2000 Subaru Legacy Automatic Transmission Problem? (Yanking Randomly)

[Janiporo]

This is 2.5 litre STW Legacy with auto transmission, yearmodel 2000. Euro model. Bought from Swizerland to Finland.

 

Recently I have been noticing that there is some kind of "yanking" (do not know the correct word for this, it seems a 0.2 second power loss and the normal again). It feels like it is changing gear but decides not to.When accelerating or speeding down there is no yanking, but when driving steadily on the flat road, this starts. It comes like 2-12 second intervals randomly (at 80 kmh / 50 mph). It gets more frequent when there is more speed.

 

Engine is running fine on idle.

 

Is there some kind of pressure sensor at the transmission that could cause this? Maybe bad connection or broken sensor? 

 

Rpm-meter does not fluctuate (is steady) when this happens.

 

There is also light yank at between gear change from 2 --> 3 but that has been there from year ago when I bought this. That is ok (just put here if it is another symptom concerning this problem)

 

Of course this could be something on the engine side, I do not know. The MIL-light does not light up.

 

I have smartphone with Torque program, but if I remember correctly, it does not read anything concerning transmission, only engine.

[Fairtax4me]

That sounds more like a misfire than a transmission issue.

 

How old are the spark plugs and wires?

[Janiporo]

Plugs have been changed about a year ago, wires have not, no idea about the age of wires. Why does it not do this when accelerating?

 

I might change the spark plug-wires and go from there. But any more help is appreciated 

[Fairtax4me]

Those are the usual culprits although the symptoms don't exactly match what I would expect from worn plugs. Typically worn/old plugs cause misfiring at low rpm under load or while accelerating like you said.

 

You might look into the things that affect fuel delivery. Replace the fuel filter if its old. If the O2 sensors are old you could be getting a false rich reading causing the ECU to pull fuel.

 

You should be able to see fuel trims with your app. Make sure trims aren't showing negative.

[Suzam]

Possible torque converter lock up issue? You should observe 4 shifts plus a 5th reduction in RPMs when the converter locks. It sounds like it's possible it is trying to engage but unsuccessfully.

Edited October 7, 2014 by Suzam

[Janiporo]

Fairtax4me: This is the screenshot of the Fuel trim status when the car started yanking. According to this site these are WAY too much on the positive side when calculated together --> http://longislandcatalytic.com/understanding-fuel-trims/

 

Which lambda does this measurement come from? If this is from the one that is after last catalytic converter, it is not measuring it correctly because there is this thing installed to remove mil light error code --> http://kuvapilvi.fi/k/yhAl.jpg

btw, that pipe works!

 

To be clear, Long time fuel trim 9,4%, short time fuel trim 10.9%, according to that side and calculated together it is 20.3% and it should be under 10% (tell me if I am wrong).

 

Driving uphill the fuel trim values dropped (especially LTFT), same at downhill. LTFT was like 3-5

 

So, is this running on too lean or rich?

 

And here is the screenshot:

 

 

Edited October 7, 2014 by Janiporo

[Janiporo]

Suzam: Torque converter is locking for shure, tested it yesterday. No slipping, and rpm stays exactly the same when this occurs. I think it is not that, what do you think, could it still be?

[Janiporo]

I read trouble codes with Torque -software, none available, the I read trouble codes with this method --> http://www.scoobypedia.co.uk/index.php/Knowledge/ReadingECUCodes and there was nothing.

 

This sounds kind of misfire, or that ecu cuts fuel for very short time. But why, is the question. Why does it not do it all the time?

 

I will check in following days the spark plugs and at the same time I will see if there is oil / some liquid around the park plug.

 

Interesting question, does ECU give more volts for the ignition coil when accelerating compared to the steady driving?

[Fairtax4me]

Positive fuel trims mean the ECU thinks the fuel mixture is too lean. Lean mixture, add fuel. Rich mixture, subtract fuel.

We would need to see what voltages the front O2 sensor is reporting to tell if its reading a lean condition.

The post cat (rear) sensor does not affect fuel mixture calculation. It's only there to check for operation of the cats. But having it pulled out of the exhaust stream with the spacer means we can't compare rear sensor values to front sensor values. Typically if a front sensor reads lean and the rear sensor reads rich, or vice versa, it's because one of them is fouled or damaged and is reporting a false condition.

 

 

If the front O2 sensor is old, you might try changing that. Before that, check over all of the PCV and breather hoses on the intake tubing, and the ones that to to the valve covers. Also make sure the PCV valve is clean.

Also check for any exhaust leaks between the engine and the cats.

The header gaskets where the y-pipe bolts to the head are prone to leak if someone has replaced them with cheap gaskets, or if the flanges were rusty and not cleaned properly.

[Janiporo]

Which lambda does the computer read fuel trim, one before, or after cat? The meter is set to --> "F/T 1x2" 

This one --> "F/T 1x1" is showing nothing as you can see from the picture above.

 

I must test all these hoses and PCV-valve if I can find the valve. Just googling where it is supposed to be.

 

What is this Y-pipe you mean? Do you have a picture so I would understand? Is it aluminium or plastic or rubber? Last year head gaskets were changed, we used Fel-Pro gaskets, I think they are good ones. Do not know about this Y-pipe, was it taken off when changing head gaskets? 

 

Thank you for your help 

[lmdew]

cylinder head > "Y pipe" is the exhaust pipe that bolts to the head on each side and joins at the front of the CAT

[Fairtax4me]

The bank 1 sensor 1 (pre-cat) is the one that affects fuel trim. I did notice the FT 1x1 does not show anything in the photo. That seems odd to me because the bank 1 sensor 2 (post cat) will not affect trim.

 

If you have a bank 2 sensor 1, (I'm not sure how many sensors the euro market cars have) then you should be able to view that.

 

I'm not familiar with the layout of the torque app though, so I'm no sure if there is a setting or selection that can/should be changed.

 

Something you may want to look at though is sensor voltage values, rather than just looking at fuel trims.

[Janiporo]

Tomorrow I will check also these lambda volt-values and be back here. It might be that first lambda is sending wrong signal. I found out that there are only two Lambdas, one before and one after cat.

[Janiporo]

Now I have tested it.

 

I forgot to mention, two weeks ago car was in emission test, and all emissions were in limits. It was tested with this tester where you put pipe in the exhaust. Do not remember what it said about O2 leaks, in limits anyway. For example HC was 0.000 

And the car was misbehaving then allready.

 

I learned that this lambda might be wide range voltage lambda, so I looked at that with my software and there is some values there. It is the top left one (O2S1LV) in the picture below (1,53 volts / highest 4.2, lowest 0.6) What is the range it should show?

(in Torque software this is marked as "O2 Sensor1 wide-range voltage) 

 

O2 1x1 -meter does not show anything (in the middle of the picture)

 

Yes, there are some meters that measure the exactly same things, I did not remember which meters I had put on there so there are some duplicates.

 

I have checked almost all of the connectors in engine bay and put some correct CRC to them and blew with air to get excess out.

 

I cleaned spark plugs, there were some drops of oil beside some of the plugs but not enough to have any effect to anything.

 

There was an minor air leak in the hose that vent from intake manifold to fuel pressure-regulator, that I fixed but it did not help for this problem.

 

Any ideas?

 

Edited October 12, 2014 by Janiporo

[Fairtax4me]

Low HC is good, 0 HC means its running lean.

 

Even with a lean mixture the HC will never fully burn during the combustion process. The catalytic converter takes care of the small amount that's left over by combining it with oxygen that the cat stores from the lean exhaust pulses that travel through it.

Your cats are probably working, or that HC number would have been much higher.

 

I'm going to suggest that you remove the anti-fouler from the rear O2 sensor, then we can get an idea of the actual oxygen content of the exhaust leaving the cats.

 

It does appear that the front sensor is a wide range sensor. The normal voltage for that type of sensor varies depending on the design. The wide range sensors use a reference voltage from the computer, that's typically from 1.5-3.3 volts.

With yours showing a Max of 4.2v, and a minimum of 0.6v, the sensor output is varying much more than normal for a wide range sensor. Wide range sensor outputs typically don't change more than 0.5 volts in either direction. But that could just be showing the range that the scanner software is able to display. How much does that signal vary while the car is running?

[Janiporo]

I am going to take this next week to one repair shop (I have a week left to figure this out on my own), they will remove the anti-fouler for time beeing, and test the car with their testers. If no luck, I will take this to the one Subaru specialist workshop that was recommended. 

 

After the car is fixed, the anti-fouler will be put back.

 

I just went to test drive and came up with these values. Wide range voltage was almost steadily around 1.5 volts, but when breaking with engine it went as high as 4.3, is this normal? This has automatic transmission.

 

The weird thing was that it did not do the yanking at all first, we had to drive like ~5 kilometers / ~3 miles before the symptom was there. First test drive at night, only thing I can think of is that the engine was quite cool and air moisture was WAY up (foggy) at temperature of 8ºC / ~46ºF. It is almost the same temperature at the daytime (when we did the last test run), about 12ºC / 54ºF

 

I added the MAF-sensor output.

 

Edited October 12, 2014 by Janiporo

[Janiporo]

Oh My god, this can not be this simple. Back tyres were deflated over 50%... 

 

There was also other symptom, ABS-light would randomly light up, it is easy to reset, just restart engine, BUT after we refilled the tyres --> ABS-light did not come back, and no yanking anymore. (Tested two weeks now at same weather than before)

 

Front tyres were in normal pressure, 2.2 Bar, rear left was 1.0 Bar and rear right ~0.9 Bar.

 

My theory is that Traction Control did not work correctly as the wheels rotated at slightly different speeds, and it activated randomly. Does anyone have a better idea why this happened?