Transmission, Automatic

Flexplate Misadjustment Results in Serious Damage

I hope that this experience will allow you folks to appreciate the "minor" maladies that afflict your babies.

background: New owner exactly 30 days yesterday! very excited. Day one- drive 145 miles home, very carefully I might add. Day two- drive to store. Come out of store and try to start. No go... battery seems dead. Call AAA. Get a reluctant jump. Car dies upon coasting. Get a tow home. Put car up for remainder of nasty weather buy new battery etc. Day 30- nice mechanic and get car in for diagnosis (drive part way, get towed the rest.. very weird) I'm thinking fuel problems ie. brain, idle stabilizer valve or maybe MAF. Its a little unsettling but hey, part of ownership. I think. Now I'm really a member of the club. But I didn't expect this: Metal found in oil pan...aluminum shavings. Its very difficult to turn car over. What the hell is it. This guy is saying rebuild and I wet my pants. Any ideas or suggestions or maybe a prayer or two from the sidelines? (before anyone says buyer beware.. I'm extremely confident I bought the right car..  Full records and not problem one) Its an 87 S4 66,500 miles and a suicidal owner...any ideas? Thanks everyone,
Brendan (praying to the 928 gods as you read)

Bad news, very bad news . The center main bearing thrust surface is probably gone . The only thing that locates the crank in the block is the thrust surface if the car is an automatic improper preloading of the flexplate at the flywheel will force the crank forward, take out the bearing, the counter weights then start to re-machine the webbing in the block. If this has happened the block is ruined--not rebuild able. The engine will run until it gets hot then it seizes after a cool down it will crank and run . I have heard of several new owners who inherited this problem . People who may be perfectly honest about most things will trade in a bad car to the auto dealers, unfortunately the car still ends up in an individuals hands Porsche was aware of a deficiency in the thrust area and increased the bearing shoulder and contact surface in 1985--It is a less than perfect fix.

Jim Bailey
928 International
79 928 5spd
80 928 5spd
68 911 5spd

Here's the test for the crankshaft thrust "end play". You need a dial indicator mount it to the block, put the pointer on the end of the crank then push the crank in and out . Measure the maximum movement. New bearings should read .110 to .312 mm - the maximum allowed is .40 as a wear limit. If wear is detected at this early stage the block should be rebuildable. As to the cause of this problem, it occurs on automatics far more often. I believe it may be related to excessive preload of the front flex plate at the flywheel. Porsche at one time had a procedure to check preload when replacing the central tube later they changed the recommendation stating it was no longer necessary to check this. This problem is not a common one. I should point out that I talk to 928 owners 9 to 10 hrs per day 5 days a week and have done so for 7 years. However, it is like the Medical Doctor--every one he sees is sick. When you see my posts just remember most people only call me when they have a problem. It does make it interesting for me when I drive my wife's 79 , or my 80. I hear every noise, feel every vibration and think of all the possible things that could be happening. In all fairness, both have been great cars, very few problems each has over 170,000 miles--

Jim Bailey
928 International

The standard way of handling end thrust on the crankshaft is to have an outward facing flange on each side of one of the main bearings, with the outer surface of each flange being a bearing surface. The mating surface of the crankshaft is finished as a bearing surface. Look at the photo on page 13-4 of the shop manual - The center bearing is the thrust bearing, and you can see the finished bearing area on the crank. The max 0.4 mm or 0.015" that Jim mentioned is the sum of the clearances on each side between the crank and the bearing - In other words, the crank should be free to slide back and forth a maximum of 0.4 mm or 0.015 ". This is shown on page 13-8.

The thrust bearings on a 5-speed car have to absorb the thrust caused by the clutch throwout bearing pushing on the pressure plate to release the clutch. Otherwise, there is normally no load on the thrust bearing.

The thrust bearings in an automatic car should have a very slight preload. The procedure for checking and adjusting this is given on pages 39-141 thru 39-143 in the shop manual.


I recently replaced my torque tube and noticed that the manual mentioned prying the flywheel to the rear to make sure that there was no load on the thrust bearing before tightening the clamp on the front of the driveshaft. I verified that the splined end of the driveshaft was moving freely in the hub of the drive plate by prying the plate slightly rearward with a screwdriver. I then pried the flywheel back, (you can hear the crank come up against the forward thrust bearing face), before tightening the clamp. It's been OK so

Don Burrows
87 S 4, A/T

A mechanic up at the local Porsche dealer was telling me he had seen this MANY times and the failure is as you idle the motor just quits. It is seizing, that is why it stops.

It usually happens because some one takes out the trans and then puts it back in and FORGETS to check the clearance at the FRONT of the torque tube, they check the back but not the front and it pushes on the crank and thus pushes the crank THROUGH the block! Pretty bad deal all the way around, but some people just don't think when that put a tranny back in as to what other things might be getting pushed on....

I would love to hear if others on this list have had this problem, or if others have had a trans done and the mechanic did or did NOT check the clearance at the front of the Torque Tube.

James Morrison
Austin, TX
'80 Euro S GT-2 Race Car
'80 Euro S donor car
'81 Hellblaumetallic/Braun-Beige
'88 S4 Dunkelblau/Linnen PCA Parade Concours Class Winner
'92 GTS Polar Silver/Navy
'94 GTS Grandprixweiss/Classic Gray HCR Porschedillo Concours Winner

OK, dumb question of the day, what exactly is the thrust bearing?

Charles Willis, 88S4

What we are talking about is the crankshaft main bearing in the center of engine block. Only one main bearing has shoulders on each side (the thrust surface). It is the only thing that limits the front to back movement of the crankshaft . The width of this shoulder was increased as of eng # 81 E 05344 model year 1984 this would indicate that Porsche felt there was a need to modify this bearing . Please note this issue is not a very common problem , being in the parts business we hear about all the problems ; few customers call to tell us the car is running great (maybe more should) that they have had no problems etc.

Jim Bailey
928 International

All of these reports about driveshaft failures due to stress begs the obvious question; is there any way that the rebuilder can tell if the shaft in an exchange tube has been weakened by previous stress?

PS. Still looking for the procedure to set the drive train so that it neither pushes nor pulls the crank against the thrust bearing surface?

Don Burrows
87 S4, A/T (Parked)

The front flexplate a nearly triangular flat plate bolts to the flywheel with six bolts two at each point of the triangle (23-28 ftlbs torque ) Only then after the central tube, forward housing and engine have all been bolted together do you install the clamping screw on the center hub of the flexplate and tighten to 54-62 ft lbs. The center hub can slide freely front to back until the clamping screw is tightened. This also means it can be clamped in the wrong place. For example, when changing an engine if the flexplate were left clamped to the central tube and the replacement engine dropped into place then the 6 bolts tightened to the flywheel it may be possible to preload the flexplate - as the bolts are tightened this could be either pushing or pulling. It may be possible when changing transmissions due to manufacturing tolerances to create pressure if the front clamping bolt is not released then retightened. The input shaft and clamp at the transmission can only be installed in one location because the clamping screw must pass through a machine groove on the drive shaft (all 5 spd shafts have grooves at both ends that make it very difficult to install improperly unless you take a drill and notch the shaft to make clearance for the bolt). Yes, I have seen this and also have seen the drive shaft welded to the trans input shaft - Yes, welded. When changing central tube you must drop the rear suspension, remove the trans, then remove the central tube. The reassembly is the opposite which makes it very easy to tighten the clamping screw at the flex plate before the transmission is reinstalled then when the rear clamp screw is installed it will only go in when the groove is lined-up. Then the transmission bolts are tightened into place. Once again, preload is possible.

Porsche bulletin 928 group 3 3903 -9203 may 5 1992 - "any damage to the engine which has been caused by improper central tube installation is NOT a warranty matter" Remember this is the factory bulletin to the Dealers saying "If you screw it up We ain't gonna pay for it". At the risk of being "flamed" I submit that 90% of the problems we all encounter with our 928s are the result of improper service and neglect or misinformation. I have seen over 300 928s dismantled and the creative engineering is amazing but so many people try to work on the car without the service manuals, many shops do not own the manuals or the belt tension tools. If you have someone working on your 928, ask to see the tensioning tool. Just say you wonder what a $400 tool looks like or ask to look at the workshop manuals. To those who say "I can't afford to buy the manuals" you really can not afford not to. The vast majority of the questions asked on this forum can quickly be answered by the manuals, and accurately. They are "flawed" in the sense that they assume a certain level of familiarity with basic mechanical procedures but they are the only game in town. At least they are sold in the US, our laws require that if a car is sold here the manuals must also be available, that is not true in most other countries.

Jim Bailey
928 International

Thanks for the response. When I reassembled my drive train I made sure that the drive shaft was free in the front flex plate splined hub by prying the plate away from the flywheel a little bit and observing that the hub slid on the splines. The part that I am unsure of in the reassembly is that, assuming that the torque converter would expand when hot, I pried the flywheel back until the crank was against the front thrust bearing surface. I then stopped prying and tightened the clamp. What I'm trying to figure out is exactly where the crank should be within the front/rear clearance (in the range of 0.110 to 0.712 mm) when tightening the clamp.

Don Burrows
87 S4, A/T, (Parked)

Factory Shop Manuals Volume III , Section 39, pages 137-143 has the procedure for Adjusting the Drive plate and states the following at the top of the page:

"To avoid axial pressure on crankshaft and crankshaft thrust bearings, the connections between flywheel and drive plate (distance X) MUST be checked and adjusted after replacement of engine, flywheel or central tube. This adjustment is not necessary after replacement of transmission or transmission parts"


Please note the parts differences for 85 and later referenced on pages 137-138 before proceeding.

As Jim and Mark stated earlier this is usually caused by incorrect repairs and is not necessarily a defect or a "weak spot" in your 928. This is just another reason to be sure of your mechanics abilities and/or your own familiarity with the proper procedures for repairing your 928 if you do the work yourself.

David Roberts

As some of you may remember my drive shaft broke almost two years ago. I have also complained of a resonant vibration at around 3000 rpm which I have previously listed to seek opinions.

We posted a suggestion that it may be prudent to regularly check the tension of the flex plate to ensure that the thrust bearing is not overloaded. I took my S4 into the local dealers this afternoon and we carried out this check. When the pinch bolt was released the flex plate popped back by about 4 to 5mm! I shuddered when I saw this.

We also checked the crank end float which was approximately 0.5mm or on the outer limit of the tolerance range. Two years ago we checked this and it was half the amount. We are not entirely sure about the gauge that was used so we are going to recheck this in a day or two with another dial gauge.

After the adjustment was made a trial run indicated that the vibration at 3000 rpm had been significantly attenuated albeit there was still a little roughness. I will comment on this when I have done a bit more mileage.

What does all this mean?

First of all I am confident my drive shaft was replaced correctly two years ago. So, one of two possibilities has occurred, either the drive shaft has stretched or the splined central boss has moved forward. Whether or not there is damage to my crank thrust bearing remains to be seen but without doubt it has been under more load than it might otherwise have been and, subject to further checks, may well be at the back end of its theoretical service life.

Given the apparently unreasonable incidence of torque tube failures on S4 automatics and/or destruction of the main thrust bearing I am rapidly coming to the personal conclusion that for some reason our cars may be prone to some kind of design related failure. In Oman we have seen 4 torque tube failures in a vehicle population of around 30 machines in the last two years. Interestingly all of them were S4's or GTS's, none were earlier models [of which we have plenty here] and they typically had around 50 to 60k miles on the clock. I note that the recent Rennlist failures also had this type of mileage on them.

This begats the question why does it affect the automatics? Not sure on this one but I have a theory that it may relate to the way in which the gearbox is set up. The gearbox has a vacuum modulator. Set to one extreme the change is soft but sloppy. On the other extreme it is quick but a little thumpy. My gearbox has always been quick on shifting. Remember, even though the auto box is a Mercedes Benz design, the valve body was modified by Porsche to give a sportier performance. I am starting to wonder if sharks with their box modulated [intentionally or unintentionally] on the sporty side become more vulnerable to shaft failure because of the torque impulses exceeding the fatigue life criteria the machine was designed for. If this is the cause then it may be unfair to describe the failure as being "design related".

The torque tube, besides transmitting the power, also acts as a transmission shock absorber. The shaft continually winds up and under constant load twists proportionately to the torque applied [a simple application of Young's law]. Excessive shock loading [during gear changes for instance] may well have an impact on the fatigue life of the material.

Hence, if the shaft is exposed to more torque than it can absorb with the elastic limit of the material or if the torque cycles applied to it mean that the fatigue life is significantly reduced, one can reasonably predict that the torque tube will stretch and ultimately fail. As it stretches, it thrusts in a forward direction and la voila- the thrust bearing saga.

If you are lucky, the torque tube goes pop and if not the whole engine goes due to thrust bearing failure.

It only takes half an hour for the mechanic to carry out the check for the tension on the flex plate and about an hour if you want to check the end float on the crank while you are at it [recommended that you record this].

If you own an S4 or GTS automatic and you have not had this checked I strongly recommend that you consider doing this immediately and ask for the check to be built into your services. If there is a design weakness we cannot prevent the torque tube failure but we can protect the thrust bearing.

There are other reasons as to why the clamp may have moved on the splines. Incorrect torque being the obvious one, maybe vibration, but this would not explain the torque tube failures.


Fred R
1990 S4 auto- questionable end float!

I would like to thank you for your post on what seems to be a growing concern that most of us were blissfully ignorant of just a short time ago. I cracked open the manuals on replacing the torque tube and saw the procedure for screwing down the clamp *after* bolting the six "drive plate attaching bolts." Undoing the clamp and retorquing seems to be a straight forward procedure that can be done by most at home. Am I right about this or will simply unscrewing the clamp, letting it adjust, then retorquing it not suffice as periodic preventive maintenance? This would seem better than just ripping open the oil filter to look for shavings. As far as the torque tube, how about softer settings for the gear changes? Just trying to find some proactive solutions to these concerns. Any of the enlightened want to share their thoughts?

1989 928S4, Samtrot Metallic, AT (and nervous)

I've only driven a hundred miles or so after replacing my torque tube and am about to try the check that you did. You said that you saw that the flex plate popped "back" by about 4 to 5mm when the clamp bolt was loosened. When you say back do you mean that the hub of the flex plate moved toward the rear of the car when released, or that it moved toward the flywheel?
I have heard that torque converters, in general, tend to expand when they get up to operating temperature. This may explain why most of the reported failures are in automatics. This would make the failures occur at the rear thrust bearing surface due to the thermal expansion of the converter added to that of the driveshaft being greater than the increased expansion of the tube resulting from it's alloy construction, resulting in a net expansion of the shaft relative to the tube.
When I assembled my drivetrain I first confirmed that the shaft was riding free in the flex plate hub, then pried the flywheel to the rear of its play before tightening the clamp bolt. If the above amateur analysis is correct then I should be OK. I'll let you know what I find.

PS. Wouldn't it seem more logical that the torque converter, being a fluid coupling would absorb torque spikes to some extent and thus be easier on the driveshaft than the "hard" connection produced by the manual setup?

Don Burrows
87 S4, A/T

Here is the TSB from Porsche and outlines their theory... or should I say facts.

While it specifically about 89 and later autos, I feel it also applies to earlier twin coil engines that have an ignition coil (or wire) go bad.


Broken Central Tube Shaft Vehicles With A/T

July 7, 1992
Model 928 S4
Group 3
Part Identifier 3903
Number 9206

Subject: Broken Central Tube Shaft Vehicles with Automatic Transmission

ATTENTION: Service Manager/Service Technician

Models Affected: 928 S4 Model Year 1989 to 1991

Concern: Activation of the ignition monitoring system (injection circuit switched off fault codes 1131 or 1231) can lead to breakage of the central tube drive shaft.

General Information: Oscillations in the central tube system that occur when the ignition monitoring system has switched (flywheel effect) can lead to breakage of the central tube drive shaft. This can occur only in vehicles with automatic transmission where the vehicle is operated at an engine speed of approximately 1000 RPM. Possible causes are:

- Damaged or defective exhaust gas temperature sensors.

- Possible damage to the ignition coil wire left side (in driving direction) between the ignition coil and distributor cap.

- Poor grounding of the mounting plate for the ignition final stages.

- Poor physical connection of the electrical plugs on the ignition final stages.

Parts Information: A new version temperature sensor, central tube and coil wire are installed in production.
Temperature senso, Part Number 928 606 155 02
Central tube with changed material for drive shaft, Part Number 960 421 012 07
New coil wire with hose covering (for left side), Part Number 928 602 040 01

Repair Information:
1. If the central tube drive shaft is broken, replace the central tube with the new version part (see parts information). Refer to Technical Bulletin Group 3, Number 9203, dated May 5, 1992 for hints on central tube installation.

2. Replace the left side ignition coil wire with the part number listed in this bulletin. Be certain the coil wire is routed freely and not under tension.

3. Check all ignition components and connectors for corrosion, tightness, correct connection and damage. Repair or replace as necessary. If an ignition circuit has malfunctioned, an LED indication will be given by the ignition monitor relay located on the L-H control unit mounting plate.

Ignition circuit I (cyl. 1-7-6-4) Red diode
Ignition circuit II (cyl. 3-2-5-8) Green diode

4. Replace both temperature sensors located in the exhaust ports:
Model '89-'90, cyl. 4 and 8
Model '91 cyl. 3 and 7
Use the new version temperature sensors listed in this bulletin.

The function of the ignition circuits and light diodes located in the ignition monitor relay remains unchanged. It is not possible to determine from the LED display of the ignition monitor relay if one or both temperature sensors are defective or which temperature sensor has failed.

When installing temperature sensors, coat the sensor adapter threads with molykote paste HTP (white) and torque sensors to 10 N-m.

5. After the temperature sensors are installed, the voltage difference of the sensors must be checked in order to ensure proper operation of the ignition monitor system.

Checking sensor voltage difference:

- Start engine and bring to operating temperature.

- Loosen the mounting bolt for the ignition monitor relay and pivot the relay up to gain access to the plug terminals Do not disconnect the relay plug.

- Set volt meter to the millivolt range and connect leads between E1 and E2 of ignition monitor relay (Figure 3, white wires). A digital volt meter must be used. Polarity is not important.

- Measure voltage with the engine idling and again at approximately 2000 RPM. A maximum difference of +2.5 mV or -2.5 mV (depending on polarity) is permitted. If the voltage difference is above 2.5 mV, stop the engine, loosen and rotate one temperature sensor. Retighten sensor and check voltage difference. If the difference is above 2.5 mV, stop the engine and rotate the other sensor. If after rotating the sensors to different positions, the voltage difference is too high (above 2.5 mV) the temperature sensors are defective and must be replaced.

David Roberts

When I finished installing the torque tube about three hundred miles ago, I pried the flywheel back from the engine, which was stone cold, until I heard the crank make contact with the thrust bearing, stopped prying and tightened the clamp bolt. Yesterday evening we put it on a lift while still at operating temp., fitted a dial indicating micrometer on the flex plate hub and released the clamp bolt whereupon the hub moved about 0.025" to the rear. We the checked the end play and found it to be 0.127 mm, well within the new bearing spec of 0.110-0.312 mm. We then pried the flywheel away from the engine to set the crank several thousandths off of the front TB and tightened the clamp bolt. We let it set all night and then found that the hub, still clamped, had moved to the rear about 0.005" and that's where we left it.

Don Burrows
87 S4, A/T

I have been assisting 3 friends with automatic transmissions in trying to determine if crankshaft thrust bearing wear is excessive on their cars. For simplicity, I have named the cars A,B and C. Here are the findings. "Hub movement" means: with a dial indicator on the center of the thrust plate hub, loosen the hub clamping screw and record movement of the hub. This shows the preload on the crankshaft. Presumably, no preload is desirable.

Car Year Mileage Hub movement CS End play

"A" '89 85K .120" .040" .009"

"B" '89 114K .100" .002"

"C" '88 44K .105" .016 .007"

We also checked the difference in preload caused by a cold car and a hot car using car "C". After making the initial check and finding .105" hub movement, we tightened the hub clamping bolt with no preload. The car was then driven ~2000 miles including one DE. We then checked the car cold and found .016" preload. Hub migration? We reclamped the hub and drove the car ~30 miles and rechecked the preload hot and found .003" preload. It appears that the shaft/torque tube do not change enough in length from cold to hot to affect the preload. Or was the .003" change caused by migration? Car A was also rechecked after 2,000 miles and the hub moved .040".

It also appears that the hub does migrate forward enough to put significant pressure on the thrust bearing, although the wear on cars A and C seems acceptable.

To try to stop migration on car "C", we used Loctite 290 penetrating formula on the hub splines and 242 on the threads of the clamping bolt, although the bolt does not appear to loosen. We will check periodically to see if this stops migration.

On cars "A and "B" we used 271 on the threads. All bolts were replaced and torqued to 110% as per advise from Porsche.

One thing that we don't understand is: why does car "B" only have .002" crankshaft end play? On car "A"and "C" the crankshaft "clunks" back and forth when we pry them. Car "B" does not appear to move, but the dial indicator shows .002" movement. The specification is .00236" (.06mm) minimum. It is hard to believe that a car with over 114,000 miles has less than the minimum axial play.

Is this a symptom of the dreaded thrust bearing failure? Is there another way to check for thrust bearing failure without tearing the engine down? (Update: 12-05-00 an oil analysis on "B" showed no problems.)

Earl Gillstrom '88 5 speed (.007 CS end play) 99,000 miles

I have some possible good news for automatic owners that are worried about crankshaft thrust bearing failure. One of the cars that I have been monitoring for drive shaft movement in the front hub of the thrust plate has stopped moving. We did an experiment on Gary Knox's '88 automatic (car "C" in my previous email) that seems to have stopped the movement of the shaft. Here is what we did:

Removed the pinch (clamp) bolt in the front driveshaft hub. This relieved the thrust plate tension and allowed the shaft to attain the proper position.

Poured Loctite 290 (penetrating formula) into the three slots in the hub. The intent was to Loctite the shaft to the hub without disassembly. Disassembly requires removal of the torque tube. You should "prime" parts before assembly, but it is pretty time consuming to R&R the TT.

Installed a new bolt P/N 931.421.240.00. using Loctite 242 before assembly, torqued to 65 Porsche had recommended to Fred Rourke's mechanics, replacing the bolt and tightening to 110% of recommended torque. The standard torque is 59 The Loctite was our idea. I doubt that the Loctite on the bolt was necessary, but it can't hurt.

Let the car sit for 24 hours. Loctite says that a full cure takes 24 hours.

The car now has about 2000 miles on it since we did this, including two days of drivers education with NO hub movement. You can say that this is not a long term test, which is true. But we checked this car previously and found that the hub had moved .016 in about 2,000 miles and one DE. The first check we did, there was .105" movement (40,000 miles). So it is very encouraging. This is not an expensive procedure. About $6.00 for the bolt and $6.00 for the Loctite. It takes about an hour or two depending on experience. Check the crankshaft end play while you are in there. We painted the shaft splines white where they enter the clamping hub. After reassembly we could look through the hole in the bottom of the bell housing cover and see the painted splines. We think that we can do a quick spot check for shaft movement without any disassembly at every oil change by looking in the hole. Maybe a small diameter fiber optic light would make the job easier but the white paint helps a lot.

Car "D" was done on 11/11 using Loctite on the splines. Unfortunately we used red paint, so it may be harder to check hub movement without disassembly.

Car Year Mileage Hub movement CS End play

"D" '89 58K .134" .005" (added 12-05-00)

Here is the procedure:

Drop the exhaust at the manifolds.
Remove the bell housing bottom cover.
Install a dial indicator on the rear side of the forward Drive Shaft Hub.
Loosen The hub clamping screw and watch for hub movement.
On the three automatics that I have done this on, the hub moved back between .100"and .120". This preload puts pressure on the rear face of the crankshaft thrust bearing. If the hub does not move more than a few thousandths of an inch when you loosen the pinch bolt then all is OK. (Update 12-05-00, Car "D" had .134" movement.)

After the pinch bolt is loose then put the dial indicator on the flywheel, (not the flexplate). Pry the flywheel back and forth to check the crankshaft end play. The spec for later engines is .0024" to .0076". Wear limit is .016".

If your end play is more than .016", then you are experiencing "Thrust Bearing Failure".

If you are installing a TT, clean the shaft and hub splines with Loctite primer before assembly. Tighten the forward hub pinch bolt last.

Pour Loctite 290 (penetrating formula) into the three slots in the hub. It doesn't matter if hot or cold or the position of the crankshaft.

Install a new bolt P/N 931.421.240.00. using Loctite 242 before assembly, torque to 65 The standard torque is 59 I doubt that the Loctite on the bolt is necessary, but it can't hurt.

Paint the shaft splines white where it enters the hub for future verification.

Reassemble cover and exhaust.

Let the car sit for 24 hours. Loctite says that a full cure takes 24 hours.

Look through the hole in the bottom of the bell housing every oil change to check for hub migration.

The pinch bolt is 10 mm with an 8 mm "Allen" head. Hardness is 12.9.
Part number is 931,421,240,00.

I will try to assemble a list of results if you try this experiment. Let me know how you make out.

Earl Gillstrom '88 5 speed (.007 CS end play) 99,000 miles

Thanks for the note. Initial checks on my S4 showed that the pinch bolt technique was working however I have booked a slot for tomorrow morning and we are going to re-check the situation. Painting the shaft is essential to easy monitoring and white is probably the best colour [we used gold because it was there]. Interesting the vibration period that I experienced at 3050 rpm has attenuated significantly since we the modification I advised you about. Before that, re-setting the bolt seemed to smooth things up but then the vibration would return as did the tension from the hub.

I am more convinced than ever that the cause could simply be that the pinch bolt creeps with age and at some critical point it starts to let slip. There is no service inspection/replacement of this bolt or its tension and eventually disaster strikes for want of a 10 minute$6 job. Porsche need a good arse kicking for not owning up to this, if that is what the cause is.

I still do not know what causes the drive shaft to snap although that is more forgivable [but not by much]. Maybe it is a flaw but not connected to the thrust problem.

I'll let you know what I find tomorrow. Keep up the good work.

Fred R
1990 S4 auto

OK, I'm really get confused by all this thrust bearing failure talk. Having just taken my TT out and seeing all these components it is still fresh in my mind.

My question/confusion IS....if the pinch bolt, clamping the collar, were to slip/loosen, the grip it has on the splined shaft would also. CORRECT??

If the grip on the splined shaft is loosened at the collar then the shaft is theoretically free to slide in and out of the collar that clamps it. .... CORRECT??

If the shaft slides in and out of the collar(the collar is attatched to the the the crank...making contact with the trust bearing) how can the shaft POSSIBLY EXERT ANY LOAD ON THE THRUST BEARING IN THIS CONDITION??? Thats where im confused??

In my eyes the perfect solution to this would be to have the shaft "free floating: in the collar!! The problem that creeps up with this is the torque involved. It would be a fine line where the collar is tight enough to keep the splines from "jumping teeth" in the collar and stripping them almost instantly and where it is to tight to allow the shaft to slide back and forth at all.

My take on this...
Perhaps the shaft for some odd reason slides in one direction but for what ever reason wont RELEASE in the other. This one way progressive movement is how this pressure is built up. Im i more correct in that scenario???

I'm sure im wrong..but my solution would be along the lines of slapping a load of grease on the splined shaft where it makes contact with the collar...THEN TORQUE'N down the pinch collar bolt...adding Loctite to it. The grease on the shaft would perhaps provide a medium on which the shaft could slide IN...AND...OUT of the collar, YET the pinch bolt is tight enough to keep the collar around the splines on the shaft and prevent it from being stripped


I'm darn sure taking the litle amount of time it takes to check the preload now and then. my collar jumped back 3-4mm when i undid the collar for the very first time!!! I DONT EVER WANT TO SEE THAT AGAIN!!!

Also..shed some light on this vibration at 3050rpm???? I have a vibration in this area as well....but I can almost certainly attribute it to my shot engine mounts!


Dear Tony,

Perhaps I can enlighten you on what we think is going on. When a shaft is under torsion it naturally wants to shorten in length. The torque tube is given this name because its principal function is to transmit torque from the engine to the gearbox and in the case of the automatic, via the torque converter. I would presume that in the case of the automatic transmission, the drive train requires a solid link simply because of the amount of torque being transmitted. Porsche probably felt this would be more reliable than a floating collar which would tend to wear.

Consequently, the flex plate acts as its name suggests, as a flexible power coupling. Nothing unusual with that. Flexible couplings are normally used to take up any slight misalignment and hence stop any transverse loading from knocking out the main bearings of the crank. In the case of the 928 drive train the thing that is fundamentally different from most conventional cars is that the drive shaft is spinning at engine speed al the time. So far so good I trust.

When Porsche designed the 928 they wanted perfect weight balance hence they stuck the gearbox in the rear. No big deal eh? Well not exactly. Up until the 928 was introduced the only marque I know of that had that configuration was the Alfa Sud although maybe there have been more. But I am pretty sure that no manufacturer ever attempted to put that amount of power and torque into that configuration, remember we are talking about a design that commenced in around 1975.

When the 928 S4 was introduced the torque derived was increased quite significantly. The torque tube, besides transmitting power, also acts as a transmission shock absorber and it does this by winding itself up to some extent. Hence as torque is applied the drive shaft will try to shorten itself and the flex plate does its stuff.

Having established that logic, imagine now that for some reason the pinch bolt holding the flex plate hub onto the torque tube splines loses its grip slightly. You are doing 60 mph and you floor it. If your car is running properly, that kickdown will instantaneously throw your head back into the headrest, such is the shock load impulse. Now if our perceived scenario is correct, this pulls out the hub ever so slightly. You then take your foot off the throttle but that does not lurch your head forward. Hence you can reasonably assume that there is not a comparable reaction in the other direction. Thus the spline takes up a new position in the hub until the next time it is yanked forward. Bit by bit the total displacement increases until it reaches a point where the thrust on the crank causes the main thrust bearing to give up the ghost and then you need a new engine and a healthy bank account. To this day Porsche have not issued issued a service bulletin telling their dealers to check the flex plate position on automatics but interestingly they very quickly proposed a solution to a problem that does not exist. Hmmmm!

Unfortunately to date Porsche do not seem willing to agree with me that a problem exists on the automatic.

It may well be possible to design a joint that slides but I presume the extreme load conditions in these cars ruled this out. I would like to see how GM have tackled this one on the new Corvettes now that they have slipped the gearbox to the rear of the car 25 years after Porsche thought of the idea.

With regard to the vibration I have asked the listers several times if anyone else has experienced the same vibration that I did but did not receive much response. My car is very smooth up to this engine speed [3050 rpm] and very smooth above it. My instincts tell me that while my flex plate is positioned correctly the vibration is very much attenuated to the point where it is hardly noticeable. I think the S4 engine has a resonant frequency period at this speed hence the great lengths they went to in designing the engine mounts. I recently drove a 928 GTS that vibrated like a dog at this speed. Whether it is related to the mounts I do not know but mine were replaced just before I bought the car 3 years ago. Other listers report bad mounts cause a lot of vibration at tick over up to around 2800 rpm. Your guess is a good as mine.


Fred Rourke
1990 S4 auto-

[The next day]

Bad news I'm afraid!

Checked the hub position this morning and compared to the paint index marks, the damn thing had migrated again back to the 3mm of thrust position. We slackened off the pinch bolt and back it popped. So much for the new bolt plus 110% torque theory. I am now going to try your method with the loctite 290. Unfortunately they did not have any of the stuff in stock and they ordered some from Dubai. How long do you recommend to allow the stuff to cure? Overnight, 24 hours ??

Having read your post again and having had a close look at the assembly and the Loctite catalogue I now understand fully what you are saying as the Loctite 290 is designed to migrate along the shaft hence it can be used without disassembly.

In your post, if I understood it correctly, you indicated that the flex plate could not be disassembled without removing the TT. Please correct me if I am wrong, but looking at the design and having spoken to the Porsche mechanic, he indicated that by undoing the bolts that mount the hub to the flex plate, the flex plate itself can be slid out and the hub slipped forward to expose the splines. If this is the case do you think this is a better option in that the shaft and hub splines can then be better degreased and allow the loctite more complete penetration if it is applied in the disassembled condition?

If we attempt the procedure in the assembled condition we also wondered about using carb cleaner to try and flush out any dirt or grease from the assembly- did you try this or something similar?

Finally, the equilibrium position of the hub on my drive shaft is such that the rear face of the hub aligns more or less with the rear end of the splines on the drive shaft. Do the cars you have worked on have a similar set up or does the hub sit a little further forward?

Sorry to trouble you and look forward to hearing from you.


Fred R

Please understand that I don't have an automatic transmission car to play with, as David and I have a pretty good collection between us, but no automatics! This means that, "Not being hampered by facts, I can speak freely on the subject."

Would it be feasible to drill the hub, penetrating perhaps one mm into the splines, tap the hole using a bottoming tap, and install a cone point set screw (grub screw to Fred)? If so, the added locking provided by the set screw bottoming into the drilled recess should positively lock the hub in place. A short set screw would be close enough to the same weight as the material removed so as to not impact balance, especially at the small radius involved.

Wally Plumley
928 Specialists

I also don't have an automatic to experiment on but would be afraid of drilling into the shaft. That shaft looks very small to me to be transmitting the power of a 928 and the shaft does break sometimes on the automatics. I think that it would be wiser to wait and see if the "Loctite fix" is successful on other cars before trying the setscrew route.

On any other "normal" drive this would be a very good solution.

Earl Gillstrom '88 5 Speed


On the package, it says 10 minutes for 290 and 20 minutes for others and full cure in 24 hours for all. The car that I checked and found no slippage was cured for 24 hours. Car "D" was cured for 1 hour, hopefully we will check this car soon. So I don't know if the 24 hour cure is necessary or not.

>In your post, if I understood it correctly, you indicated that the flex
>plate could not be disassembled without removing the TT.

I don't know what is on the flywheel side of the hub. Maybe someone else can answer the question of hub removal without TT removal? We thought of trying to "prime" (clean) the parts while assembled but did not. I guess it depends on the condition of the parts when they were assembled. We think the car that we had success with had never been apart since new. We should try to get an answer to the question of cleaning while assembled, or disassembly without TT removal. The Loctite has to work better when the parts are cleaned first. On the other hand, it worked for us (so far) doing it the "quick and dirty" way. Anyone?

>Finally, the equilibrium position of the hub on my drive shaft is such that
>the rear face of the hub aligns more or less with the rear end of the
>splines on the drive shaft. Do the cars you have worked on have a similar
>set up or does the hub sit a little further forward?

I only checked this on the success car and the hub is .158" forward of the beginning of the splines. I do remember that the other cars did seem to be in about the same position after unclamping. It may depend on where the rear end of the driveshaft is clamped and may vary from car to car.

Earl Gillstrom '88 5 Speed


I'll give the Loctite procedure a go and then see what happens. As Constantine has reminded me, the other thing that Porsche suggested in my dialogue with them was to replace the whole flex plate. For a non service component I thought that was a bit cheeky but then maybe it is a factor, although how the splines can wear on a non moving part is beyond me.

Wally has suggested fitting a grub screw which may be a good idea as well and if all else fails maybe we should weld the bloody thing! Either way we will get there one way or another and hopefully our posts may help others to avoid engine destruction.

Let's see what happens


Fred R

The only difference (other than hp/torque, which may be a contributory factor) between the autos having (or being prone to) this problem ('85 and up), and the ones not ('84 and older) is how the driveshaft (and crank end play) is attached/adjusted at the _front_ end of the system. Volume II, Section 39, pages 47-48. See Note at bottom of page 48. Notice the parts deleted as of '85. Porsche's claim that "Adjustment of the drive plate is no longer required." Obviously this turned out to be _wrong_, hence their later recommendation the to replace the pinch collar bolt and torque to 110%. A "fix" that is apparently not adequate. Their claim of ignorance of any TBF problem is, of course, to limit their liability for a change (made no doubt to cut service time/costs/complexity) that eats motors. Their "fix" mentioned above is to correct a problem they claim doesn't exist. I doesn't get much plainer...... When looking for the cause, you must look at what the difference is between the cars not having the problem, and the cars that do have the problem. And that, is mentioned above, quoted on the pages referenced.

Best regards,