yet another tranny question

[gl1674]

edit: Most cars have a plugged hole for running the hose from gauge to tranny - if the SVX has one I can't find it - anybody know where it might be hiding?

Location: Left (drivers side), on a side wall approx 1/3 - 1/2 height up, just behind (~1 inch) the line between tranny and front diff casing. Access is from under the car, exhaust pipe is in the way - make sure it is not hot. You can get to the ports without removing exhaust, but getting there and looking at the same time is a bit problematic.

You are looking for 3 10mm bolts in a vertical row, all 3 are recessed. You need the middle one. The bolts should have rubber o-rings on them. The other two bolts are also test ports, but I'm not sure what lines they are in... AWD transfer pressure port is somewhere else - at the back of the tranny.

[UberRoo]

Originally posted by gl1674

That would prove precisely nothing. TCU does not wear out, it either work or it does not. But solenoids are mechanical things and they fail.
Mine came to a stage where it would stay closed as long as there were no signal applied to it. Any duty signal (even 5% ratio) and it would flip fully open. I don't know why it was happening, I cut it open and did not find anything obviously broken, but it was not working as it should.

Even if your were the engineer who designed the TCU, it would be highly presumptuous to say that it would either work, or not work. Electrical systems seem to invoke a certain amount of black magic. A recent engineering graduate with no experience will claim that every electrical problem can be explained rationally, but any experienced technician would beg to differ. Electronics are mechanical too - at a quantum level - and they exhibit bizarre traits that science cannot explain. ...let alone schmoes like us who just want the stinkin' thing to just friggin' work already. (Your computer is made from the same junk the TCU is made from. It's never had a hiccup?)

Besides, the output of the TCU is based on input signals, and the inputs come from mechanical sources. Shoot, they travel down wires. So do the output signals. Wires can be faulty. Loose connections. High resistance. Ground faults. Inductance. Pathing. Or just plain broken.

Don't get me wrong though, the solenoid is a very guilty looking suspect and the TCU probably won't even need to testify. Just don't scratch it off the list too soon.

[lee]

Originally posted by oab_au
I am not sure, but I would think that as the Manual Mode is for slippery surfaces, the TCU may tell the ECU to turn the power down, on the take off.

Sorry, I can't post the diagrams of the manual as my main computed has died, due to a servere virus.

Harvey.

P.S. Fluke DVMs are the best, bullet proof.

Thanks Harvey. As usual, my command of the written (and probably spoken) English fails. After all, I'm a physicist, the folks on here who read what advice I can offer are lucky I know how to spell Kat.

So, when I stated there was poor acceleration, I was really only talking about the time/speed frame wherein the transmssion would have been in 1st gear if Manual mode hadn't been selected. Once the car was up to normal 2nd gear speed the response was normal.

I agree on Fluke, unfortunately just a bit out of my current price range given the other demands I have on my wallet. Please keep in mind I have a wife and 4 kids. And one of those kids is a female in college.

[lee]

Originally posted by gl1674


Location: Left (drivers side), on a side wall approx 1/3 - 1/2 height up, just behind (~1 inch) the line between tranny and front diff casing. Access is from under the car, exhaust pipe is in the way - make sure it is not hot. You can get to the ports without removing exhaust, but getting there and looking at the same time is a bit problematic.

You are looking for 3 10mm bolts in a vertical row, all 3 are recessed. You need the middle one. The bolts should have rubber o-rings on them. The other two bolts are also test ports, but I'm not sure what lines they are in... AWD transfer pressure port is somewhere else - at the back of the tranny.

Thank you very much. that is much needed information.

However, as I already told Harvey, my English usage sucks. I was actually wondering about a hole from the interior to the transmission so a tech could watch the gauge will in-car. see attached for an example from the Legacy manual.

[lee]

What I hope is my last post on this today. My transmission continues to work well to/from work today (75 miles round trip).

From this I further assume the TCU had a glitch - although why it took several weeks to work itself out is a mystery.

or

The solenoid was never reaching its potential (perhaps stuck/leaking) and something that happened while at the tranny shop (the stall tests?) has corrected the problem (albeit highly likely to be a temporary fix).

So, I'm under much less pressure to hook up the gauge and test voltages (since it seems to be working). If the problem comes back then I'm going to bet on the solenoid. If it stays healthy, then my bet would be the TCU glitch - a few posts back I mentioned the battery had been disconnected for a while due to a bad terminal - I may (???) have made the reconnection just as the last electrons were fading from the system memory (on the way to the reset that occurs by leaving the battery unplugged over night).

Hate to sound like a broken record, but I want to thank all who have attempted to help me here. If any of you should decide to visit Mickey's place (Disney World) - I'm about 60 miles east - head directly to the nearest beach, at water's edge you'll be within 4 blocks of my house. Beer is on me.

[oab_au]

quote:
--------------------------------------------------------------------------------
guess you could do that, though I would be a bit reluctant to do it, in Manual mode due to the lower pressure that this mode would use.
--------------------------------------------------------------------------------


Harvey, why would that be? At full throttle TCU will give full pressure - no matter what gear you are in. At least my FSM gives the same numbers for all forward gears.



quote:
--------------------------------------------------------------------------------
I was thinking, that another way to see the line pressure performance, would be to connect a DVM up to the A solenoid to watch the level of the modulated signal, that the solenoid is receving
--------------------------------------------------------------------------------


That would prove precisely nothing. TCU does not wear out, it either work or it does not. But solenoids are mechanical things and they fail.
Mine came to a stage where it would stay closed as long as there were no signal applied to it. Any duty signal (even 5% ratio) and it would flip fully open. I don't know why it was happening, I cut it open and did not find anything obviously broken, but it was not working as it should.


quote:
--------------------------------------------------------------------------------
the TCU may tell the ECU to turn the power down, on the take off
--------------------------------------------------------------------------------


Hhmm, do you know this for sure?

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Gid'ay, The A solenoid has two feeds going to it. One from the throttle line, and one from the TCU. These are independent of each other. The throttle sensor line can be at full throttle,(no signal applied), and the TCU feed, can be at full signal, which will turn the line pressure down, regardless of the throttle position.

This is like in the old auto box where the shift valves are ballanced between, the governor pressure, and the throttle pressure, to decide the shift point. We don't have a 'governor pressure', so this is created by the TCU, based on the speed sensors and the data written it the TCUs memory.

So if you checked the signal on the solenoid side of the dropping resistor, while doing a full throttle run, you would see the moulated signal being reduced by the TCU on the change points. This is done to make the changes softer, like when it is changing from 3rd to 4th, on a light throttle (low throttle pressure) the TCU still will reduce the line pressure to smooth the change. I think this is why Lees band, is loosing grip, the line pressure is going too low.

In manual mode the TCU will reduce the change pressure even more to prevent the chance of wheel spin. It is this same thinking that makes me belive that the TCU would signal the ECU, through the Torque Controll line, to reduce the torque by retarding the spark. Don't know for sure, just an educated guess.

The solenoids do go down, I also have pulled failed units down to find nothing, but Lees trouble is only on the one change, so I find it hard to bring this back to a solenoid, but nothing is impossible

Harvey.

[gl1674]

Gid'ay, The A solenoid has two feeds going to it. One from the throttle line, and one from the TCU.

Gidday mate,

The two feeds thing is not exactly correct. Solenoid A has exactly one wire coming into it. This wire is coming from the TCU.
TCU knows the throttle position and can adjust the signal to the solenoid as it sees fit. It could do different pressures in different gears or in manual mode if it was programmed to do that, I just do not see any evidence that it does that.

The potentially confusing thing is with the dropping resistor. TCU actually has two outputs that control the Solenoid A - one goes from TCU to Solenoid A directly, the other via the dropping resistor. Both outputs are outputs - they do drive the solenoid. This is used to create a combination of a constant bias (via dropping resistor) and pulse-coded modulation. The reasons for doing such complex constant bias + pcm adjustments escape me; it seems everything could be achieved with just the proper pcm ratio. In the later models Subaru has actually dropped the constant bias part - no more dropping resistors.

The power down via spark retardation has negative consequences on fuel economy and cat temperatures - that's why I am asking. It seems that just the higher gear in manual mode (no first) should take care of the extra torque...

[oab_au]

Yes I can see that I have caused confusion by calling the feed from pin 7 of the TCU the "throttle line". I'll explain.

The TCU has two lines running to the A solenoid, they join in the loom somewhere. The line from pin 7 runs through the dropping resistor, to the solenoid. I have been calling this feed the "throttle feed" as it mimics the throttle sensor output, just at a higher voltage. It rises and falls with throttle movement, I have not seen it deviate from this. So the line pressure will rise and fall with throttle movement.

The other feed from pin 8 is the TCUs modulated signal, this is also run to the A solenoid and modifies the throttle movement signal in accordance with the program written in memory. The effect this has is, under full throttle, the pin 7 feed turns down the voltage sent to the solenoid, to give full pressure. when the change is to happen, the pin 8 signal, raises the voltage at the solenoid to drop the line pressure, while the change takes place then turns it back off so that pin 7 can again control the pressure.

Why it is done this way, I have no idea. The dropping resistor is necessary to prevent the higher voltage on pin 7 from flowing too much current through the 1.5/4.5 ohm solenoid resistance, it is also necessary to allow the two different signals to operate inderpendently, otherwise one would tend to short the other out.

Yes all this could be done in the TCU with out the dropping resistor line, but we are stuck with it.

The Torque Control line to the ECU does turn the power down by retarding the spark. It is done to make the changes smoother. It only happends for a few revolutions, then it advances back to normal. As I said, I don't know for sure, that this is done in the manual mode, but as they have gone to the trouble to include this function, I would think they would have gone all out, to make sure that the wheels don't break traction. Turning it down on take off and the 2/3 change would seem to be on the cards.

One day, mate, we will get to the bottom of how it does it and why

Harvey.

[oab_au]

Lee has the slipping happend since you adjusted the band before the stall test.

Harvey.

[lee]

Originally posted by oab_au
Lee has the slipping happend since you adjusted the band before the stall test.

Harvey.

If I understand the question, yes it did.

I adjusted the bands and checked the tps, reset the ECU, hooked back up the dropping resistor, etc. Basically trying to go from what should be the best possible baseline.

Later I went to the local tranny shop where they did the static stall test. It has been working very well since then. Only the occasional glitch - that may be just my nerves - as I've been watching/feeling for the slightest sensation of "wrongness".

As a side note: I noticed almost no difference in shift "harshness" with the resistor unplugged compared to plugged. However, I'm assuming the TCU knows it's there (or not) as I get the blinking power light and proper code when it is disconnected.

Today I checked a few voltage numbers at the TCU for the voltage going to solenoid A. I will admit the engine/transmission was cold (which goes against the stated test procedure), but I got strange numbers. My data says the figures should be 1.5 to 3 volts with throttle closed, and less than 0.5 volts throttle fully open. What I saw was about 0.4 volts closed and about 0.25 fully open. the interesting thing is watching the voltage as I opened the throttle. The voltage rose (to about 0.9) then began to fall. although I didn't record numbers I think it would look like a traditional bell curve. The data set also says I should see about 1.5 to 4.5 ohms between pin and body ground. I got zero. I hope it's a case of me violating test procedures, or not having the wires hooked correctly. I got so confused I just quit the test for the day rather than carry on in a possibly mistaken way.

I will test tomorrow when the engine and tranny are warm and I have a fresh outlook.

As I can't find a factory porthole, I'm considering drilling a hole through the transmission tunnel (that I can later plug and seal with a dab of silicon) so I can run the transmission pressure gauge inside while driving to/from work.

[oab_au]

It depends on which pin in the plug you were measuring. If it was the pin 7, green/yellow wire, then that is strange. If it was the 8 pin, green/red, that can be ok. you can't measure the A solenoid resistance with it connected, as there are other lines connected. You have to unplug the loom to the box and measure the gearbox side of the plug, green/yellow.

You may have something there, If you can say which line you are measuring and the voltage, we may be able to work it out. The fact that the shift feels the same, with or without the resistor connected, points to the throttle movement not changing the line pressure, that could be causing the slip.

Harvey.

[gl1674]

I noticed almost no difference in shift "harshness" with the resistor unplugged compared to plugged.

That's not right.

Also the voltage on solenoid should decrease continously as you are opening throttle. Max voltage should be at idle - solenoid should be fully open and provide minimal line pressure.

You can check the voltage on the middle pin of TPS - it should increase from 0.5 to approx 4.5 volts as throttle is opened, the increase should be linear.

As I can't find a factory porthole, I'm considering drilling a hole through the transmission tunnel (that I can later plug and seal with a dab of silicon) so I can run the transmission pressure gauge inside while driving to/from work

How long is the gauge hose? I would suggest using generous amount of duct tape and attaching the gauge to the outside mirror. Less intrusive than drilling holes...

[lee]

It was the line to the solenoid, not the TPS. But like I said, I was tired and frustrated, working in the dark under the dash with a drop light. Tonite is no better, so I've not even tried anything.

The TPS voltage at the TCU tracks with the voltage at the throttle body. That I have checked on a good day - as I wanted to make sure something wasn't happening to the signal between the TPS and the TCU. And it does move up in voltage smoothly - at least as best I can tell.

Harvey, makes me feel a little better to have you tell me that's maybe not a problem - just my technique in measuring (didn't know it should be unplugged to measure).

gl1674, The lack of change in "harshness" of the shift was something I thought worth mentioning. But I don't know what it means except to think the solenoid is never going to "full strength" position (hi-line pressure). Need to get that gauge hooked up.

BTW, My problems have semi-returned. When the fluid is good & hot (170-190F) the problems seem to be gone. But from cold up to that point I the problem has returned. How do I know this??? Below 150F you just have to take my word for knowing how it should feel. As soon as it gets warm enough for the TC to lockup (150F - by spec, and by my temp gauge), I can tap the throttle and see the revs climb; however, when it warms up a bit more (170-180F) and I tap the throttle, there's no movement of the rev counter (acting & feeling like it should).

The hose is 6-feet length. My plan (unless somebody tells me this is really dumb) was to find a spot under the carpet near the gas pedal, drill a 1/2" hole, and put the gauge hose through there. When done I would put in a rubber/plastic plug with a silicon caulk to seal it all up. I would feel better about this than having it hang on my mirror. I'd like to watch it for a few days, including at least one to/from work jaunt (about 75 miles round trip).

Is there a recommended wire off the TCU I should hook to a VOM & monitor along with the gauge? I'm guessing pin #8 on the 16 pin connector that joins the dropping resistor with the TCU and solenoid A (green-yellow wire)???

[oab_au]

Yes Lee, the pin 8, green/yellow wire will give the total signal, that A solenoid is getting. The lower the voltage the higher the line pressure. The other pin 7 green/red will show the throttle movement voltage. Both these wires are up at the dropping resistor.

Harvey.

[lee]

Originally posted by oab_au
Yes Lee, the pin 8, green/yellow wire will give the total signal, that A solenoid is getting. The lower the voltage the higher the line pressure. The other pin 7 green/red will show the throttle movement voltage. Both these wires are up at the dropping resistor.

Harvey.

Thanks, your last line makes it much easier/convenient to tap into. I hadn't thought about where the wire came from/went to. I was thinking of tapping at the TCU, but the dropping resistor is a much nicer work area.