4.44 Swap Finished; Complete with horrible problematic sound

[svxfiles]

Originally Posted by dannmarr

I was with my mechanic while he was using the camera. When he spotted the crack, he showed it to me. It wasn't the normal crack between the bolts, it was from the bolt to the outer ridge of the flex plate.

Wow! Did it sound anything like the video?

[Dynamic Entry]

I had heard about using a screwdriver as a stethoscope before, but had never tried it. It worked very well for me (I used a footlong philips), and I could isolate the noise fairly well. It is quiet on the motor case, and quiet on the starter. But very loud on the bellhousing, and equally loud on the bottom of the front diff.

I am under the impression it is NOT coming from any of the diff or transmission internals, since I have been told only the TQ, flexplate, and front pump should be rotating while in Park, and the noise definitely is present at that time.

Originally Posted by svxfiles

If a starter is incorrectly installed, (at a slight angle, etc.)
if the starter drive goes bad due to wear,
if the bell housing is not completely flush due to too many,
or not enough locating pins,
if the bellhousing is not flush due to a ground wire getting the way of the seating surfaces,
all can cause a problem.

- starter seemed to be installed fine, and after removing it and installing it again there was no change in noise
- not sure what starters look like when worn out, but I think the teeth look ok (and of course this noise is new and the starter condition has not changed)
- the gear seems to be in the correct retracted position, and it moves freely
- the bell housing is completely flush all the way around (I will triple check tomorrow)

Originally Posted by Hocrest

With the engine off, can you hear the noise by turning the engine manually? If so, remove the starter and try again to turn it manually.

No, turning the motor over at the crank does not make any noise, whether with the starter in or out.
None of the bolts are loose.
I cannot see any crack along the outside portion that is viewable thru the access hole
http://www.youtube.com/watch?v=X58wy...ature=youtu.be

[dannmarr]

Originally Posted by Dynamic Entry

- not sure what starters look like when worn out

Here is a photo of the inside of my starter. The teeth were fine but the inside was bad two of the brushes were not making a complete contact on the commutator.

[Dynamic Entry]

Originally Posted by dannmarr

Here is a photo of the inside of my starter. The teeth were fine but the inside was bad two of the brushes were not making a complete contact on the commutator.

Was your starter making a noise similar to what I described above?


At any rate, I think I can rule out the flexplate...





That scratch/gouge presumably goes all the way around the TQ.

I am surprised that I never felt any resistance or heard any scraping when rotating it by hand. The TQ spun freely before I bolted it to flexplate, and now it moves normally when turning the crank pulley. No noise, no binding, no scraping.

I'm taking bets on what exactly is contacting the torque converter and scraping it.

[Sean486]

wow, looks pretty deep

[dannmarr]

Originally Posted by svxfiles

Wow! Did it sound anything like the video?

My flexplate crack did not sound like that! That sounds like something else.

[Dynamic Entry]

I am completely stumped.

I have not separated the trans from the engine yet. I am really dreading that (winter is here with inches of snow and the car is outside).

Part of me is afraid that even if I do drop the tranny again, I won't even find anything.

One might ask, with noise and damage like THAT, how could you NOT find anything?

Well, I offer a review.

Video: Flexplate&TQ spinning with motor being turned at crank. This is WITH the starter installed. No noise or resistance coming from TQ, only normal compression resistance and the noise of the ratchet wrench.

So, I removed the TQ to Flexplate bolts.

Video: The TQ spins freely (as it also did during the install before I ever mated to the flexplate) The scrape is clearly visible all the way around, although it does seem to waver or squiggle (for lack of better word).
***Disclaimer, the noise you hear in this video is not being made by any scraping of the TQ, it is spinning freely.

An important note about seating the TQ fully. I know that I got the TQ seated all the way before I began installation.

The easiest way to tell for me is that the teeth on the TQ only clear the bell housing by like 1/8 of an inch.

It is common for the TQ to move while raising and shifting the transmission to hang it on the engine. This happened to me.

But, I did get the TQ back to fully seated, and I cite this as my evidence:

When the transmission and engine are mated (bellhousing to engine block, no gaps) and the torque converter still engaged into the front pump, there will be a small gap between the torque converter and the flex plate (you should be able to freely spin the torque converter with the engine/transmission bolted up tight)

I had a gap between my TQ and the flexplate, and the TQ freely spun.

Bolting the flex plate to the torque converter actually draws the torque converter out of the transmission (about 1/4") and completes the mechanical attatchment.

This was the case for me.

So. I think I had the TQ fully seated.


Here is a picture of the Outback torque converter sitting in the scrapped SVX bellhousing. You can see that there is ample room all around the TQ casing, and the narrowest margins seem to be at corners of the bottom opening.



But that said, there is nothing in there that is at all close to touching.



Also, remember that the scrape is present on the casing inbetween the teeth and the flexplate bolt flanges. Something has to be contacting the casing and NOT contacting these.


So, again, it seems like the starter is the only thing capable of protruding into the bellhousing and making contact. (Unless I dropped/lost some random shrapnel inside and it is caught in a bizarre way that allows it to contact the TQ with engine running only).

But I have removed the starter twice, and can find zero evidence of contact or damage.

I wish there was some way to run the car with the starter uninstalled.

Is push starting an automatic transmission a myth?


I am at a loss.

[Hocrest]

Use a crayon or paint pen to fill in the gouge. Put everything back together, run the engine, then pull the starter to see if your markers are scratched away. If so, there is something in there that shouldn't be.

You may not have to drop the trans? If you slide it back on the bottom studs, you may have room to check what's going on. Have a clean floor or a magnet under the trans when you take it apart, to help ID if any thing falls out.

[Tapani]

Have you checked the steering rack for rub marks? It's pretty close for any foreign stuff to be left there and at the correct location ...

See photo

Tapani

[Dynamic Entry]

I was able to slide the transmission back almost an inch without dropping the exhaust or touching the axles.

I went as far as buying an LED inspection camera to try and see all around the torque converter.

I can't see anything foreign in the bellhousing, and I could not see any marks that indicate contact.

I am going to paint/mark the TQ and put everything together again.

I am also going to paint/mark the starter surfaces. I really couldn't see any sign of damage on the starter, but I can't think of anything else that could be causing this.

I did check the steering rack, because it does line up with the location of the problem. But there is clearance between all components, and nothing foreign.


At this point I am sure I have made some ridiculous mistake, but I really cannot understand it.

[Tapani]

How deep is the gouge? In the photo you posted it looks like the TC shell is bent... I wonder how big the internal radial clearances are? I have never actually seen one machined open in a lathe - only photos...

Make sure the locating dowels have not been pushed out when installing the tranny.... happened to me one. This may cause misalignment.

You said the sound changes from reverse to D.... I trust yo checked the tranny rear mount - the loading of does change with forward / reverse change even when standing still. Nothing but the pump should rotate within the tranny itself if the vehicle is not moving in D or R.

Interesting,

T

[Dynamic Entry]

Originally Posted by Tapani

How deep is the gouge? In the photo you posted it looks like the TC shell is bent... I wonder how big the internal radial clearances are? I have never actually seen one machined open in a lathe - only photos...

You mean it looks bent here? I think that is an optical illusion caused by the fact that the scrape does not follow a straight path at that point. I have a video below that shows the TC spinning and it is easier to see the condition of the casing. I will verify, but so far looking at it in person I have not noticed any dents in the casing. The only variance on the shell topography (for lack of better word) is caused by the flanges that are welded on for the flex plate bolts.


I have not tried to measure the depth of the gouge, but it appears to be about 1 millimeter (1/32 of an inch maybe).

Originally Posted by Tapani

Make sure the locating dowels have not been pushed out when installing the tranny.... happened to me one. This may cause misalignment.

When I slid the trans back today to look around I visually confirmed that both sides have the locating dowels in place. Both are still partially inserted into their respective locating holes (transmission was not backed fully off the dowel pins).

Originally Posted by Tapani

You said the sound changes from reverse to D.... I trust yo checked the tranny rear mount - the loading of does change with forward / reverse change even when standing still. Nothing but the pump should rotate within the tranny itself if the vehicle is not moving in D or R.

Please expand on what I should be checking on the rear mount.

Are you suggesting that the rear mount could be responsible for WHY my TC is being damaged? (Which would still leave the question of WHAT is contacting my TC)

Interesting that you mention it, because I used a Legacy rear trans mount.

The SVX rear trans mount was broken in half when I uninstalled it. The rubber had sheared completely apart, so the trans mount came out in two pieces, top half and bottom half.

The Legacy mount was the exact size and shape of the SVX mount. The only difference was that the SVX mount had only 2 studs and the Legacy mount had 3. So I cut the extra one off, and used it.
I wasn't sure if the rubber would be a different density, but figured if the SVX had been running with a sheared mount, this could only be an improvement.


If the rear mount could somehow be responsible (I don't really see how the mount could affect the TC position), that would be good to know, but it still wouldn't solve my mystery of what exactly is contacting the TC.

Originally Posted by Tapani

Interesting

Indeed.

[svxfiles]

The rear mount has nothing to do with your noise.
Can you measure how far back the teeth on the TQ are from the opening at the back of the engine?
I think that you do not have the TQ installed far enough.
As you install the TQ it takes THREE steps, that is to say the TQ drops into place three times as you load it into the transmission bell housing.
If it not drop all the way back, then it will be under pressure, enough for the TQ to bind/leal/scrape.

[SVXRide]

yeah, what OT said!

Bill

[Dynamic Entry]

Originally Posted by svxfiles

The rear mount has nothing to do with your noise.
Can you measure how far back the teeth on the TQ are from the opening at the back of the engine?
I think that you do not have the TQ installed far enough.
As you install the TQ it takes THREE steps, that is to say the TQ drops into place three times as you load it into the transmission bell housing.

I wish that I had taken pictures of my TQ install. I compared pictures of what is wrong and what is right, and I was sure that I had it right.

But, I am willing to entertain the idea that I have it wrong.

I will try to get some meaningful pictures and measurements.

But please indulge me:

Originally Posted by svxfiles

If it not drop all the way back, then it will be under pressure, enough for the TQ to bind/leal/scrape.

It will bind/leal/scrape on WHAT? The starter?



Cheers