Aluminum Flywheel keeps coming loose!!

[jebbach911]

Hi Guys, need advice! Here's the story. Two times in a row, my aluminum flywheel (Findaza) has come loose. The first time I disassembled and cleaned everything with carb cleaner and retorqued the bolts to 120 ft. pounds. I got 2 autoX's out of it and it has happened again. I have purchased new bolts from our host and I'm removing the motor now. Next race is this coming Saturday so I have at least till then to "get er done".

Now the important info. Before this started happening, I changed my rev-limiter pill from 7000 to 8000. It made a huge difference in my lap speeds because I wasn't bouncing off the rev-limiter before a turn. I'm putting the 7000 pill back in for this next race but can you guys tell me anything else I could do to keep this from happening again?

Thanks in advance for any wisdom

[Henry Schmidt]

LockTite and a rev limit appropriate for the cams you're running.

[jebbach911]

Thanks Henry! I have been using Loctite red on the bolts. I think I will order 7200,7300, 7400 and 7500 rev limiter pills.

[tadd]

Wow, thats really weird loosing an aluminum flywheel. THe MOI reduction is supposed to help with the shedding problem. Herr Weiner says it is a 8000 rpm problem on the long strokes. Not 7900, but 8k. Its supposed to be a 4th harmonic of the gas torques.

Walt from this board has proffered excellent luck with overtorquing the flywheel bolts (suggested to him by bruce anderson IIRC) to 8K. Search pelican under flywheel bolts and youll get a bunch of hits with his procedure. He uses a stock plate.

tadd

[JJ 911SC]

Originally Posted by Henry Schmidt

LockTite and a rev limit appropriate for the cams you're running.

Henry

Was going to recommend Loctite from your answer to my question

Did go dry upon your recommendation.

Thanks again

JJ

[tadd]

Dumb question, but have you had your tach calibrated? I've seen them +/- 1k out with age.

t

[Steve@Rennsport]

Keep the peak RPM's below 8K and it won't come loose (provided its installed correctly).

[KTL]

Originally Posted by tadd

Dumb question, but have you had your tach calibrated? I've seen them +/- 1k out with age.

t

That's a good point. Check the rpms with another source, like a timing light with built in tach for instance.

[Walt Fricke]

Jeb

The next thing for you to do is put your motor on a chassis dyno and get your torque curve. You can then use that to see where the optimim shift points are. It is possible that you are winding up the motor beyond where it makes sense. Though perhaps not - you are autocrossing, so I am guessing you seldom, if ever, get into 3d gear. If you have stock gears, you usually can't rev high enough in first to hit the optimal shift RPM into 2d. Plus in autox it is often worth reving above nominal best shift points in order to avoid a shift. An autocrossing buddy with a penchant for data once calculated that each upshift in an autocross cost 1/2 of a second!

But firstly, you have come across something which bedeviled the factory race engineers in the early 1970s when they tried using the 70.4mm crank in race motors, so don't feel you made any mistakes in assembling your motor.

I have had good results with 150 lbs/ft and carefully applied (so none can get between the crank and flywheel mating faces) Loctite red on my 2.7 race motor, which routinely went above 8K, and shook itself loose promptly before I took these precautions. This has been on steel flywheels, but lightened ones, including a genuine Porsche RSR flywheel with the 4 pound RSR pressure plate.

Based on that, I am dubious that the MOI has much to do with things, even though you'd think it would. But I have not used a stock flywheel, much less the 8 pound or 12 pound stock pressure plates on this race motor. Of course, for autox you want the assembly to be a light as you can get it.

I am facing a similar question with my aluminum flywheel (don't know if it is a one-off or an early Fidanza, as I traded a roll bar for it). It is going on a 66mm crank for a 2.8SS, and that will be reved above 8,000 if the DC80 cam performs with this motor at all the way it does with my 2.8 long stroke 2.8. I am inclined to keep doing what I have been doing because I see no downshide. My instincts tell me that the extra compression on the aluminum flywheel won't squish it enough more than 110 or 120 lbs/ft to cause a problem. Porsche certainly ran its 66mm crank race motors up pretty high. But I am basically guessing here.

So do the 66 mm cranks need extra help up above 8,000 RPM, or is stock torque on the bolts fine despite 8-8,500 RPMs?

[chris_seven]

If the problem is caused by a torsional vibration, which I agree with Tadd are likely to be 4th or maybe even 5th order gas torques then the impact of a flywheel with a lower than stock MOI are likely to be twofold.

The resonant frequency of any torsional vibration will change as the mass-elastics of the crank will now be different and if the runing speed approaches the resonant frequency then there will most likley be problems.

The stiffness of the crank is likely to be the dominant factor, however, and as the 66mm stroke crank is stiffer then the 70.4mm crank I would expect it to have a higher critical speed and the effect of the light flywheel to be less.

I would want to see a 'headroom' of resonant frequency being at least 1.4 or perhaps even 1.7 times the maximum operating speed.

The second effect will be that if the change in MOI doesn't significantly effect the 'critical' speed then the inertia forces that are trying to 'shuffle' the flywheel will be less so bolt tightness shouldbe less important.

I have to say of you get too close to critical speed this would tend to be irrelevant and I wouldn't be surprised to see bolts shear.

It would be interesting to look at the mating face of the crank/flywheel to see if any 'fretting' has ocurred as this would be a good witness that here is some torsional vibration and shuffling taking place.

I recently saw a 240Z crank that had hollowed out crank pins in an effort to reduce weight and there was a significant amount of 'brown powder' in the interface which shows that fretting had taken place. You won't see brown powder with a aluminium flywheel but there should be a witness of some type.

I agree with using stronger bolts and higher preloads as this will increase clamping force and make shuffling less likely.

The best way to asses the torsional vibration problem is to rig up an engine dyno with a device such as a Bently Nevada Key Phasor and look at torsional behaviour. Without being able to measure torque phase this is virtually impossible as dynos generally only measure mean torque.

[Walt Fricke]

Chris

My mention of 4th harmonics, especially in the 70.4 crank, comes from articles in Panorama, which are based on something from the factory at some point - a "node" ends up right at the flywheel/crank mating surface. Apparently a high point rather than a low point, or whatever is bad for this purpose. Frere mentions the effects, but not the cause.

I can attest to the fact that fretting occurs once the two surfaces move against eachother at all. Metal transfers back and forth between the two, destroying the flywheel for reuse. Since this movement also backs out the bolts, it promptly becomes very destructive. Did the same to one crank, rendering it useless, and the second time I caught it earlier and the end of the crank could be trimmed, though the flywheel mating surface was cratered. I still have it, but suspect that any repair would not be cost effective for a $400 part.

After upping the clamping pressure I have never seen any evidence, be it marks or any powder dusting, of any movement.

The 917 motors (interestingly with casting numbers of "912."), take their power via a gear in the middle of their 12 cylinders. It was explained that this precise point also was at a point of inflection or something for harmonics, but a good one rather than a bad one, which left these powerful motors immune to these issues. There was an article in the Porsche Club of America's Panorama magazine describing this aspect of the 917 (which is where I picked up 4th harmonic), but I can't find it in the index to the PCA's Upfixin series of collected technical articles.

[Flieger]

The 917 crank was basically an enlarged 911 crank I believe, since there were two rods to each throw. I also think I heard that if they took the drive from the end then there would be a bad harmonic vibration just in the rev range they wanted to use- 8000rpm and thereabouts. So I think the good node on our engines is in the middle.

[EBS_Don]

Maybe not the ultimate "fix" for this situation but we now have ARP flywheel bolts for 6 bolt cranks...

[chris_seven]

There is a very good book which has all the basic math for torsional analysis and decribes all of the modes you see in various engine designs - I have had the view that the counterweights on the 911 cranks are to deal with these high order vibrations for some time now and if fretting occurs - as witnessed- this is likely to be the main cause.

There used to be a detailed ASME paper discussing the 917 engine design and it was for some time available on the web but I have't seen it for some time now but I will search again.

There are two basic types of engine vibration to worry about and that need analysing.

I think the 917 suffered quite badly from lateral vibrations and this is where the descriptions are most appropriate. Lateral vibration is affected by balance and are best described as whirling. If you only had two bearing (Austin 7) then the crank would tend to behave as a skipping rope and this is why 3-bearing engines can wipe out the centre main at high rpm.

Higher modes will result in the nodes as described and if the system reaches resonance it would generally result in a catastrophic failure.

Taking the drive from the centre of long shaft has been done before and some early Alfa Grand pix engine had a central drive to the cams.

Lateral vibration is easy to measure with accelerometers fitted to the motor when run on engine dynos .

Torsional vibration is much harder to detect and this type of vibration is generally referred to as 'invisible'. This kind of vibration will not result in the apperance of nodes as it is more of a wringing effect.

In-line 6 cylinder motors suffer badly from this type of problem and this is why they all have 'crankshaft dampers' fitted to help absorb these damaging vibrations.

The Hirth coupling is another way of helping to elimiate this type of problem.

I would agree that bolting up the flywheel can be effective and I would almost certainly use ARP or similar high strength fasteners.

The other point to consider, however, is that bolting on the flywheel with a higher clamping force will only eliminate fretting and/or the bolts coming loose.

It won't reduce the level of torsional vibration or the stresses generated in the crank pins - which are the 'softest' part of the structure.

If these stresses are high enough then they will 'eat' the fatigue life of the crank fairly rapidly.

I think I read an article concerning the 2.8RSR having new cranks after every 24hr Endurance Race to avoid just such failures but I can accept I may have been dreaming.

I still think it would be a great idea to rig up a dyno with a torsional vibration measuring system and look at what is happening in detail.

[jebbach911]

Thanks Guys for all the input!

Using all of everyones suggestions I cleaned and tapped the crank threads to get the old red loctite out, used new crank bolts from our host, put red loctite on threads and a tiny bit of oil under bolt heads to aid in accurate torque, torqued crank bolts to 150 ft. pounds, installed 7800 rpm pill in MSD box.

Ran first AutoX yesterday with awesome results! Won my class, won PAX and got TTOD.

[Flieger]

Did you actually use a tap or a thread chaser? If you used a tap you have cut the threads deeper and may have issues with looser bolts. Though, this effect is not as pronounced as it would be in Aluminum.

[Walt Fricke]

Yes, I just run the bolts in to clean up things in the crank, take them out, spray with brake cleaner, and clean/dry out with air from the compressor.

For heavier duty threaded hole chasing I just cut four slots up an old bolt with a hacksaw, but haven't found the need to do that on a crank.