[XP] Sporadic aircon

incorrect gas quantity will give over high or over low pressures, either will turn the system off. get the system vacuumed and properly refilled.

Colin:

I can't convert bars to psi (perhaps you can help) but 7.3 vs 7.1 is too small of a difference to be the cause. It is interesting that it changes but that kind of a pressure change is very small.

Also, the readings that matter are the high side and the low side, measured at the fittings. If you get a set of gauges and some long hoses I think you could read this while driving.

Have you considered a possible mixing valve issue? Maybe the A/C is working but warm air is being sent to you?

running: a high of 10 -15 bar and a low of 2 is about right. the standing pressure will always be the same related to ambient temperature. 5.6 bar at

25 degrees c

But the system isn't turned off - in the failed state, the compressor relay is pulled in continuously. Also, the pressures seem to tally with the spec for the system. The fan supposedly cuts in on speed 1 at

9 bar and goes to speed 2 at 18 bar. The compressor gets shut down below 2 bar and above 24 bar. According to that, I have a maximum pressure somewhere between the two fan speeds, and a minimum well within spec.

The aircon service I had done did purportedly include a re-gas. I suppose they might have botched it...

Cheers,

Colin.

It was that closeness that made me think the compressor might not be running. i.e. the pressure during my failure mode is nearly the same as when the system has been turned off for hours.

I might end-up doing that before I've sussed out what's going on...

Yep. It's definitely the evaporator warming up. The system passes a self-test which includes exercising the flap-valve motors, and the windows mist-up during a failure.

Cheers,

Colin.

Colin Stamp wrote on 02/09/2007 :

Just a suggestion....

On the evaporator coils there should be a temperature sensor. The sensor is there to check the temperature and make sure it does not fall below 0deg C, such that ice can form. Could it be that either this sensor is faulty, or that it is detection the formation of ice?

Shortly after the a/c stops working, do you perhaps see a discharge of water from the drain?

Most likely a knackered clutch on the compressor. Connect a test lamp from the clutch power wires on the compressor, and see if it's still getting power when the system stops working.

The anti-freeze sensor was top of my list when I only had the pressure readings to go on, but since I put the LED in, it turns out that the system is trying to run the compressor all the time during a failure. None of the sensors or the control system are suspect any more.

Cheers,

Colin.

then the clutch probably needs attention

Could be. The next part of my cunning plan is to check the clutch is pulling current.

I'm still wondering if there might be a refrigeration problem which could cause the high-side pressure to intermittently drop to a "switched off" level even with the compressor running though.

Cheers,

Colin.

That's my favorite at this stage too. I'll eliminate the rest of the wiring first.

If only it would fail whilst parked with the bonnet open. I feel a strange urge to bodge in a video camera, pointing at the pulley, with an LCD telly on the dash :o)

Cheers,

Colin.

moisture in the system can cause this, it freezes to form a plug, then melts etc.

But when you get a blockage in the system high pressure goes through the roof and low pressure drops very low (or even to vacuum if the switch doesn't cut the compressor off). His pressure is going back to an "at rest" pressure.

You need to know if the compressor is turning or not when the problem happens. The LED is a good idea but an LED will light with virtually no current. A compressor clutch needs a few amps to pull it in. If you have a poor connection / bad switch somewhere in the loop it is possible for there to still be enough oommmppf to light an LED and not pull in the compressor. Kind of like having a dead battery... The interior lights will still come on but the starter won't spin.

Steve B.

I have news!. It's the clutch for certain.

I finally managed to get the thing to fail whilst parked today, by letting it idle for ages on the drive after a longish run. It stayed failed when I opened the bonnet too, so I could finally check stuff out properly.

The compressor isn't turning even though there's 15mV across its fuse, which is the same voltage as when the syetem works properly and the compressor is on. So the clutch is definitely pulling normal current, but it isn't engaged.

Right. Now I need to fix the bloody thing. A quick Google suggests that I might be able to fix it by taking some shims out to compensate for the wear. I suppose the next step is to take the clutch apart and see if that's possible.

Anyone done this on a 9-3 before?

Cheers,

Colin.

No, but I bet its similar clutch to a Volvo 70 series, which also need periodic re-shimming when the air gap becomes too great for the magnet to pull in the clutch.

You'll be needing a clutch puller to do it.

Tim..

Thanks for the tip. I haven't identified the compressor yet, but the service manual says it's a Sanden TRS105R 3211. I think I'll try to dig around a bit further to see if I can find out what adjustments are possible and what tools I need...

Cheers,

Colin.