is the problem. It is possible that the ECM and a sensor control a relay, that controls the clutch. You are going to need a wiring diagram and someone with knowledge of how the electrical side of the system operates.

[Tom Greenleaf]

Welcome Kevin. Wish I could say I knew lots about Subarus in general but can't. As Cornbinder said be real careful jumping anything. This is highly likely all computer controlled. To add testing relays for power may be backwards as many ECM set ups are adding ground not power.

Switches may also be "thermal resistors" not like what you may be used to - a screw in switch for either high or low pressure monitoring plus act as the fail safes to shut system down.

You said you thought the charge was correct and truth is you really can't know for sure without evacuating/recovering all of it, measure what came out and charged back up to exact amount listed (should be) on a sticker under the hood. Then you can take pressure observations that it is in fact working within the ranges it should.

Get gauges. Make sure if cheaper ones they also read vacuum - some don't.

Basics first up to there should be codes stored in wait if system will not stay on help lead at least to what it's seeing that it won't stay on despite working when jumped. It should have a thermal switch to cut it off if pressures are too high as well which you can see with gauges.

Observe that fan(s) are working when A/C compressor is engaged or it would stop compressor with high pressure over the limits.

More thoughts as more info can be had on pressures and think about code reading that includes a reader than can scan for A/C,

Tom

[Nacho]

There are several factors involved in getting power to the compressor. The first thing to check are fuses. If OK, you have to trace all wires to key components to determine where is the failure. That's the way we do it. Feel free to email me for a system diagram, specifying your Impreza's engine size and if it has automatic or manual A/C.

[Gnarsinski]

Thanks for the replies. I'm planning on ordering Robinair gauges from Amazon, about $80. Seem like they should be pretty good quality, what do you experts think?

Fans are not engaging, but I think they also get their signal from the pressure switch, which is pointing me in that direction.

All the fuses I could find that are associated with the system seem ok.

I'll see if my Scan tool can look up AC codes.

Nacho, PM sent for those diagrams, much appreciated!

[Tom Greenleaf]

If you have wiring diagrams now from Nacho see if fan(s) are triggered by a pressure switch only or if thru ECM - some are detecting high pressure enough to need fans at all or don't waste the effort like for high speeds with enough air flow naturally. I can't know which car do that and which ones don't but more and more do that.

It can mean there just isn't enough high pressure to indicate it needs cooling either it really doesn't or incorrect info, switch or connection. Can also be plain low refrigerant charge.

Pressures are a great diagnostic way to determine if in limits or not for the conditions while being observed. Do also have accurate and fast thermometer(s) as pressures are totally dependent on the actual temps at the time plus RPM while compressor engaged. If not engaged static pressure also matters to know what that is at the temp of engine area where parts are that contain refrigerant,

[Cornbinder89]

Fans are not engaging, but I think they also get their signal from the pressure switch, which is pointing me in that direction.

All the fuses I could find that are associated with the system seem ok.

I'll see if my Scan tool can look up AC codes.

Nacho, PM sent for those diagrams, much appreciated!

You can't just assume the switch (if it is a switch and not a sensor) is the problem. IF it only has two wires, a switch, can't preform 2 functions with different perimeters, Fan control is high pressure related, compressor cut-out is a low pressure condition. A sensor that can "talk" to a electronic module, and it would be the module that would make the call on compressor cut out and fan based on the pressure the sensor sees. A sensor CAN preform two functions with different perimeters when combined with a module.
You have to understand how the system works so you can make a valid test of the components. Anything less is throwing parts at a problem and hoping to get it in the 1st or 2nd try.

[Cornbinder89]

To put it another way, a bad connection will give the exact same symptoms as a bad switch, only a valid test will tell the difference.

[Gnarsinski]

The pressure switch has a 4 pin connection. I haven't received diagrams from Nacho yet, but I found one for a similar model online. One connection is 12v, the other sends power out the rest of the circuit when the switch is activated, another goes to ECU and one goes to ground. I understand what you're saying. Until I get the actual diagram I can't fully diagnose and test. Any my gauges are on their way, should have them by Monday.

[Cornbinder89]

What ever you find, please post what it is, that is how we all learn and can help others.

[AC_Chris]

The pressure switch you have is called a trinary switch. If you ask for a trinary switch that may narrow the search. Sometimes they are called fan switches but trinary is a better term since they do three different things.. The switch has low medium and high pressure functions The compressor is shut off for very low pressure conditions, very high pressure conditions and the mid range part of the switch controls when the fan comes on. The diagram will make all the difference. May be some other component like a ambient temp sensor or evaporator temperature sensor, etc.. Good luck.

[Cornbinder89]

The trinary's I am used to never connect to the ECM. They are strictly for an Electro-mechanical system. They have 4 terminals, two control the compressor (power in, and out to the compressor clutch) which connect to a low AND high pressure cut-out switch, the other two terminal are strictly for fan control and can be NO or NC depending on the switch.
I am guessing that a Late model Subaru has something different esp if it is tied to an ECM.

[Gnarsinski]

Just got the wiring diagrams from Nacho, looking forward to troubleshooting tomorrow and Monday, I'll post back with my findings.

[Gnarsinski]

So here is what I think I've figure out so far. With gauges connected, system has 80psi on hi and low side. When AC compressor is jumped and running low side is about 20psi and high side is 110psi.

Per the wiring diagram, the pressure switch has 4 wires.
Terminal 1 is power from fuse 22. Its shared with the intake door actuator, the blower motor relay, and also goes to sub fan and AC relay. Voltage measures about 5 volts when the key is turned whats up with this??
Terminal 2 is is connected to terminal 1 through a normally closed switch, it goes to the evap sensor and to the AC relay. Checking continuity between 1 and 2 checks out good. Switch is operating per the wiring diagram
Terminal 3 goes to ground
Terminal 4 goes to the ECM and is connected to Terminal 3 via a normally open switch. Which shows open with a continuity check.

If I apply 12v to terminal 2, both fans come on and the compressor engages. So maybe my issue lies with why I'm only getting 5v into the switch?

[Cornbinder89]

Sure sounds like the feed is the problem. Are you checking with a digital meter? Often they detect a "nominal" back feed as they don't place enough "load" on the circuit being checked.
If I am understanding what you are saying, power comes in on terminal 1 when the blower is on, it then goes thru a NC frost switch and on to the control side of the A/C relay.
So trace back from terminal 1 until you get a good solid battery voltage.
Electrical is all about breaking into sections and testing until you isolate the problem between the smallest two sections.

[Gnarsinski]

Sounds like a plan. Ill see what I can find.

[huntindog]

So here is what I think I've figure out so far. With gauges connected, system has 80psi on hi and low side. When AC compressor is jumped and running low side is about 20psi and high side is 110psi.

Both the high and low pressures are too low. Looks to be under charged. A leak some where is a virtual certainty. Under charged and the compressor protections are keeping it from running, and doing damage. Don't jumper it any more.
Find and fix the leak, and put a proper charge in it....

[Cornbinder89]

Pressure can only tell you temp at the point measured! static pressure of 80 psi equals a temp around 77 deg, running pressure of 110 (high side) equal a condensing temp of around 94, or 17 deg above ambient, not bad, remember, the condenser fans are running. 20 psi low side is 24 deg F, a bit cool, but if the evaporator itself hasn't cooled yet (system run long enough stabilize temps and pressures), that can throw off things a bit. Low, yes but not out of the ballpark..
It may be a little light on refrigerant, but we don't have enough run time or info to say that for sure, At the pressures indicated I know of no safety that would shut the system down with those readings. It is classic proof to why state of charge can't be determined by pressure alone.
1st and foremost solve the electrical feed issue that you observed, once that is solved, then move on to any remaining issue. Without a copy of the wiring, I can not give more advice.

[Cornbinder89]

When low side is used for a cut-off the psi setting is quite low, IHC, for example, uses a 7 psi switch setting in the low side to shut the system down. High side shut downs a much higher but still well below the 80 psi range, and these would not trip with the compressor running until it has expelled virtually all the refrigerant. High side cut-out often are as low as 35 psi, anything above that and the compressor runs.
The one exception is one CCOT systems the clutch cycling switch is set around 28 psi, When one of these systems are low, the clutch will cycle more and more frequently until it reaches a point where the static pressure falls below the switch setting and the compressor never comes on. On these systems there is no separate "low refrigerant" cut-out, rather they let the system cycle shorter and shorter until it stops all together.

[huntindog]

Todays computers changed everything in the automobile world. Many systems will not allow the vehicle to hurt itself... Past systems that allowed a compressor to run as low as 7 PSI can hardly be called a protection system.

Is that the case in this instance? Don't know. I do know that nobody yet claims to know how it works for this car. And nobody has any specicfics on just what the suspect switch does...It cannot even be found.

I also know that 110/20 pressures are too low. Running the compressor longer will not improve that... It could wreck the system worse than it already is... Perhaps the computer is stopping it from hurting itself.

At any rate, I would find a shop that can evacuate and measure what is reclaimed.... It will most assuradly be less than it should be. If it holds a vacuum, then put the proper charge in it. If the system then works, chasing all these electrical gremlins is not necessary.

If it was low, then a slow leak must be found. The most likely places are the schrader valves, or the front compressor seal.