GregBBRD

Every engineering company on the planet has top tier engineers and lower level engineers....Porsche is no different.

Due to the "problems" that "show up" starting about 1990, I've always assumed that the majority of the top tier 928 engineers were either gone or moved to other projects, by that time.

By the time the GTS models were made, this became very evident. Some of the changes made bordered on sheer stupidity.

My own studies, when I decided to built my own torque tube shafts, showed that the 28mm shaft diameter concentrated the "twist" right at the base of the junction between the actual shaft and the reduced area leading to the splines. Note that this "junction" is not undercut to reduce the stress risers (like the 25mm shafts) but is simply a constant angle from the 28mm size down to the OD required to cut the splines.

Not Porsche's very best effort of engineering....

GregBBRD

It will probably work fine.

I don't preemptively remove 28mm shafts (I've probably got over a couple of hundred clients that still have these shafts), but I never leave one in, when rebuilding. My attitude is that by the time the bearings go bad, the shafts are probably bad, also. (It's my theory that Porsche felt the exact same way, which is x why they never offered "rebuilt" torque tubes.)

Like I do with everyone I offer advice to, I wish you the best of luck.

Dave H.

same.

FredR

Greg,

That the 28mm shaft has a weak point is not in doubt but if it was an inherent weakness that is failing because it is "relatively weak" then my thinking tells me they would all be failing and it is just a question of when they decide to let go. On the other hand if an external defect was causing these shafts to fail [i.e. the clamp slipping] would that better explain what we see? My original S4 shaft failed after some 84k km [approx 50k miles] which was outrageous. Once I managed to fix the flex plate issue with the Loctite solution no issues to date since. Although I have no data, I am quite sure the number of S4's and GTS's driving around on 28mm shafts far exceeds those that have the 25mm shaft retrofitted. That being the case then logic says shafts should still be snapping with gay abandon and my query is whether or not this is in fact happening given that most enlightened owners will have some form of flex plate clamp mitigation in place.

Some clamps never show a problem others [like mine] do [or did in my case]. Of course for all I know mine could go pop tomorrow but unlikely I suspect. Thus my question as to whether the 28mm shaft is the culprit or the victim. A flexible coupling is designed to take out radial misalignment to a small extent and not intended to cope with forces that can deform the leaves of the coupling. As I have advised many times, when my flex plate coupling had slipped by 2mm I knew about it because of the vibration it generated at exactly 3050 rpm. Interesting a couple of years ago a fellow lister and new S4 owner posted about his issues. I advised him about the flex plate issue and TBF possibility and how a slipping clamp can lead to other issues. Some of our more experienced bretheren thought I was being alarmist. The owner involved did an excellent job of rectifying various other issues, took his S4 out for a spin upon completion and within a matter of a few miles his drive shaft duly snapped.

Will a 28mm drive shaft snap if it has no flex plate slippage issues ? Why did Porsche even change to the 28mm design even though early 87 S4's had the 25mm shaft fitted? Something must have triggered this surprising decision- could it have had something to do with the noise issue that requires a damper? Could the S4 design have caught them by surprise when they released it and some bright spark thought it might attenuate the problem given the change to the natural frequency of the shaft? who knows?

Rgds

Fred

GregBBRD

Fred:

I'm not sure of your point or how you want me to respond.

I spend time here, pass on what I've learned, and make honest evaluations and suggestions, free of charge.

Interestingly enough, when it comes right down to it, what people believe or choose to do, has zero affect on me....unless those people are actually my own clients.

I guess I should be grateful of the mistakes people make.....I'm repairing more and more of these cars, every single day....and a huge percentage of what I do is repairing things that have already been repaired once or twice before. The business has grown so large and there are always so many people waiting for a chance to get their cars into my shop, that we are moving to a larger facility in December, to be able to better service client's vehicles.

Here's the reality.....I've never seen a 25mm automatic torque tube shaft break. Zero. Never.

I've seen many 28mm shafts fail. I've taken out many others, preemptively, when I rebuild torque tubes.

It sounds like you think I should be telling people to leave these shafts in, when they have the torque tube apart and have the opportunity to change them.

I just don't think that is going to happen.....

chart928s4

As a TT breakage victim I concur with Greg's assessment but I feel like the odds are a new 28mm shaft won't break right away. Also it will cost you many hundreds extra to upgrade Greg's 300M shaft from a regular shaft.

The big counterweight is that the failure is devastating and disabling. I personally was so traumatized by it that I upgraded all the possible parts. I never want to worry about that particular problem again. Also the Super Clamp provides a similar near zero chance of future TBF.

FredR

Greg,

You appear to be going down a negative avenue discussion wise when there is simply no need. I have raised a query about what is going on - not a statement of fact and what I am questioning is whether the apparent weakness of the 28mm shaft is the culprit or the victim as it were.

If my drive shaft were to snap [once more] I would be on the phone to either you or Constantine to suss out availability of a 25mm shaft- no need for further discussion on that one. If my torque tub bearings were to fail [as expected any time] I would go for Constantine's super bearings [three of them] no need for discussion on that one. When I have the 28mm shaft out I would have it thoroughly examined for signs of cracking [dye pen/UT or anything else available at the time] and if there are no indications of cracking I would consider re-using it. If at such time I had by then covered a highish mileage on the shaft I would probably replace it altogether. That your shaft is a sound long term proposition nothing to debate on that one.

Where I have a degree of uncertainty is at what age/mileage should one throw in the towel?

The technical questions I ask in general are:
1. Why did Porsche retain the 25mm shaft for the early S4's and then ditch it for the 28mm variant [I do not buy it was to use up what was left in the earlier parts bin but...]? The only possible answer I could come up with was that given the automatic has two bearings did some engineer suddenly perceive the shaft was deflecting more than it should so beefed up the body diameter to increase stiffness across the bearing span? For sure they could have designed a new flexplate clamp for the front had they been bothered- more a problem for changing the rear coupling I suspect. Whatever the reason we all seem to think it was a bad move design wise.
2. Why does the 28mm flexplate clamp seemingly fail? If it was inherently too weak then all clamps would slip and have the problems I had 18 years ago [or worse] but this is just not the case. Many examples [i.e. a large majority] seem to survive without the lengths some of us have gone to.
3. Some 28mm drive shafts survive very high mileages- how can this be if they are inherently weak design wise?
4. Could it be that the 28mm shaft is somehow inducing vibrations that in some cases cause the flex plate clamp to slip?
5. Could spurious vibrations cause the bearing mounts to move? We assume it is caused by degraded rubber but..?

I am convinced my S4 28mm drive shaft failed because of the clamp slippage, this induced vibrations that caused fatigue at the perceived "weakest point". Did it slip because of an inherently weak clamp design or did other factors induce it to slip thus overcoming what should have been adequate clamping force? Increasing the clamping force solves the immediate problem but not necessarily the root cause.

I doubt we will ever know the real answers to the above and in engineering the answer often lies in fixing the symptoms rather than the cause. Some folks believe that the Loctite solution does not work- I have 18 years experience telling me it does but what I cannot say with 100% certainty is whether it will still be holding tomorrow. What I can say is that I had a problem that Porsche [or anyone else] could solve that was going to wreck my engine and/or cause another drive shaft to fail- today 18 years on everything is still in one piece thanks to a $6 bottle of glue. Will my drive shaft fail anytime soon- who knows?- but if it does expect a long distance phone call.

Regards

Fred

GregBBRD

Some notes, which may/may not help.

Facts:

Both the 25mm and 28mm shaft couplers (front and rear) are identical.

The clamps do not migrate. The front clamp is welded to the splined piece that attaches to the flexplates. The shafts migrate.

The Modulus of Elasticity says that all steels twist the same amount.

The stock shafts are made from a special steel that has a good "memory"....it has the ability to "untwist" to it's original position better than other steels. (I'm not near my data, but as I recall, it is 1011.)

Observations:

Shaft migration does not occur at lower mileage....regardless of how a car is driven, I've never seen shaft migration on a car with less than 30,000 miles. It rarely occurs on cars that are driven gently....even at very high mileage. Cars that have been driven more aggressively will have shaft migration earlier. As these cars have passed from "higher end" owners into the hands of enthusiasts, shaft migration has become more prevalent. .

Beliefs:

The shafts twist more, as they age. This occurs because of micro cracks (fatique) in the shafts. 28mm shafts fatigue and fail because their design concentrates the twist in the "weakest" area....where the shaft diameter transitions from the splined area to the 28mm shaft diameter.

25mm shafts COULD potentially fail from the same micro cracks, however because these micro cracks are formed over a much larger area, thrust bearing damage "stops" the vehicle long before the shafts can actually break.

Porsche knew that the shafts twisted, from the very beginning.They knew that when the shafts twisted, the shaft would get shorter. They also knew that the front clamp would not keep migration from occurring, which is why ever automatic shaft (prior to 1987) had an additional "migration limiter" assemblage of parts consisting of circlip, thrust washer, and shims.....to ensure that the shaft could not migrate.

Porsche engineers dropped the "migration limiter" pieces in 1987...presumably because the new flywheel needed for the new injection system greatly limited installation of these pieces. Shafts, with the extra length and required circlip groove, continued to be made into early 1989 models.

Porsche knew, from long terms testing, that shaft migration became a serious problem by 1990. Their "solution" was to introduce a larger 28mm shaft in 1990.

The 28mm shaft still had twist, but that twist was initially less than that of the 25mm shaft. Unfortunately, the twist was concentrated, as described above.


Conclusions:

All "additional" clamps, bottles of Locktite, etc. on the market do not solve the twist problem from fatigued shafts....they are a "patch", not a solution.

Clamping the shaft more securely transfers the thrust problem to the front of the crankshaft thrust bearing....away from the rear of the bearing.

Long term, the only correct solution is to replace the shaft. Porsche knows this....this is why they never "rebuilt" torque tubes with $5.00 worth of stock bearings....the shaft is the most expensive part of the entire torque tube.

I make all my shafts with the "early" circlip groove. Instead of buying a very expensive "clamp". It is possible (I've done this for over 20 years) to simply fit the original "thrust migration" circlip, thrust washer, and shims, for just a few dollars.

Be 5% smarter than what you are working on. Instead of buying an expensive clamp to mask the problem, buy a shaft, and cure the problem. Use the stock "thrust migration" pieces and forget ever ruining an engine.

Constantine

Greg,

Sorry but disagree with your hypothesis on the root cause of TBF. You guessed this would be coming, right?

Drive shafts getting more twisty as they age due to micro cracks forming is not correct. If that were the case you would have many failures of both the 25mm and 28mm varieties since their "modulus of elasticity" would be reached in short order if there were micro cracks forming in the metal. Micro cracks would not just stop forming at a certain point, but would continue on to larger cracks and so on until failure. Micro cracks are small cracks that weaken the metal causing further cracking.

It would also stand to reason that if the drive shafts are not concluding their stretching to failure due to the engine first suffering TBF, you would have a lot more TBF'd 928s out there. I know you have rebuilt many 928 engines, uprating their HP/TQ figures and re-used old drive shafts in the torque tubes when you put the 928 back together. If your hypothesis is correct, then those 928s should have broken those drive shafts in short order. Let alone the many owners who have boosted their 928s and have been running them with no drive shaft failures, except maybe the 28mm drive shafts due to their built in stress riser.

The pieces you talk about (circlip, bearing and washers) that were placed on the front of the 25mm automatic drive shafts were discontinued to be used in 1984 according to the Porsche WSMs. However the WSMs instructions for placing them must be followed to the letter or the problem really hasn't been fixed. Not many people get it right since the instructions are a bit complicated, to include Porsche field techs and is why their use was discontinued.

It is great you are making 300M drive shafts, we have used a few in our TT rebuilds for customers as you well know. But to pin TBF of the 928 automatic engines just on the drive shafts getting more twisty as they age is a bit of a stretch, no pun intended.

Cheers,
Constantine

GregBBRD

OK. Not going to argue this stuff....just takes too much energy.

Few stumbling blocks that interfere with your theory.

Have you ever seen a broken 28mm shaft close up? They don't just shear....there are many obvious cracks lengthwise through the shaft. These are larger cracks that start as micro cracks.

Grab any used 28mm shaft and MagnaFlux it....every single one done this to has cracks, right where I say they are at.

Grab a 25mm shaft that has been in a car that has had TBF and MagnaFlux it....it will be completely riddled with micro cracks....

Why don't original cars have shaft migration until 80,000 miles, except in very rare cases where cars have been driven extremely hard? (I've never seen a shaft migrate before 80,000 miles, personally, but have heard of this happening.)

The shimmimg of the stock shafts with the circlip and thrust washer requires grade school addition and subtaction. Super simple. Getting the circlip onto a shaft already installed, in front of that late flywheel....very difficult.

I had to come up with a "solution" to shaft migration as early as 1993 or so, on higher mileage 928s. Like you found, Locktite and even over tightening the bolt did nothing. For me. The stock early pieces work perfectly....for about $5.00. I've been doing this, for over 20 years.

I do use your clamp, when clients insist. Also works good! Costs a bit more than $5.00.

FredR

Greg,

Thank you for sharing that- as I am concerned I am aligned with most your thinking. I concluded many years ago that driving style influences the chances of clamp slippage leading to mechanical failure as does the initial running period- I have opined on this list many times that when failures occur, they seem to happen in the 60k km to 100k km interval.

Where I perceive a difference in our views is that this interval is needed to induce some kind of failure in the clamping system that permits the shaft to start slipping through the [by then inadequate?] clamp. Once the shaft migrates through the clamp it then becomes an issue of whether the thrust bearing goes or the shaft breaks due to the micro cracking mechanism you outlined above. The two factors acting jointly cause" premature failure in examples driven hard. My previous owner was a Dutch chap who did not hang about and the failure occurred shortly after I acquired the 90 S4 in 1998. Thus in order to prevent these low mileage "premature failures" the clamp integrity is absolutely essential. How this integrity is achieved is academic compared to the consequences of loss of joint integrity. At the time this happened on my late S4 in 1999 there were seemingly no options to solve this albeit reversion to the earlier system would have been possible as I later learned.

After I lost the S4 around 2005, I then acquired the GTS. Again this car had been driven wlith some serious intent by the original owner who I know well. When the original motor had covered something just over 110k km the motor experienced TBF. In this case the motor failed failed rather than the drive shaft and the flex plate was seriously mangled- I saw the damage on the car! At the time the car was then owned by its second owner who was somewhat distraught. I made him a fair offer for the remaining "detrius" given it had a superb chassis and I had the S4 motor to put in it. Upon acquisition we pulled the motor and the thrust bearing web was cracked- game set and match for that motor as expected. I then had to plan the recovery project and after the motor was removed we puled the torque tube for inspection. I had a mechanical inspector friend of mine organised some NDT on the shaft at the base of the taper. We could find no signs of distress so decided to reuse the shaft. I do not drive at high speeds these days but I do push the car hard on acceleration and through the bends. The shaft now has some 160k km on it and last time I inspected it, it had no signs of distress. The interesting question now is "how long will it last"? As stated earlier I will drive it until it fails or until the bearings give and then will make a [hopefully] informed decision what to do shaft wise.

Given what your customers spend to acquire your services spending 600 bucks or whatever to acquire a new shaft when things are apart and thus get a superior solution makes every sense. Had I had this option in 1999 I would have done just that. If an owner has a stock shaft in seemingly good condition then clearly it is up to the owner to make a considered decision whether he retains it or not.

Do shafts fail that have not had any issues with the clamp letting go? - I have no position on that one all I can say is that the half a dozen examples I have seen fail all had issues with the clamp not holding the shaft and I have not seen one fail when the clamp was holding.

What would be nice to know is just how many shafts have failed and why and to put that in context with the sample population. I suspect that circa some 20,000 examples were built with the 28mm shaft. How many of those have failed and how many are still intact? If local experience is anything to go by some 10% may have failed. although that is alarming, some 90% would still be driving aorund the original 28mm shaft. Of those that survive how many are driven hard and what mileage have they survived to date? It would be good to hear some thoughts on that assuming some listers have such.

rgds

Fred

Constantine

Hi Greg,

Yes, have seen my share of sheared 28mm drive shafts. They sheared because of the necked down stress riser, not because of micro cracks.

No one has kept any information to support your statement about drive shaft migration happening around 80k miles, not even Porsche.

Bottom line is that 928 owners are fortunate to have choices to choose from when rebuilding their drive line.

Jerry Feather

Greg, I have just started to do some "engineering" of my own about the 928 drive shaft(s) mainly in advance of my need for a shorter one or ones in the Radical Custom 928 Project. I have been looking into the tooling necessary to cut the splines and the design of the shafts in general. I read through much of the old 2013 thread about your replacement short shafts and much of the misinformation provided by others, but that project and some of the current discussion give me some concern.

In one of these recent threads I posted a question about why the drive shaft is turned down to what appears to be the minor OD for the splines and which appears to be the weak spot in the shafts, and which appears to be the location of what I suspect is the most common or even only point of failure of any of these driveshafts. No one took any notice of my question; but upon further consideration of that issue I think I have figured out the answer myself.

In the earlier discussion (2013) much was apparently made of the thought that rolled spines are stronger than cut splines. I think that is true; and from what I found on-line(?) they may be about 35% stronger. However, I have concluded, as I think you probably did also, that Porsche only rolled the spines in question because that is much easier, quicker and cheaper than cutting them. From the videos online and from my own vision of cutting these splines I suspect it will take between 15 and 30 minutes to cut one set of splines, and I suspect that is even true in any kind of CNC setting. That is much too expensive and time consuming to do several thousand of them. Thus Porsche had them rolled - - not because they are stronger, but simply because that was cheaper.

However, the downside of having them rolled is that it appears that the shaft must be turned down to the minor size of the splines on each side of the splines for what appears to be about 3/4 inch, probably because of the process, machine and/.or tooling used to roll them. That requirement is what in fact has given rise to the weak points in the shafts; and is very apparent in being exactly the point that all the failures occur.

The reason I quoted your statement above is that I think it is not correct in two respects. First is that the taper in the 28mm shaft reduces the size from the 28mm down to the minor spline diameter but not directly to the splines, since the pictures above in this thread show the taper ending about 3/4 inch away from the splines. Then the shaft is a constant diameter up to the base of the splines.

The other thing is your conclusion that the shaft must be reduced in size for some distance next to the splines to allow for cutting them. That is, I think, only needed for rolled splines not for cut splines. The cut splines can be cut into the body of the shaft outside of the splined area and that will leave much more shaft material to keep from weakening the shaft at the point it is otherwise the weakest from having rolled splines.

My planned drive shaft will not be turned down to the small diameter next to the splines I will be cutting in it, and my short shaft, if I have to make one of those too, will not be reduced in size in the middle between the sets of splines on each end. There is simply no reason for it.

GregBBRD

Interesting observations.

I'm not in town and able to look at a shaft with your observations, although I certainly will.

My splines are not CNC cut, but cut on very old, very sturdy machines that are custom made to only do this job. My machinist has spent years locating, purchasing, and rebuilding these specialty machines. (He has many CNC machines for all the other operations, so it actually never occurred to me that spines could be cut on a CNC.) BTW....This operation is very time consuming....the splines are cut slowly, requiring many passes and resetting of the machine to get down to the proper depth.

I "moved past" the whole cut versus rolled spline debate when it was pointed out that jet engines had cut splines. If cut versus rolled was such a significant problem, I doubt that multi-million dollar jet engines, with the huge liability concerns, would use cut splines....

For what it is worth, I've had absolutely zero issues with any of my short shafts or torque tube shafts....in any application.....and some of these pieces have been subjected to high torque testing behind some engines with the highest torque of any other 928 engines.

Also, for what it is worth, I've got some of these shafts "clamped" with the stone stock Porsche clamp....just for experimental purposes. Zero shaft migration....

Lastly, since I build so many engines, I've had the opportunity to see and rebuild several engines which have been clamped with Constantine's clamp for several years (while retaining the stock torque tube shaft.) The wear that occurs on the FRONT of the thrust bearings from these old shafts twisting and getting short is very concerning.

Constantine

Greg,

Did you see the front of the thrust bearings of these engines with which you are very concerned about before the installation of our clamp? This seems to be a bit of hyper ventilation about a non-existent problem. Let us know when you run across a 928 engine with "reverse" TBF due to your theory of the stock drive shafts twisting so much. I have yet to hear about that ever occurring.

Are you backing up your claims of your drive shafts twisting less by tests? Before we stood up the Super Clamp we did some testing to verify the clamping force of it against stock clamps.

Seems like a lot of anecdotal information being presented as established fact. That you stood up 300M drive shafts is great, like I've said before. We are standing up a new version of our own drive shafts with a few teaks to make them better in a few areas than the originals. However we will never float out the idea that they will twist less than stock versions or help stop TBF.