92 Legacy Jerks When AC Is On

[ralphl]

About a month ago my 92 Legacy (manual transmission, non turbo) began to jerk, when the AC is on. When the AC is off, I never get the jerk. At first, when driving, the jerk made me think I had a bad spark plug wire or even a bad coil (causing 2 cylinders to malfunction), the jerk was so severe. However, then I noticed that the jerk happened only when the AC was on. Has anybody else had the AC cause the whole car to jerk?????????

Normally, when driving, I never noticed when the AC compressor cycled on and off, it happened so smoothly. If I was idling at a stop light, I could notice that the tac would show the engine speed up to about 1100 rpm,when the compressor came on, and drop to about 650, when the compressor turned off. At idle the compressor mostly STILL smoothly cycles that way with no jerk on the engine.

I have tried several things. First,I checked the sight glass on top of the filter drier. Saw no bubbles. Waited for the compressor to cycle off. Saw a few bubbles when the compressor was off, but none when it cycled on. AC still gets as cold as it always did, so I don't think I am low on R12.

I unplugged the connector on top of the compressor so the compressor would not turn on. I drove around and experienced no jerks, so I think the jerk must have to do with the compressor cycling. So I reconnected the compressor so it would run for the following tests.

I swapped the fuse box compressor relay (right most relay looking at car from driver's side) with the one next to it. Drove around; car still jerked. So I don't think its a bad fuse box relay.

I unplugged the bypass air valve (idle air valve) and drove around. Car ran rough, but still jerked. So I don't think the bypass air valve is involved in the jerk. I reconnected it.

I unplugged the throttle sensor and drove around. Car ran rough, but it still jerked. I also ran all the voltage/ohm checks specified in my old Chilton paying particular attention to throttle closed/open lead. The voltage turned off and on as it should, when I manually opened and closed the throttle. I also checked the rheostat wiper lead to see if the voltages were in spec (full open and full closed) and checked to see if the voltage changed smoothly with the opening and closing of the throttle. Everything checked out fine. So I don't think the the throttle position sensor is involved in the jerk. Reconnected throttle sensor.

I have considered the possibility that the compressor clutch could be grabbing and causing the jerk on the engine. However, I never get the jerk on the engine, when the engine is at idle and the compressor cycles. Is it possible that the clutch grabs only when the engine is at driving speed?

While I know that the compressor must be running for the jerk to occur, I don't know if the jerk occurs when the compressor turns on or when it turns off. Maybe the ECU, in trying to control engine speed, sends out a signal to some place (injectors are all I can think off) to cut engine speed and then decides it too slow and immediately sends more gas to the injectors causing a jerk.

I have been reading 3 threads on this website dealing with AC compressor cycling problems. These are:
viewtopic.php?t=36434
viewtopic.php?t=32066&highlight=thermostatic
viewtopic.php?t=40114
They discuss problems with the evaporator thermostatic switch and the evaporator cut relay switch both located above the glove compartment and on top of the evaporator box (note that I have not yet looked at them). As I understand it the thermostatic switch is attached to a thermocouple or thermistor attached to a fin in the evaporator box. This sensor switches on and off according to temperature and feeds its output to the ECU. The ECU uses this information (along with other information such as the pressure switch output and the throttle position sensor) to cause the cut relay to turn the compressor on/off. Does anybody think one of these switches could be the problem? Also, have I correctly described the workings of these 2 switches?? The only way I could see these switches causing a jerk is if they "chattered" causing the compressor to very quickly turn off, then on, then off again and the ECU to feed more, then less, then more gasoline to the injectors for engine speed control.

Also, I see only a single pressure switch (on top of the filter drying. I presume this is a high pressure switch. I looked for a low pressure switch that would be used to say when the R12 was low. However, I couldn't find one on the low pressure pipes. Is it possible there is a low pressure switch under the dash, inside the compressor, or some other place. Maybe there is no low pressure switch???? If there is one, maybe it could be causing the problem???

[Legacy777]

Welcome to the BBS.

I moved your post to the engine forum.

What AC system do you have, Zexel or Calsonic?

Even if your AC is still blowing cold it's possible you could be low on refrigerant and as the compressor rpm's increase the suction pressure lowers quite a bit putting more load on the engine, and that could be the jerking you're feeling.

Could you describe the jerking a little more? If you accelerate lightly in 3rd or 4th gear do you experience the jerking or is it more prevalent when the engine rpm's are higher?

[ralphl]

Josh, Thank you for moving my post to the correct spot. I had some difficulty learning the web site.
I have a Calsonic ac system.
The car will jerk in 3rd or 4th gear even if accelerating only lightly. Sometimes I can even feel a slight jerk, when going mildly downhill.
I discovered some more about the jerk. First it happens when the compressor disengages, not when it engages... Surprise!
Second, I found that if I pull the leftmost of the 4 ac relays (looking at fusebox from driver's side) and drive, the car no longer jerks. The wiring diagram I have says that is the sub-fan relay and upon closure of the relay, it appears to drive the passenger side fan. The passenger side fan never seems to turn on, even when the engine is hot. So I unplugged the connector to the passenger side fan, put the relay back in, and the car still jerked. What the heck could cause that.
I took the passenger side fan completely off the car. It is sitting on my workbench. The plug for the fan has 3 leads--one ground and two that should have 12 volts to drive the fan. I presume the fans are each 2 speed fans, because of the 2 leads. With the fan on my workbench I connected a battery (negative to ground lead, and positive to one or the other of the 12 volt leads). Fan did not turn on. I can easily turn the fan by hand so it isn't frozen. My conclusion is that I have a bad passenger side fan motor. However, because the jerk still happens, when the left most relay is installed, but the fan is disconnected, I conclude that the passenger fan has nothing to do with my jerking problem.
So now I am really puzzled and am grasping at straws for how to go forward.

Any help is appreciated........

[arse_sidewards]

The passenger side fan never seems to turn on, even when the engine is hot. So I unplugged the connector to the passenger side fan, put the relay back in, and the car still jerked. What the heck could cause that.

The fan is locked up solid (more likely) or has some internal electrical fault (less likely). It's not drawing enough current to blow a fuse or light stuff on fire but it's drawing enough current to add a pretty substantial amount of drag to the alternator which itself has a 3:1ish advantage over the crank because of the pulley ratio.

[ralphl]

Thanks for the response... I appreciate it.
No, the fan isn't locked up. It turns freely by hand. Moreover, I took the passenger side sub fan out of the car, and the car still jerks with the sub fan relay (left most relay) installed. It does not jerk, when I take the sub fan relay out. That's the big puzzle. How could the relay for the sub fan (but not driving the fan) cause the car to jerk?
Also, the air conditioner wiring diagram is at http://ww2.justanswer.com/uploads/2paul ... 42_FAN.pdf

Thanks again..

[ralphl]

Complete address for wiring diagram for air conditioner is: http://ww2.justanswer.com/uploads/2paul ... 42_FAN.pdf . This is needed if you just copy and paste to bring up the wiring diagram.

[arse_sidewards]

Disconnect the alt or pull the belt and see if the problem persists.

[ralphl]

Sideways: Thanks for the suggestion. I will try that. Right now I am going to take a 2 week vacation and will not be able to try it until after that, but I will. I will get back to you in 2 weeks.

[Legacy777]

I still recommend you get a set of gauges and look at the pressures. If the differential pressure across the compressor (due to low refrigerant) is too high the compressor could be drawing excessive power from the engine and causing the "jerking" you're experiencing.

[arse_sidewards]

Gauges are pretty cheap. Anyone maintaining an old A/C system should have a set

[ralphl]

Thank you both for responses Josh and Arse_sidewards. I appreciate it very much! Well, I am back from vacation and had a little time to work on my Subaru ac and tried (and screwed up) what you guys said to do.
A. I unplugged the alternator and drove. Most of the jerks on compressor cycle shutoff went away, when I did that. I still noticed a couple of jerks, but they may have be caused by something else. So I think you are right that the alternator is what is jerking the car. However, I still haven't found out what the load change on the alternator is causing it.

B. I have a friend with a set of a/c gages so I visited him. However, his gage hoses (both high and low pressure) are for 7/16" schrader valves. The legacy has a 7/16"-20(fine) low pressure schrader valve so we could hook to that. However, the high pressure schrader valve is3/8"-24(fine) valve, so we could not hook to that and find the high pressure. My friend also has an (about) 2 gallon tank of R12, so we hooked that up and tried to charge the system. We tried to monitor the sight glass, but it appeared clear like there was no refrigerant at all. So my friend turned the tank upside down, which I think caused us to over charge the system. Suddenly the compressor, which had been cycling every 3 minutes for so, began to cycle every second or two.
I put a volt meter on the brown-yellow wire leading from the pressure sensor to the compressor relay and sure enough it was the pressure sensor that was turning the compressor on and off. But I don't know if this was because the pressure was too high or too low. I pulled the compressor relay from the under hood fuse block and put a paper clip across the switched terminals to make the compressor stay on. The compressor groaned and suddenly a big puff of (I guess, refrigerant) came out the back side of the compressor, making me again think that we had overcharged the system. However, with a hose with a 3/8 schrader fitting I can't check the high pressure side.
C. Can anybody tell me where to get a gage hose (or adapter) that would fit the 3/8 schrader valve? I couldn't find one in any of the 3 auto parts stores in town. The gage end of the hose needs to b 7/16.
D. I tried to bleed off a little R12 on the low pressure side, but I didn't to empty the system and I don't know how much pressure should be there.
E. Does anybody know what the high and low pressure readings should be, when the compressor is running, and also when the compressor has cycled off.

[ralphl]

The sentence above that reads "However, with a hose with a 3/8 schrader fitting I can't check the high pressure side." should read "However, without a hose with a 3/8 schrader fitting I can't check the high pressure side."
The sentence above that reads " I tried to bleed off a little R12 on the low pressure side, but I didn't to empty the system and I don't know how much pressure should be there." should read " I tried to bleed off a little R12 on the low pressure side, but I didn't want to empty the system and I don't know how much pressure should be there."
Sorry that I mis-typed.......

[Legacy777]

It definitely sounds like you overpressured the system. What was the suction pressure that you saw?

The proper way to fix this is to have the system evacuated, recapture the existing R12, and then refill the system with the proper amount of R12 refrigerant. There will be a label on the radiator core support with the AC refrigerant amount, type, etc.

[arse_sidewards]

FWIW I wouldn't go tracking down expensive R12 when a propane-isobutane mixture costs basically nothing.

[Blulian]

R12 defiantly works better, but yeah finding some is impractical most of the time.
I have a R134a conversion adapters on my 91 and it is plenty cold.

[ralphl]

Just a little update on my problem. While on vacation my 92 Legacy sat unused for a couple of weeks. My friend who has ac gauges found that he had an adapter to go from a 7/16" to 3/8" schrader valve. We attached the gauges started the engine and the ac worked great! The pressures on the high and low pressure fittings were within bounds. The compressor cycled appropriately for the temperature and engine speed. I really don't know what happened to make the ac work again. I think it may have been the 2 weeks sitting idle. I noticed that, unlike some other ac systems, the Legacy ac was very slow to equalize the the high and low pressures after the engine was shut off. When the pressure was so high that the compressor immediately shut off after it cycled on, I think that we didn't wait long enough to let the system equalize. Hence the 2 week idle let everything equalize and the system worked again. That's my theory and I am sticking to it!!

Now that the ac works again I am back to the original problem that the car jerks when the compressor turns off.

[Legacy777]

I'm not sure what to say. Did you ever use the AC previously and not have this issue? If so when was the last time?

[ralphl]

Josh: Thanks for responding. Regarding your question about when the AC first started jerking. It was fine for 30 some years. About 4 or 5 months ago, I first started getting the jerk.

I have been further investigating the jerk. Because my passenger side sub-fan motor was bad, I put in a new one from the junkyard, and now both fans seem to work perfectly. They stop and start the way the wiring diagram shows. So I am back to trying to figure out how the ac sub-fan relay being in place, causes the car to jerk. I unplugged the sub-fan (passenger side fan), put a voltmeter on the wire leading to the relay, put the voltmeter under the windshield wiper (so I could see it when I drove), and drove around. With the ac sub ran relay installed the car jerked. The voltage, when the compressor was running on the ac sub fan wire, was around 12 volts. As soon as the compressor kicked off, the voltage dropped to near zero (jumped around a little) as it should. The car jerked.

How could a relay driving a piece of wire with nothing connected to it, cause the car to jerk????????????????????

With the ac sub-fan relay removed, of course there was no voltage on the sub-fan wire. The car DID NOT jerk.

Looking at the wiring diagram for the ac and fans I see there is a diode, which is connected to the Temperature sub-fan relay and the ECU. I don't know the reason for the diode, but it allows current to flow from the temp sub-fan relay to the ECU, but not from the ECU to the temp sub-fan relay. I am grasping at straws, but maybe a bad diode (with current flowing the wrong way) could cause the ECU to jerk the car.

The wiring diagram says the diode is the "corner of the left fender" I looked in both fenders, but couldn't find a diode. I did find a connector on the passenger side fender that is supposed to cause both fans to run. I haven't tried it yet, because I currently have the battery out looking for the diode. However, the wiring diagram says its on the left side. Maybe the diode is also on the passenger side. Maybe to access it requires removal of the passenger side fender liner, since I see the fan wires to in there.

Does anybody know where the sub-fan diode is and how to access it?????????????????

[Blulian]

Was is this "Jerking motion" ?
In my head I picture a rapid change in velocity of something in the drivetrain. Can you film it?

[ralphl]

Austin, Thank you very much for the response. I appreciate it. The jerk feels like the car had a bad spark plug wire and missed. Or maybe had 2 bad spark plug wires, since it is a fairly significant jerk. And it really seems strange that it does this, when the ac compressor cycles OFF, and that pulling the relay that powers the passenger side sub-fan ac mode prevents it from jerking. I have a wiring diagram of the ac and fan system that shows a diode in the left fender, that is the only thing I can think of that might cause this. However, I can't find it. Maybe it requires the removal of the fender lining to see inside the fender.

[Blulian]

So you would describe the engine runs rough and possibly is misfiring when the AC compressor cycles on?

[Legacy777]

Sorry for the delayed response.

Any chance you could take a video of the jerking occurring or have someone ride along with you take a video?

The diode is wrapped up in the wiring loom so it may be kind of difficult to find. You should be able to use a multi-meter with the function to check a diode to see if the diode is functioning properly. It should essentially just show continuity one way but not the other way.