Cold AC at start, not cold after 10 minutes, repeatable

Is it normal for the high side to be almost too hot to touch after 3 or 4 compressor cycles?

here's a video of how it acts at first:

Sounds like it's slipping to me.

I wonder if the high pressure safety switch is tripping and killing the compressor. The lines should be warm to the touch but if they are hot enough to burn you something is very wrong. I would bet the system is clogged or overfilled with oil/stop leak.

The problem with most of the store bought cans is that they include oil and stop leak in the charge. Well, when AC systems leak they lose refrigerant, but the oil stays in the system. People just keep cramming the store bought cans in and after repeated extra shots of 3oz of oil, the system is compressing more oil than it is refrigerant, system pressures skyrocket, burning hot lines, performance suffers, other damage, etc. Stop leak is also generally not great for any refrigerant systems, though it has worked for me in the past.

I hate to be the "take it to a shop" guy but at this point I would really take it to a pro with a real AC recovery/charge machine. You have no way of knowing how under or overcharged the system is, if it is triple filled on oil, etc. They will be able to pull all the refrigerant and oil from the system and recharge with the correct amounts of both. From there, you know you are starting with the correct amounts and can DIY repairs as needed, and use recharge cans with NO OIL.

AngryCorvair (Forum Supporter) said:

Is it normal for the high side to be almost too hot to touch after 3 or 4 compressor cycles?

Assuming the high side is getting to 200psi (rough guesstimate), 200psi of R-134a is about 130F, so... what do you consider too hot to touch?

pointofdeparture said:From there, you know you are starting with the correct amounts and can DIY repairs as needed, and use recharge cans with NO OIL.

+1 on excessive oil.  I've found that a simple evacuation will not remove excessive oil, one needs to do a thorough system flush. 

And if PO was a dingbat and used those handy dandy cans of refrigerant with sealant in them, the car is now a rolling air conditioning service equipment destruction device, and the refrigerant should be vented to atmosphere and all components replaced.  It's cheaper than buying someone a new Robinair machine because your car wrecked their equipment.

That E36 M3 should be banned.

So I popped the hood and was gonna hook up my manifold and gauges. I heard a tiny "pffft" when I removed each valve cap, and I saw this happening on the high side:

In reply to AngryCorvair (Forum Supporter) :

Honestly?  Ignore it unless it's chronic.  That can be an artifact from the last time it was serviced.  Some cold refrigerant still in the trapped space between the valve and cap, heats up to ambient temp (and beyond!!), some goes pfft, the rest slowly boils off while you watch.  The boiling point of R-134a at ambient pressure is about -14F.  The pressure at 80F (guesstimate for your locality) is 87psi.

Blow it out with compressed air, drip a little water in there, watch for bubbles after the water is able to find its way to the bottom.  Proper bubble solution is better, but it's a bit much to expect you have that laying around, and I hate soapy water, it doesn't have the right bubble persistence.  (Personal prejudice)

Ambient temp 90F, parked in direct sun, black interior, gauges on, start engine, 50 psi low and 250 psi high, compressor has been running about 5 minutes straight, AC is blowing cold, fans are cycling properly.

It's overcharged.  If the low side is 50psi and the high side is that high.

Here's what the system is doing after 20 minutes of idling. Compressor clutch still hasn't disengaged.

10 minutes later, clutch still hasn't disengaged, AC is still cold. Fan kicks up when high hits 270 psi and kicks off when high drops to 225.  Low side is oscillating between 40 - 50 psi on about the same period as shown in above video.

I'd be a lot happier if the low side was lower, that seems overcharged to me.  I can't see that being very cold, either - I'm not happy unless vent temps are 40-45f with the fan on medium, set to fresh air.  With low side pressures that high it shouldn't be possible to get temps that low.

Fans controlled by high side temp does throw a monkey wrench into the works, though.

With AC still ON, I turned temp control from MIN to MAX and waited til I felt heat out the vents, then turned it back to MIN. Clutch never disengaged, low side dropped to 25 psi and high side to 225 psi. Took about a minute for pressures to return to oscillating at prior values.

THEN (cue mysterious music), I pressed the AC button to turn off compressor. Almost immediately, low side went to 100 psi and high side dropped to about 130 psi, and the compressor would not restart when I pressed the AC button again. Does that tell us something?

Is the clutch engaging?  I bet not.  Do you have power to the clutch?  If yes, figure out why the clutch goes out when it's hot.  If no, figure out why the control head or ecu isn't turning it on.

In reply to AngryCorvair (Forum Supporter) :

When the compressor is not re-engaging, whack it lightly with a hammer, or dump some cold water on the compressor clutch.

That is indeed a symptom of a clutch starting to fail: as it gets hot and resistance goes up, it has less magnetic strength to pull the clutch engaged, even if it has enough strength to keep it engaged.

Percussive assistance or a little cooling can allow it to re-engage.

Too much gap:

Problem solved:

In reply to Pete. (l33t FS) :

when the compressor won't restart, it is because it is not getting +12v. I determined this with a test light. I let the system run until things were "hot enough", then I turned off via AC button. When I pressed the button again, the idle speed bumped up and the cooling fan kicked on, but the compressor did not engage.

So I disconnected the clutch wire  and checked with test light, no juice.

So I jumped from B+ to clutch and it engaged.

then I shut everything off and let it cool. Came back later and restarted with clutch connected, and it worked. Pulled the wire and checked with test light, and it lights up. Reconnected it and it engaged. Did this a few more times and it stopped engaging. Checked with test light, no +12v.

I want to test the AC relay but I need to study the schematic first, make sure I'm not on the wrong path. Seems to me that if the idle speed is bumping up and the cooling fan is coming on, maybe the relay is OK? I'm not sure what components and functions are on that relay.

In reply to Woody (Forum Supportum) :

I checked that site, good info but not my problem. For some reason I'm losing the +12v to the clutch.

In reply to AngryCorvair (Forum Supporter) :

Ugh.... I admit to never having had to do electrical diag on a Subaru.

But this is a common failure mode for Hondas:  They have separate relays for the A/C clutch and fan.  For whatever reason, the relay for the fan doesn't fail but the one for the clutch does.  The clutch relays fail in both directions, too: sometimes they stop engaging, sometimes they stick on.  "Stick on" is a fun one, because the computer has shut off the fan, so A/C high side pressures go right on up to the point where the safety blowoff on the compressor does its job at 450psi or so.  If there is dye in the system, it looks like a Predator had an accident in the engine bay.

I'd swap the A/C clutch relay for the horn relay,  but Murphy suggests that the A/C relay is going to be something different from anything else you may want to swap with.

STICKING RELAY FTMFW!!!

Marked and Switched AC clutch relay with secondary cooling fan relay. Noticed fan not spinning, gave the relay a solid flick, fan came on. About 10 minutes later, same thing, solid flick got it going again. Swapped them back, went for a drive. 2 blocks out, AC not AC'ing. Lift hood, give relay a flick, AC clutch engages. After that, it worked like normal the rest of the drive. Thanks Pete (l33t FS)! Your knowledge was very helpful on this whole AC ordeal. Tomorrow I will buy a new relay when I return the unused rear wheel bearing.

I had the bubbling at the port with our '05 Odyssey, and the A/C wasn't working well.. I replaced the valve with a special tool that allowed you to unscrew and replace the valve without evacuating the system. In my case, I also noticed the lines around the valve were dirty with some kind of oil (A/C oil).

I replaced the valve, and took it to the dealer for a vacuuming/refill and all was well with the A/C again.

Scientific method plus expertise for the win

Nice repair and problem solving! Love when it's something simple like that. 

Woo Hoo!  Nice work all!  Great thread.