Problem with A/C compressor randomly not engaging

[inquirer]

The car is a Mitsubishi Lancer, but I guess it doesn't really matter.

Here is the thing. Randomly the A/C blows warm air, mostly for a couple of minutes max and then works fine. Maybe it's just a coincidence, but so far the problem mostly arises when the car is parked (obviously with the A/C turned on) and less often when driving.
When I turn on the A/C the radiator fan turns on, but the compressor does not engage.

The system was recently recharged. It had lost about 150 grams of R134a out of 500 grams during a period of more than 10 years. The recharge did not make any difference regarding the aforementioned issue.

 

[Nukeman7]

I suspect the air gap is excessive on the Lancer's A/C compressor clutch plate or you have a rusted clutch plate. Check out the video below that shows you how to adjust the air gap. If you see rust on the mating surfaces of the clutch plate, you should also wire brush it before reassembling.

 

[John105]

It could be that it didn't take 10 years to lose 150 grams, it lost it in x days, during the 10th year.

Generally speaking the output temp is lower at higher rpms so maybe the system is not efficient, but very noticeable at idle.

There's really no way to know online, likely system needs to be inspected. Pretty sure the car knows not to engage the compressor under certain conditions related to the refrigerant.

 

[inquirer]

@Nukeman7

If that's the issue, it is something that can be fixed relatively easily. I wonder if it can be something more serious, like a problem with the magnetic clutch.


@John105

I don't think it's very likely. If it was losing refrigerant relatively rapidly, I guess its performance would also start to suffer and the disengagements would be getting more frequent as its pressure was dropping over time. So far there is no noticeable drop of performance and the problem occurs with the same sparse frequency.
Furthermore the system passed a 20 minutes vacuum test before recharging. I don't know if 20 minutes of vacuum testing are 100% adequate, but I believe its another indication against a leak.

Pretty sure the car knows not to engage the compressor under certain conditions related to the refrigerant.

If the car didn't give the command for the compressor to engage, shouldn't at the same time keep the radiator fan inactive? In my case when the A/C turns on the radiator's fan starts blowing air, but the A/C compressor stays inactive. I think that the activation of the radiator fan means that the car also tries to engage the compressor, but for some reason the later fails.

 

[Jetronic]

I've found that AC systems always lose charge gradually, some more than others though. 150 grams in 10 years is very low.

The system needs inspected indeed, I'd be looking at the frost sensor first.

 

[inquirer]

I've found that AC systems always lose charge gradually, some more than others though. 150 grams in 10 years is very low.

The system needs inspected indeed, I'd be looking at the frost sensor first.

Yeah, I know. 150 grams loss (actually in more than ten years in my case, close to 15 to be more precise) it's something expected. The recharge though added some noticeable performance to the system, especially when the car is stationary.

I think it's almost certain that the issue is unrelated to the refrigerant level.

 

[Nukeman7]

@Nukeman7

If that's the issue, it is something that can be fixed relatively easily. I wonder if it can be something more serious, like a problem with the magnetic clutch.

Here is the thing. Randomly the A/C blows warm air, mostly for a couple of minutes max and then works fine. Maybe it's just a coincidence, but so far the problem mostly arises when the car is parked (obviously with the A/C turned on) and less often when driving.
When I turn on the A/C the radiator fan turns on, but the compressor does not engage.

It is not the magnetic coil in the A/C clutch. If it was, the A/C would not work intermittently as you described in Post #1. The A/C clutch magnetic coil should have two wires going to it. Disconnect the wires at the closest plug and jumper 12 volts to the clutch wires from the battery momentarily and see if the clutch plate will engage/disengage consistently as you cycle the power on and off. If the clutch doesn't engage at all, power up the coil and check for magnetism with a steel screwdriver. If magnetism is present and the clutch faceplate doesn't engage consistently, the air gap is too wide.

 

[inquirer]

@Nukeman7

I have read that some times the magnetic field could get weaker. Is this true? If yes it could explain the behavior I'm describing, if not then of course the magnetic coil can be ruled out as a possible cause.

 

[Mechanician]

Check for either battery voltage at the compressor wire, or the magnetism (usually easier) as described above, when the problem occurs.
If there is no magnetism, try a new compressor relay.
Usually there is another relay nearby, maybe for one of the fans, that can be borrowed for testing.
The compressor coil can be tested with an ohm meter, should be 3-4 ohms.

 

[vwmaniaman]

I am with Nukeman on this one. It happened on my daughter's Honda CRV and it also had a corroded ground connection

 

[Nukeman7]

@Nukeman7

I have read that some times the magnetic field could get weaker. Is this true? If yes it could explain the behavior I'm describing, if not then of course the magnetic coil can be ruled out as a possible cause.

I have also read that the magnetic field can get weaker over numerous years/decades. I suspect that the internal resistance can increase in the coil windings with heat, age and corrosion, resulting in a degraded field.

I have repaired A/C issues on two vehicles this summer due to A/C clutch problems. The first was a burned out coil where the plastic housing melted and the second one was an excessive air gap with a rusted faceplate.

 

[inquirer]

@Mechanician
Can't the relay get tested. I think it can. If yes and it isn't a complicated procedure I can test it anyway.

@vwmaniaman
Measuring the air gap is something I will probably do. I have to find the specs first though. By the way, it can be measured with the compressor installed on the car, right?

@Nukeman7
I guess I will start by measuring the air gap first. Especially if it's something that can be done without uninstalling the compressor there is no reason to not do it as a first step. As I said, I have to find the specs for my compressor though.

 

[SHOZ]

If you think it is the clutch, when it is on and suppose to be engaged but not, use a stick or something to bump the clutch face. If it's worn it will engage.

 

[Mechanician]

Typical air gap symptoms are reverse of yours, works when cool, then stops when everything gets warmed up.
But like every problem - not always.
Yes you can test the relay - usually the problem is the contacts get burnt, and have a problem connecting at times.
It's easier to just swap a relay with the same numbers from a different circuit, than it is to try and catch that happening.

 

[inquirer]

Typical air gap symptoms are reverse of yours, works when cool, then stops when everything gets warmed up.
But like every problem - not always.
Yes you can test the relay - usually the problem is the contacts get burnt, and have a problem connecting at times.
It's easier to just swap a relay with the same numbers from a different circuit, than it is to try and catch that happening.

The problem doesn't occur with the engine cold, but so far it has always taken place with it fully warmed up. When I mentioned the car being parked, that was always after a drive with the engine fully warmed up.

 

[Nukeman7]

@Mechanician
Can't the relay get tested. I think it can. If yes and it isn't a complicated procedure I can test it anyway.

Like Mechanician suggested, just swap the relay with an identical one in the fuse box.

@vwmaniaman
Measuring the air gap is something I will probably do. I have to find the specs first though. By the way, it can be measured with the compressor installed on the car, right?

It depends on the amount of clearance that you have on the vehicle. The Lancer is FWD with a transverse engine, so you should be able to access it from the wheel well opening after removing the splash shield.

@Nukeman7
I guess I will start by measuring the air gap first. Especially if it's something that can be done without uninstalling the compressor there is no reason to not do it as a first step. As I said, I have to find the specs for my compressor though.

It depends on the style/design of the A/C clutch on your compressor. Unlike the one shown in the video, some face plates have a recessed shoulder that prevents inserting a feeler gauge into the gap. Most air gaps are ~ 0.02 inches; however, you don't really need to measure it to perfection. I just set the air gap as small as possible where the outer pulley won't drag on the face plate when the coil is de-energized.

Also, check out this thread from a Mitsubishi Lancer forum:

Mitsubishi Lancer A/C Clutch/Relay

 

[Chris142]

@Nukeman7
I guess I will start by measuring the air gap first. Especially if it's something that can be done without uninstalling the compressor there is no reason to not do it as a first step. As I said, I have to find the specs for my compressor though.

every clutch I ever saw was .018-.028.... when it does it tap the outer clutch hub with a stick. See if it engages

 

[inquirer]

@Nukeman7

The access to the compressor is pretty much unrestricted. Lancer's engine is very spacious. I was asking if there is any other specific reason to do the measurement on a bench. If not then there is no problem, there is more than enough room to do it on the car.

 

[Nukeman7]

@Nukeman7

The access to the compressor is pretty much unrestricted. Lancer's engine is very spacious. I was asking if there is any other specific reason to do the measurement on a bench. If not then there is no problem, there is more than enough room to do it on the car.

It is really best to measure the air gap while installed in the car to eliminate any other variables (however, remote they may be).

 

[inquirer]

Here are 2 photos with the clutch disengaged and engaged.

I suppose that the gap gets tighter towards the axle of the clutch, right? Otherwise if we are going to judge it from the circumference of the clutch it probably is out of specs.