yet another tranny question

[UberRoo]

As I started reading the symptoms I felt kinda stumped at first. Then I thought about it. It only slips during that transition, right? Once fourth gear is fully engaged, it never slips, right? It's pretty important that these two statements are fact because they're the rails that my train of thought is riding on.

Basically, the clutch is not slipping, there's simply a delay in the transition between gears. Shift solenoid one is off in both third and fourth gear. Shift solenoid two is off in third, but on in fourth. This means to me that whatever happens after solenoid two is given power is taking too long....that, or solenoid two isn't getting power like it should. Maybe a bad connection?

I'd start with this simple test harness. You need two 12-volt light bulbs or 12-vold LEDs, two T-tap wire connectors, a ground source, and about ten feet of wire. Under the dash is the TCU and harness. Find the wires that control solenoid one and two and put a T-tap on each of them. (You'll need a wiring diagram. If you decided to do this, ask me and I can tell you which wires do what.) Run a wire from each T-tap to each light bulb, and then a ground wire from the bulbs to wherever. (There's also a ground wire in the TCU harness you could tap with a third T-tap.)

These two bulbs will light up in a different combination depending on what gear you're in. My tranny is working perfectly, but there's still a pretty lengthy delay between the time the solenoid is triggered (when the bulb lights up,) and when the shift finally happens. If the bulb for solenoid two does weird things, you've found your culprit.

Whether you do this test or not, check and wiggle the connection to the TCU under the dash. There's another connection under the hood. It's on top of the transmission right below the throttle body. There's actually two connectors there, one for the solenoids and sensors, and the other is for the gear indicator. A weak connection may be your headache.

Back to the topic of what may be causing the delay in the gear transition, I suggest two possibilities based off the fact that the only mechanical reason the clutch would not be engaged immediately is because there is no hydraulic pressure to activate it. 1) The pump is failing to produce enough flow possibly due to plugged passages. 2) The shift solenoid valve isn't opening up entirely, restricting fluid flow. In both cases there's a flow restriction. Unplugging the dropping resistor is an important step in diagnosis. If you unplug it and the problem goes away, that tells you that the problem is much more likely to be mechanical; maximum pump output can overcome fluid restrictions. If unplugging it does nothing, I'd say you're looking at some buggy electrical problem. Unfortunately, this is very speculative, but it gives you a place to start.

I've never checked to see, but I've been told, (and suspect it's true,) that there's a pressure port on the tranny for the first/fourth gear clutch fluid passage. I've always thought that the pump pressure and flow rate was far too anemic for this engine and transmission combination - simple problems from a minimum competency design. Also, it suddenly occurs to me that the internal filter/screen may be plugged.

Do you know if the Subie transmissions can handle the typical stall speed test, i.e., stand on the brakes & gas & note engine rpm? That "should" tell me if it's a TC issue.

Also wondering about the servo that tightens the band for 4th - is that a different one from 2nd?

Absolutely! Same answer for both questions actually. Unless your transmission is suffering from overheating problems or you're getting slippage in first gear, the stall speed test is harmless. The test should only require a few seconds, so as long as nothing slips and you don't carry on like that for more than say, ten seconds, you're fine. The servo for second is different than the one for fourth. Here's the control pattern:

Gear 1 = clutch/solenoid 1 on & 2 on (both)
Gear 2 = clutch/solenoid 1 off & 2 on
Gear 3 = clutch/solenoid 1 off & 2 off (neither)
Gear 4 = clutch/solenoid 1 on & 2 off

[lee]

UberRoo,

You've hit a key point I didn't explain - my bad definitely.

No, the gear change works well (or at least I think it does) - I can feel a definite change up/down in gear. The slip occurs anytime the car is in 4th, not just at the time of the gear change.

Note: I have not done extensive testing of the slippage because this is my daily driver, so there may be questions I can't answer as I've not tried too many routines that induce the problem. Here's the scenarios I've seen so far:

1-a) Pulling away from a traffic light (tranny in D) at the usual rate with car in front of me, speed limit @ 45mph. Car shifts normally from 1st up thru 4th, except switching from from 3rd to 4th I get an extended slip before reaching the 45mph.

1-b) Alternately, I give it even more gas and it downshifts back into third, until I let off the gas. Finally, here's how I drove it after a few sessions of this - hoping to preserve what I've got.

2) I put it in 3 and accelerate up to 45-50mph. I have my foot on the gas just enough to maintain speed, and I put it in D. I can feel the shift take place, and the rpms drop down like everything was OK. As long as I use only enough gas to keep up with traffic, i.e., mild slowing/acceleration everything seems fine. However, if I now give it say 1/3 throttle, to change lanes and get around one of our many senior drivers here in FLA, I can feel the slip.

BTW, when I say I can feel the slip what I'm saying is based on 20K previous driving miles and how much rpm I should be carrying & how much the car should be accelerating at a given throttle position. The engine/transmission has never slipped freely like there was nothing engaged. It's more like revving to 4K rpm - instead of holding around 3K rpm.

So, needless to say I've not tried a full throttle run to see what happens, or with heavy throttle just how much it will rev to when the "slip" occurs, etc. I have also not had the car over 60mph since the problem occurred. I will try some of those things if it will help someone diagnose my problem, but keep in mind this is my daily driver and I have to switch to a 29 year old motorcycle when my car is down.

Edit: Also, the TC will still achieve lock-up once the fluid is warm enough. Even with this problem, the fluid has never exceeded about 190 degrees F. I'm pretty confident on that temp as TC lock-up first occurs when the gauge shows around 150F.

[lee]

one more thing fellow forum members....

my wife gave me heck when I fist bought this car...she wanted me to get a Supra or a Lexus SC300. If it's the tranny on it's last legs I'm gonna catch it since it's only been two years since it was rebuilt. I don't beleive she would let me rebuild again without a LOT of grief. My only hope then is to go to a 5-speed - that might work. Problem with that is this is my daily driver - have no idea how quickly the swap couldbe made, what with few Subaru in the area (non in local wrecking yards to source manuals from), crossmember issues, etc..

[UberRoo]

Hmmm... I think you may be right on track with the stall speed test, but I'm still not quite certain I understand what you're experiencing. In the scenarios you describe, you consistently mention third gear, but you say the problem exists when you're in fourth gear. To clarify in my mind what's going on, let me ask you this (and I'll assume for the rest of this reply that the answer is yes.) When you have been in fourth gear for say, more than fifteen seconds, and when there is NO third gear downshift, have you experienced this slippage? (Those test lights would be useful, because I'd simply ask you if the light for solenoid one ever went out or flickered.)

Again, assuming that your answer is yes it has slipped at some point well after the fourth gear shift, we're on a very different diagnostic path. My first instinct screams, "unplug the resistor!" That will narrow things down so much. My second instinct is that your are simply experiencing slippage from the fourth gear clutch. Major bummer, but maybe not so bad because the clutch may be fine. It wouldn't surprise me if the filter/screen is plugged and maybe causing the pump to cavitate and suck in air. Also, a loose electrical connection is still a good suspect. I had a buddy who's car would miss in reverse because the engine tilted under the torque of the wheels and pulled on an electrical wire which controlled the ignition. A week later it did the same thing in first gear because the engine tugged on a cracked vacuum line.

Another big question: Does this happen when the torque converter is locked up?

I'd see how far you can rev the engine when it slips...just once. It should only be a few seconds of slippage. Take a good wiff of your tranny fluid too.

Sounds like you have some test driving to do.

[lee]

Originally posted by UberRoo
Hmmm... I think you may be right on track with the stall speed test, but I'm still not quite certain I understand what you're experiencing. In the scenarios you describe, you consistently mention third gear, but you say the problem exists when you're in fourth gear. To clarify in my mind what's going on, let me ask you this (and I'll assume for the rest of this reply that the answer is yes.) When you have been in fourth gear for say, more than fifteen seconds, and when there is NO third gear downshift, have you experienced this slippage? (Those test lights would be useful, because I'd simply ask you if the light for solenoid one ever went out or flickered.)

Again, assuming that your answer is yes it has slipped at some point well after the fourth gear shift, we're on a very different diagnostic path. My first instinct screams, "unplug the resistor!" That will narrow things down so much. My second instinct is that your are simply experiencing slippage from the fourth gear clutch. Major bummer, but maybe not so bad because the clutch may be fine. It wouldn't surprise me if the filter/screen is plugged and maybe causing the pump to cavitate and suck in air. Also, a loose electrical connection is still a good suspect. I had a buddy who's car would miss in reverse because the engine tilted under the torque of the wheels and pulled on an electrical wire which controlled the ignition. A week later it did the same thing in first gear because the engine tugged on a cracked vacuum line.

Another big question: Does this happen when the torque converter is locked up?

I'd see how far you can rev the engine when it slips...just once. It should only be a few seconds of slippage. Take a good wiff of your tranny fluid too.

Sounds like you have some test driving to do.

thank you very much, you've given me multiple good ideas and avenues to test. I'm hoping that one of the easier scenarios, i.e., bad connection or clogged filter is the issue - need a smilie for keeping my fingers crossed

only one avenue I eliminate off-hand - the fluid. It smells/looks fine - with no fluid lost. That's one reason that makes me hopeful it isn't the band slipping - but then not sure how much slipping it would take to get that burnt up ATF smell/look.

edit: Yes, it does slip when the TC is locked up.

[huck369]

Originally posted by lee


edit: No it does not slip when the TC is locked up.

then it almost has to be the TC...at least as I see it

[lee]

OK, I know a little bit more now. But before I begin to update the issue can anyone confirm the text from this site: http://www.babcox.com/editorial/bf/bf90222.htm

"Before conducting a stall test, check the fluid level and condition. Chock the wheels and set the parking brake. Start the engine and place the transmission in drive while holding the brake pedal under firm pressure so the vehicle doesn’t move. Then push the accelerator to the floor while holding the brakes on. Note the maximum rpm the engine reaches. This is the stall speed. If it is lower than specifications, the torque converter one-way clutch is slipping. If the stall speed is higher than specifications, the transmission is slipping. Possible causes include a low fluid level, restricted fluid filter, a sticking pressure regulator valve, slipping clutches, bands, shaft splines or one-way clutch."

I had to run an errand this AM and used my SVX. I removed the dropping resistor before driving and things seemed a bit better, but not fixed. However, since this wasn't a back-to-back test (break in driving from Friday to Sunday) I view the results to be mixed.

When I got home (engine & tranny all warmed up) I ran the stall speed test, i.e., stand on brake pedal, WOT for 5 seconds or less and what the tach. I got slightly less than 2250 rpm. From searching on the forum I conclude the correct stall speed is around 2700 rpm - can anyone with a FSM confirm that figure?

If the above quote is correct and 500 rpm difference is significant (is it?) then Huck's guess looks like a winner. That will be a lot cheaper. Will take lots of my labor, but hey, keeps me off the streets and out of trouble.

edit: Assuming it's the TC, will a TC from a Legacy, etc work or do I have to run down a SVX TC? If I get one from a junk yard should I just carry it straight to a rebuilder shop, or would you give it a try as is (assuming it looks clean, etc?

[huck369]

The Legacy TC won't bolt up to the SVX flywheel (Flexplate), so to use a Legacy TC, you need a Legacy Flexplae too.
I say have your SVX TC rebuilt, I had one done, it was $185 (The best I remember), plus labor, if you have someone remove and re-install it.

I don't have a FSM for the SVX, so I can't answer you're correct stall speed question

[UberRoo]

I'll check my stall speed here in about an hour, but I don't think 500 RPM is particularly signifigant. Moreover, I can't imagine how being 500 RPM lower can indicate a problem in the transmission. Maybe your tranny fluid generates more resistance in the torque converter, or maybe your engine just isn't pumping out as much horsepower. I kinda doubt both those scenarios, and I think 500 PRM is within specification. If something inside the transmission is slipping, your RPM should rise quickly and pretty significantly, probably by a minimum of 1000 RPM, and very likely all the way up to redline (unless, despite no wheelspeed, the TCU forces a shift to prevent overspeed, thus using a non-slipping clutch.)

I haven't read that article yet, but I'll give you my thoughts on it when I do. The SVX tranny is a newer design, and it's possible that article was written for older designs. (The primary difference I can think of is how the overrunning clutch is used.)

I would definitely not suspect the torque converter after that test.

---------------------------------------

Okay, my stall speed is about 2550 with warm fluid. (That's a pretty important factor.) That article seems to be about right. Try though I might, I couldn't find anything that positively confirmed that a low stall speed is caused by a faulty torque converter. I found two references that suggested it was possible, but one was referring strictly to diesel transmissions, and the other strongly urged that all engine performance issues be absolutely resolved before replacing the torque converter.

So, yes it may be the torque converter, but it sounds like a pretty rare problem. A more common failure of the torque converter results in a higher stall speed. What really prevents me from hopping on board with the torque converter theory is that the engine slips, or revs higher on the freeway, but bogs down, or revs very low doing a stall speed test. That it would do both just doesn't make any sense to me. Maybe there's an explanation that includes a torque converter problem, but I'm skeptical and I can't imagine how. I'd really like to know what the solenoids in your transmission are doing.

* edit: No it does not slip when the TC is locked up.

* then it almost has to be the TC...at least as I see it

I'd love to think that settled it, but remember that once the torque converter is locked up, it no longer multiplies torque and it lowers the engine RPM taking it out of it's power band which would prevent parts from slipping. In addition, the TCU basically doesn't and won't do anything while the torque converter is locked up. That means a buggy TCU is effectively turned off and won't exhibit any problems as long as it's in that mode.

Gosh, I wish I wasn't running out of ideas because I really don't want you to drop the tranny if you don't have to, and even worse, replace a perfectly good torque converter. Unfortunately the only remaining ideas yet untried aren't as much fun as theorizing about it.

In the order I'd start troubleshooting:
1) Check the connections
2) The next time it slips, see how high you can get the engine to rev.
3) Hook up the test lights to the TCU harness.
4) Change the internal filter/screens.
5) Check the 1st/4th gear clutch pressure port while driving.

Also, does it feel like the transmission suddenly releases a clutch, or does it feel like it gradually begins to slip? Is this an intermittent problem, or does it happen every time?

[lee]

I checked connections, they seem OK, but I admit I didn't check continuity at each wire, but they are clean and secure.

During the time it feels like it's slipping, I can't get the engine to go over about 4K rpm before a downshift occurs. A couple of times it seemed as if the transmission hooked up and slipping stopped without downshifting, but traffic was bad enough and I was too slow-witted to actually downshift to see if it went to third or just hooked up solid in 4th.

I was able to find several references that imply a lower than normal stall speed can be a result of stator or one-way clutch slippage in the TC. Here's one from Subaru:

http://www.ravensblade-impreza.com/t...ssion/4eat.pdf

What I don't know is what the specs are for the stall speed and how far out it can be from specs and still be quasi-OK. Online I found Mitchell's data for VW and Mercedes. VW spec for replacing the TC is only 200 rpm below spec, while Mercedes is 4-500 rpm. I guess I'll call a few places tomorrow and see if anyone will give me the info over the phone (expect most will want me to come in and give them $$ - freebies are getting rarer & rarer).

That said, I'm out of time this weekend, so unless it gets worse, I'll try to work up a light package during the week to hook up to the TCU as suggested by UberRoo. Will also order in the pan gasket/internal filter and give that a whirl - probably next weekend as my HQ is visiting this week to "help me do my job better."

I also need to pay some attention to rpms in the lower gears. I think if it is a slipping stator/one-way clutch, then it should be doing it in all gears (although I can see why 4th would be worst, what with the gear ratio/torque thing).

Thanks for staying with me on this folks, I appreciate this quite a lot.

[gl1674]

Lee,

FSM spec for stall test is 2350-2750 rpm. My car gives me 2400 rpm.
Lower reading may indicate an engine problem...

I think I've been through exactly the same problem myself.
The good news - my transmission is still alive 1.5 years later.
The bad news - solenoid pack is not cheap, $250 and I did the replacement myself.

My symptoms were exactly as you described. Stall test was normal in all gears. I did not have any error codes and all tests except line pressure were normal. The culprit was worn duty Solenoid A - it would keep line pressure too low and that caused a clutch slippage.

My advice is to get the transmission line pressure checked immediately. With this slippage the tranny is in self-destruct mode, you can't drive it like that.
You should see 60psi at idle and jumping up as soon as throttle is applied. With faulty solenoid A it stayed at 60psi all the time.

Disconnecting dropping resistor may get the pressure back up and eliminate the slippage - at least this is something you can try straight away without tools/lifting the car.

The line pressure is controlled by TCU based on the TPS input - the source for TPS code have to be resolved before you can meaningfully measure the line pressure (if you get low reading, it will confirm that the tranny is slipping, but you would not know whether it is because of TPS or Solenoid A).

The TCU code is thrown whenever the signal from TPS is too low or too high. The normal reading at idle is 0.5v, if you move it to below 0.1 v you will get the code all the time. Poor connections (either in the harness connectors or in the TPS itself) will throw the code intermittmently.

[UberRoo]

I agree with gl1674. A worn solenoid would certainly do it. I would have assumed a jammed shift valve or plugged fluid passage, but either would amount to the same thing. You know what they say: If it's happened before…

I would think that with the resistor unplugged, the line pressure test would still be valid even with a faulty throttle position sensor. Ideally it should be working, but with the resistor unplugged, solenoid A should be getting a signal for maximum pressure, and should it fail to perform, it won't be the fault of the TPS. (Solenoid A would be the last link in a very short chain. The only other link would pretty much be the wire that runs to it, and if that were bad, the TCU would let you know with a trouble code.)

$250 for a solenoid pack sounds better than dropping the entire transmission, better than replacing the torque converter, better than a clutch rebuild, and better than convincing the wife that you need a five-speed. ...unless she agrees of course. I sure wish my stick went left to right as well as back and forth.

[lee]

Some people say cars are just machines. I'm here to say they aren't, and in fact have a secret life all their own.

Story time:

I decided to take the afternoon off and just got home from visiting the place that did the rebuild for me. Since the warranty expired a year ago, I'm assuming they wouldn't send me away if there was money to be made.

Anyway, tech took it for a freebie road test. He came back and said he tried it all, including manual button runs, and found nothing wrong. No slippage, nothing.

Now admittedly I had the dropping resistor unplugged, but I was still seeing some symptoms during the drive to their business. Post the road test, the car shifts almost like it did with the new rebuild. It's the "secret life" thing. Cars always know when you have a hard to diagnosis problem and never misbehave in front of the technician.

If I'm getting around 2200-2250 stall speed, that's not far off the lower value of 2350 - doesn't seem enough to warrant buying a TC rebuild just for insurance. Also the tech told me he had never seen a one-way clutch fail that didn't cause more problems than I've got.

Maybe it is just the TPS. I do have that error I need to fix. It could well be that both of the TPSs I have on hand are bad. I checked late last nite (actually about 1AM today) and got the proper voltage at the TCU from the TPS and the voltage went up as I pressed the throttle (all this done with the engine off). I believe I'll tap into the TPS line and monitor it for a while with the VOM.

I did stop by Radio Shack and buy a couple of miniature lights and a little box so I could set up the solenoid monitoring arrangement.

I remain thankful for all the help.

[oab_au]

Lee when you check out the TPS, check the resistance through the full sweep of the wiper. Hook up the meter either OHMs or VOLTs, ignition on, engine off. Move the throttle shaft slowly by hand, watch the meter , look for a spot where the reading drops. The reading should change smooth, to show no bad spots on the resistance track.

This type of sensor is prone to the carbon tract etching away, due to the vibration that is transmitted through the throttle shaft. This sensor spends most of its life sitting in the one position, like when cruising at normal speeds. The vibration echting it away at that spot.

As the throttle is opened, the TCU reads the rising TPS voltage to increase the line pressure, in line with the increasing load, if there is a dead spot on the tract, the voltage drops and the line pressure drops also.

It is not the torque converter stator slipping or sticking as the stall test shows. A problem in this would happen in all gear positions. I would also disregard the A solenoid, for the same reasion. The shift solenoids don't actually shift the gears, they just operate the shift valves, other variables operate also.

The units that cause the different ratios are, the FORWARD CLUTCH, BRAKE BAND and the HIGH CLUTCH. The sequence that selects the ratios are;

1st. FORWARD CLUTCH applied.
2nd. FORWARD CLUTCH still applied, BRAKE BAND applied.
3rd. FORWARD CLUTCH still applied, BRAKE BAND released, HIGH CLUTCH applied.
4th. HIGH CLUTCH still applied, BRAKE BAND applied.

If you check this sequence against the shift solenoid sequence, you will find that they don't match up.

I still believe that the band is slipping as it tries to hold the front sun gear, if it does, then it will go to 4th, if it slips, it goes back to 3rd. The brake band may be out of adjustment, or there is a line pressure problem at that particular point, caused by the TPS.

The line pressure is varied with each gear change, more pressure applied in the lower gears, dropping as the ratio rises, this is besides the change in pressure due to throttle pressure. This is done to smooth the changes. The change into 4th would have the lowest pressure. This could be why the slip is not occuring in 2nd.

A drive with a line pressure gauge fitted, should show a drop in pressure when the problem occures and I would look at the TPS. If there is no drop in line pressure at that point, then I would go for the brake band adjustment and it has to be set up right. This is tightened to 6.5ftlbs, or as tight as you can, with one finger on a 3" spanner. This is to ensure that the band is fully contacted with the drum. Then back the screw off three full turns, this sets the clearence that is necessary for the right timing of the 2nd to 3rd change, to prevent a 2 to 3 flair.

All the best Harvey.

[gl1674]

Lee,

Few more things more or less at random:

- it is quite possible that tranny does not slip at stall test, but still slips when driving - at stall RPM the engine revs are too low to produce enough torque, when driving the revs are higher and more torque is available.

- What is the mileage on the transmission and where was it rebuild (the question really is: did they change the solenoid pack or not). 150k is the ballpark for duty solenoids life, but it is very rough - a better measurements is hours of operation..

- Did you check ECU codes? A faulty TPS should also set the code in ECU. And the code is also 31.

- TPS should give you continuously changing readings as you slowly move throttle - any positions where resistance/voltage jumps or any noise when it is being moved are a problem.