Review: Thrust Bearing/Crank Endplay Check at 125k
 

All,
This post is a detailed review and commentary on the testing of the crankshaft axial movement and endplay to diagnose any possible thrust bearing failure. This post does not present anything "new" regarding the procedure or the reason for it, as there are dozens of threads on the topic. However, based upon the obvious continued interest in the subject--be it due to engineering curiosity or conspiracy theory--I figured I'd post my results.

For those that just want the results and don't care about the details, here is my summary up front:
Objective: Measure flexplate deflection and crankshaft axial movement, for detection of possible crankshaft endplay outside specified parameters due to crankshaft thrust bearing wear or failure.
Subject: 1989 Porsche 928 S4, Automatic. 124,824 miles. Well maintained, stock drivetrain. No known driveshaft or crankshaft repair.
Measurements: Initial flexplate (loaded) deflection: 4.5mm. Axial crankshaft endplay: 0.21mm (0.0083inch). (Average of several tests with digital calipers and dial indicator).
Conclusion: Within factory specifications (<0.4mm/0.016inch).

Discussion: First, let me say that this forum and its contributors are an incredible asset to folks caring for the 928 and keeping it alive! While I have performed crankshaft measurements on other vehicles, I literally learned everything that I know about the 928 TBF history and this particular test from this forum, its contributors and also links that they provided. Thanks.

I took delivery of this 2-owner vehicle almost a month ago, and I had the opportunity to review all of the extensive maintenance records prior to my purchase. (I even was able to talk to the mechanic for the car the last six years.) There were no records of anyone checking the crankshaft endplay for possible thrust bearing failure, nor any records of any major service to the engine, crankshaft, torque tube or transmission. Maintenance was largely either preventative or routine R&R of worn or non-functioning parts. So, upon delivery I was very anxious to perform the TBF tests. I drove the car a total of about 50 miles before putting it on the lift for this test and many more over the last two weeks.

Accessing the flywheel/flexplate area was fairly uneventful. As expected, I had to unbolt the exhaust from the manifolds so it could drop down enough to be able to remove the two bellhousing bolts furthest back. The "worst" part of this was loosening the bolt on the engine that holds the short bracket to the manifold. It was on there good. For better or worse, I only had to struggle with that on the right side..the left side bracket was broken in the middle. When removing the exhaust bolts, the bracket bolt and the bellhousing bolts, I saw no evidence that any had been wrenched before. Also, I followed the advice of at least one of the forum members, and I just didn't replace the two back bolts. That bellhousing cover is so light, I'm not worried about it going anywhere.

With the bell housing removed, I observed a very clean inside area, despite the fact that the plug on inspection hole on the bottom was missing. Looking at the flexplate clamp and driveshaft, I could see a long portion of the splines compared to other pictures I've seen in this forum. There was no sign of any slippage along the splines. After using a prybar to carefully rotate the engine (here is where I feel like I will be banned if I don't mention "rotate the engine clockwise only, never counterclockwise") I lined up the pinch bolt for loosening. Again, no indication of any previous loosening of the pinch bolt--it still had the red factory paint dab lining it up with the clamp. I could see that the flexplate did have some load deflection, and I measured it to be 4.5mm. Not excited about that, but I've read about worse.

I loosened the pinch bolt, and while I did not actually visually see it or feel it, the flexpate and clamp did move once unloaded and returned to "flat." I could also now see less of the splines on the driveshaft. (I'm guessing about 4.5mm!) I used a prybar to first move the flywheel back (toward the rear of the car). I had read that you would hear and feel a "clunk" or "thump" letting you know you had moved it rearward enough, and I did. I took a measurement using a digital caliper across the edge of the flywheel and the edge of the upper part of the bellhousing. I then used the prybar to move the flywheel forward, again with a tactile "thump" when it had moved. I took another measurement and calculated the difference. I repeated this back-and-forth test three or four times, until I felt comfortable with my readings. (I attributed any very small deviations between measurements to being a human.)

Then, I set up a dial indicator on the flywheel and performed the back-and-forth test measurements another 3 or 4 times. (Why use both the caliper and the dial? Heck, I have both in my garage and I've already gone this far, so why not?) Satisfied with the measurements, I moved the flywheel back to the rear (unloaded), tightened the pinch bolt to 66 ft/lbs. (110% of original factory specs), and marked the location of the clamp along the splines with white paint. I did not use any thread lock.

The results are as stated at the beginning of the post. Unless I change my mind (based on something posted on this forum, I'm sure!) I don't think I will go with a different clamp or loctite, but just make a practice of inspecting, measuring and releasing the pressure on the flexplate every 5,000 miles or so, and just enjoy driving the car for a while instead of being underneath it.
Thanks,
Jason

Jason,

A good write up. The end float you measured is fine but I would advise caution about the displacement of the clamp you report. I introduced the over-tensioning of the pinch bolt to the list almost 20 years ago to the day after I had problems. I had a drive shaft snap and upon replacement a previous vibration at 3050 rpm disappeared. After two weeks and a few miles the vibration reappeared and a check revealed front clamp slippage of 3mm. We torqued up the joint and I visually checked the torque with two different torque wrenches and the vibe disappeared again only to reappear about 2 weeks later. This time I refused to accept the car unless and until Porsche could come up with a solution. They refIerred the matter to HQ and within minutes they came up with the over torque solution- not bad for a problem that they said "did not exist"!

About this time I came across the Rennlist as the internet had just become available to the public here. I immediately floated this dilemma and it just so happened that Earl Gilstrom suggested the Loctite approach. My 90S4 was the second 928 to have this approach applied to it and it held tight until I lost the car in a smash some 6 years later. My current GTS suffered a TBF failure for the previous owner. I put my S4 motor into the GTS chassis in 2006 using the GTS drive shaft and the GTS flexplate clamp that slipped and used the Loctite procedure on the installation. That joint is still holding some 13 years on but of course I cannot predict whether it will let go tomorrow but if I suddenly see a vibe at 3050 rpm I will know what to look for.

The examples that slip typically slip and survive seem to do so by 3mm. If yours has slipped by 4.5mm that would trouble me some. The additional bolt torque may help a little in some cases but in my case it made no difference whatsoever. I tend to put the slippage down to machining tolerances as it seems to be the only credible explanation as to why some slip and some do not. What seems pretty clear is that once an installation has a tendency to slip predicting when it will fail with TBF is more or less impossible but the one thing one can be sure of is that once the pressure imposed by the slipped flex plate causes the lubrication to breakdown thrust bearing failure will take place instantly and when that bearing spins in it housing [as they invariably do] terminal engine destruction is more or less guaranteed. Repairs of such damage have been achieved by folks with such means but the cost to have such done professionally would make it impractical compared to the cost of a second hand unit.

If you choose to take no mitigation's that of course is your choice but knowing what you have seen and to leave as is is tempting fate somewhat 20 years ago I had no options available to me and I had discussed the problem with my fellow engineering colleagues. We concluded that a Loctite type of approach should be feasible and lo and behold Earl presented me just when it was needed and it has worked for me. Now you have the Constantine clamp available and Roger also does the Ritech clamp that is easy to install. There is also Ken's PKlamp but no sure if it is still available.

OP: Thanks for the write-up. What was the variation in the end play measurements you obtained using the caliper vs. the dial indicator? I assume not much.

As a data point: my '87 with 127K miles seems a lot like yours - no evidence of past fiddling in that area of the car. I measured the flex plate deflection @ 0.140" (3.6mm) and end play @ 0.007" (.18mm).

What number for the Loctite did you use or recommend?

Thanks

Loctite 290 was originally specified by Earl. The second time I used it I had to settle for Loctite 270 that has a slightly higher breakdown temperature despite the lower catalogue number.

FredR,
Thank you very much for your comments. Your experience, while unfortunate, provides valuable insight into the issue. Just to clarify, you discovered the slippage issue 20 years ago, but never actually suffered TB failure, correct? What was the mileage during your episode(s) and had there been any previous work that could have been an aggravating factor?
As mentioned, I certainly will consider a clamp the next time I go in. (I didn't choose to go with a clamp right out of the gate before I had even done any measurements or inspections.)

Geza,
Wow, your starting point and results are very similar to mine. What are subsequent measurement results/mileage if any?
The measurements with the dial indicator had nominal variations once the tool was securely set up. I consider measurement with the calipers to be a less reliable option, because even though I marked the measuring locations, I had to remove and re-set the calipers in between each fore-and-aft movement procedure. So, it is virtually impossible to know if I was measuring in precisely the same place...which is critical when measuring such small tolerances. That being said, caliper measurements varied at most by +/- .05mm (0.002 inches). So, still within acceptable wear parameters.

I've only added ~1000 mile since resetting, so I haven't rechecked, and probably won't for a while. Regarding the rear bolts - I swapped them out for some slightly shorter ones (CRES M8 x 20mm), which still had plenty of thread engagement, but were able to be installed (and removed) once the exhaust was back in place.

My drive shaft snapped at about 87k km and everything became apparent during the recovery process. The same motor is in my current GTS chassis and has now covered about 160k km. As best I could fathom out the motor was unmolested until removed from my late 90S4 after the smash back in 2005.

As I am concerned it is most unwise to drive the automatics with the later model clamp without additional mitigation. When it goes invariably there is no warning- just a lot of tears and many $$$'s down the toilet. The general consensus is that the problem tends to manifest itself in the 60k km to 120k km mileage range but that is not a reliable performance indicator.

If I missed it did you use a caliper or a mounted dial indicator for the measurements? I read as using a caliper. For most accurate results it is recommended to use a mounted dial indicator.

See Post #1 and #6.
Thanks, Jason

Excuse my ignorance in advance, but what did you determine is the correlation between a "snapped driveshaft" and crankshaft thrust bearing wear/failure? A "snapped driveshaft" could result from numerous factors disparate from excessive crankshaft endplay.
Thanks, Jason

I always thought the idea was to clamp the bolt after moving the flywheel to the rear, since it wants to creep forward? But I'm always learning new things about this car!

You are correct. That is what I did. I edited the original post for correction.
Thanks, Jason

At some point in the manufacturing process Porsche changed the design of the automatic drive shaft for a second time. The original design was shimmed and no issues reported with it, then around 1984 they went with the flexplate design used until the end of production and with the arrival of the S4 they went with an increased diameter shaft with a taper either end wherein the splined shafts were the same as the previous models. It seems some early S4's were supplied with the straight shafts presumably until stocks ran out and then used the later version.

There is a popular conception that the later tapered shafts fail because of the taper and associated stress raiser. I have a feeling this is a false perception . If the stress raiser were the sole cause of shaft failure then logic suggests that all drive shafts would fail at some finite point when fatigue sets in but this is just not the case. My original 90 S4 tapered shaft failed at 87k km. My current GTS tapered shaft is still going strong at 160k km and we regularly hear of tapered shafts doing way more than this. My theory is that when the flexplate clamp slips it is experiencing something it was not intended to do by design. The flexplate is intended to absorb axial misalignment, not a false positioning of the clamp on the shaft. When my drive shaft failed, prior to that I had noticed a vibration at 3050 rpm but did not pay too much heed to it. When the replacement shaft was installed initially there was no such vibration but shortly after installation the vibration reappeared as I reported earlier. After spending quite a lot of time analysing the situation I formed the conclusion that maybe the larger diameter shaft was getting a bum rap on a false basis.

My reasoning was that when the clamp slips it must do so for a reason. That some work and some do not was the puzzling bit. This has never been resolved but I figured it just might be a simple as production machining tolerances. Get a combination of the shaft at the lower end and the clamp at the upper end of the acceptable range and maybe that creates a combination where eventually the clamp for some undetermined reason no longer can hold- maybe due to fretting or whatever. When the shaft was vibrating at 3050 rpm, I reckon this may have caused the onset of fatigue failure. Not nice, but a better option than TBF. The 928 is known to experience vibration when the engine mounts start failing as characterised by vibrations at 1400 rpm and 2800 rpms. Thus not inconceivable that the clamp slippage modulated the harmonic frequency a little.

Bottom line- clamp slippage might just explain incidences of tapered shaft failure. If the tapered shaft design was inherently flawed then they all would fail and they do not. A thread I raised about this subject matter produced some interesting responses. The most surprising was that some of our more experienced professionals reported that they did not see too many of these shafts fail despite the perception that they do fail. Thus my conclusion that something must be causing some of the tapered shafts to fail and it was not their inherent design. Thus the tapered shaft may be the victim rather than the culprit.

In a discussion with my long time 928 friend and TBF correspondent Constantine I asked him if he had ever heard of a tapered drive shaft failure when fitted with his bullet proof clamp and the response was "never".

Purest hearsay, but I don't think that the crank end play and the shaft snapping have much to do with each other. I have one data point: A car with factory spec end-play and a snapped shaft. Beyond that it's just a guess.

As with airplane crashes, I suspect a bunch of factors combined caused the failure. Here are a few possible candidates: Shaft redesign, torque tube internal bearing or damper failure/migration, "limp-home" mode, strategic corrosion. The prevailing theory is the shaft resonates in a "whippy" fashion and any of the above that cause additional whippiness or fatigue can contribute. Certainly demonic possession has not been eliminated either.