66mm stoke cranks

[chris_seven]

I am trying to figure out some of the details with regard to crank design and looking through various threads there seems to be a difference in opinion.

My main interest is 66mm stroke 2.0 litre cranks and particularly 906 Types.

Looking at a standard 66mm 911S/E crank the counterweights appear to be relatively small and not a full counter weight.

The latest 66mm Supertech crank seems to have full counterweights that are larger than the 2.0 litres and the 906.

As we only want the crank for 2.0 litre engines is there any reason to increase the counterweight size compared to the 2.0 litre?

Rod bearing diameters are also interesting. The large rod bearing of the 66mm crank will, of course have more losses than the smaller diameter used on the 70.4mm crank and it may be interesting to use a smaller diameter.

I have read that the 70.4mm crank has a wider journal?? And also that web thickness was increased but these comments seem to contradict each other.

I thought that bearing widths and hence web thicknesses would be the same for both cranks.

If this is the case then the reduction in stiffness in using a smaller diameter rod bearing is unlikely to cause edge problems as the crank would still be stiffer than the 70.4mm stroke.

As we will be needing to make rods we can use the NASCAR Bearings recommended by Steve Weiner and get away for Glycos.

The final question is concerned with cross drilling. I was considering grooving both end journals and cross drilling and grooving No4 main.

I realise Henry’s crank has all mains cross drilled but this is quite costly. Is just cross drilling No4 good enough?

The crank will be made in EN40B and gas nitrided so I think it will be relatively strong.

We have just modelled an H Beam Rod



We are planning to model an I beam and then compare weight to stiffness ratios before deciding on which design to adopt.

All comments gratefully received.

[HawgRyder]

The counter weights on the crank do not necessarily have to be bigger.
With the advance of technology over the years...the rod/piston combo has gotten lighter.
This means less weight to balance.
If you have a clean sheet of paper to start with...you are ahead of the curve...and the design can be changed to suit your needs.
BTW...are the bearings (NASCAR type) going to be Clevite or Michigan 77 ?
I have had great luck with those over the years in a lot of different engines.
In Chevy sizes...they are available grooved or pinned and even micro-grooved (looks like the veins in a leaf) for oil dispersion.
Bob

[BURN-BROS]

The weak area of the 911 crank is the thickness of the throws. 22mm rod width yields a 7mm throw thickness. This is acceptable for the 2.0/2.2/3.0 cranks utilized in high rpm applications. However,when Porsche changed to 24mm width, thy picked up a nasty harmonic above 7500 rpm and destructive @ 8000 rpms.

Also, I spoke to an Formula 1 Engineer that studied the Porsche crank. He stated that the crank did not have sufficient counterweight to offset the weight of the throw by itself. He concluded that due to the boxer layout and the empirical data that this "problem" must not cause harm.

[chris_seven]

Bob,

I have to say that NASCAR Bearings are a complete mystery to me and I am more than happy to take advice on this subject.

I have contacted Vandervell Racing Bearings about their Copper/Lead Race Bearings but the investment may be too high.

I agree with you comments about counterweights and the CAD model shows a rod weight of about 550 gms for the H beam and the I beam isn't finished yet.

I think that the 10cmg maximum imbalance relates to an ISO 1940 G1.5 - ish classification which would be extremely good as most car cranks tend to run at around G40 so balance shouldn't be much of an issue.

[Flieger]

From what I have heard the reduction in journal size on the 70.4 mm crank and the corresponding reduction in torsional stiffness was what led to the torsional harmonic mentioned above moving into the range of engine speeds which were regularly seen by the racing engines. This caused the flywheels to be shed.

I thought I read that the 917 engine had a bad torsional vibration but that this was below the rev range normally used.

Perhaps smaller counterweights can push the torsional harmonic to a higher rev range?

[BURN-BROS]

Originally Posted by Flieger

From what I have heard the reduction in journal size on the 70.4 mm crank and the corresponding reduction in torsional stiffness was what led to the torsional harmonic mentioned above moving into the range of engine speeds which were regularly seen by the racing engines. This caused the flywheels to be shed.

If that were true then the 3.0 liter SC crank would have suffered the same fate as the 2.4/2.7 crank. Porsche went back to the 22mm wide journal and all was well.

Originally Posted by Flieger

I thought I read that the 917 engine had a bad torsional vibration but that this was below the rev range normally used.

917 had the same 24mm wide journal as the 2.4/2.7/2.5ST/2.8RSR cranks and indeed had a torsional issue. And that was with a Ti rod....

The harmonics did not rear up again until the 964. The larger piston and increased stroke was too much...They solved the problem for the street car by decreasing the journal width to 21mm. The downside was there is not enough bearing material to withstand the requirements on the competition engines and were forced to run the 22mm wide cranks in the 993 RSR's.

[Flieger]

OK, I will take your word for it. I never hear of a 3.0 SC revving to 8,000 without a missed shift and bent valves.

Reducing length will have a similar effect to increasing the journal diameter in regards to stiffness.

[BURN-BROS]

Originally Posted by Flieger

OK, I will take your word for it. I never hear of a 3.0 SC revving to 8,000 without a missed shift and bent valves.

O.K. later 935's/962/GT1 etc...

[chris_seven]

Originally Posted by BURN-BROS

The weak area of the 911 crank is the thickness of the throws. 22mm rod width yields a 7mm throw thickness. This is acceptable for the 2.0/2.2/3.0 cranks utilized in high rpm applications. However,when Porsche changed to 24mm width, thy picked up a nasty harmonic above 7500 rpm and destructive @ 8000 rpms.

Does this mean that the web is only 5mm thick for the 70.4mm cranks with the 24mm rod?

Do you know what type of vibration occusrs is it a torsional or a lateral? Shedding flywheels sounds like torsionals

Max,

Do you think reducing the big end journal diameter has much effect on the stiffness other than to increase the effective length of the web when its diameter s reduced? When you combine this with the stroke increase and the possible thinning of the web then there must be an influence on the mass-elastics.

All engine have the potential to suffer from torsional vibration and as long as it is well away from the running range it shouldn't be an issue.

If the crank is very flexible in this area then there could be a witness from the edges of the rod bearings which would possibly show increased wear as a side effect.

[BURN-BROS]

Originally Posted by chris_seven

Does this mean that the web is only 5mm thick for the 70.4mm cranks with the 24mm rod?

Do you know what type of vibration occusrs is it a torsional or a lateral? Shedding flywheels sounds like torsionals .

1mm per side so 6mm.

I am no Engineer but I believe it is lateral. I imagine that the throws are bending along an axis from the center of one rod journal to it's opposite journal.

Porsche played with many journal diameters on the 911 engine since 1965. The The 3.2 production car had a bump in journal size along with the 3.6 to 35mm. But it did not help to control the harmonics that plagued the 3.6. They simply ran out of room.
Engineers settled upon 53mm diameter. And ran that dia. exclusively from the inception of the 9 bolt crank on the competition cars and the GT-3's.

[HawgRyder]

Chris...for a complete list..try Mibearings.com for all the available types for chevy sizes (their biggest line).
If the throws and mains are of Chev size...the world is yours.
They even have rod shells that have extra bevel on the outsides to shed oil and lube the cheeks of the rods.
The spider vien types flow oil all across the face of the bearing and create a hydro-type barrier to prevent wear.
We used these types in the drag angines for lower friction.
Bob

[Walt Fricke]

The 9 bolt 70.4 cranks will, indeed, shed the flywheel if reved about about 7,500 rpm without special measures being taken. Porsche had this problem with the 2.8 RSR, which used this crank (although it had a narrower bearing to allow a larger fillet radius).

The books say this was due to the wrong kind of 4th harmonic node (destructive?) unfortunately ending up right at the crank/flywheel mating surface. Caused the 6 bolts to loosen and back out. Happened to me before I had read a bit more than I have now. Per Frere, Porsche never solved this satisfactorily for that engine. I know they solved it pragmatically be using new bolts for every race. A friend knew of a team which tack welded the bolt heads in place, and they still broke loose!

US hot rodders solved it by using red Loctite and torquing to 150 pounds foot. That has worked for me for years, ever since Bruce Anderson told me to do it.

Would the fact that the bolts loosen suggest lateral vibration, overcoming the preload that the stock 110 pounds/ft produces? Me, I I couldn't tell a 4th order harmonic from a nunnery. But some of you guys know all about these effects and their causes.

The 917 takes its power off in the middle of the crank from a gear. Writers on the subject say that the harmonic node there is of the benign kind. I don't know if this was a happy accident or clever design, or some combination of both.

The 9 bolt 3.0 70.4 cranks don't seem to have this problem at all, at least not that one hears about as often as you hear about in its 6 bolt parent. Plenty of 3.0 race motors out there spinning over 8,000 rpm. Similarly, the 66mm cranks don't have this problem either, which was perhaps why it caught them out with the 2.8?

Next time I build a race motor I will send a 66mm crank with a damaged rod bearing to the grinders to have the journals turned down so I can use Clevite or the like. Will need new rods, too, of course. Guys like Bob have been praising this approach for a long time, and with the bad reputation Glyco has developed, spending this extra money is beginning to look like more of a necessity than a luxury.

[Flieger]

The 917 had the central power take off by design. It had been done before, however. It was not a new idea but Hans Metzger was still one smart guy.

[chris_seven]

Originally Posted by BURN-BROS

I believe it is lateral. I imagine that the throws are bending along an axis from the center of one rod journal to it's opposite journal.

I would tend to agree that the main vibration is not a torsional and as you say the throws are bending.

2nd Order vibrations (2 shakes per rev) are nomal behaviour for V6 engines including 180 degree configurations. These vibrations would be longitudinal and are associated with crankshaft behaviour. I can see them breaking bolts if the resonant frequency coincides with running speed. Softening the crank in the longitudinal direction - increased throws and thinner webs- would depress the critical speed at which resonance occurred.

A 4th order vibration (4 shakes per rev) is much more likely to be a torsional vibration due to gas torques and I am not sure thinning the web by 1 mm has much influence on the torsional resonance of the structure - I do realise that I could be wrong- but it is quite difficult to model accurately.

If the resonance were longitudinal and the flywheel bolts didn't have enough preload to overcome the axial forces caused by the vibration then the bolts could suffer from fatigue failure.

By increasing the preload to a value greater than the axial force you would eliminate the effect of fatigue - same theory as con rod bolts - so it all seems to stack up.

[Flieger]

Chris, I have heard the 917 motor described as a "180 degree V 12" because it has two rods per throw, while a boxer 12 would have each rod on its own throw as does the 911 engine. So in that regard the 917 motor is sort of an enlarged, twinned 911 engine. I would think that this distinction would significantly change the mass-elastic behavior. If that is correct, then I do not think that what applies to V6's could be transferred to boxer 6's but I could easily be wrong.

[Walt Fricke]

The crank bolt issue does not seem to have anything much to do with fatigue failure. The bolts themselves look more or less OK (if you discount where the bolt heads were slightly machined by the center of the clutch disk during shifts and other times when they could move relative to the disk). Threads sometimes looked a bit worn, but basically the bolts are intact and about normal diameter. Not like the tree ring fractures from a stress riser (because not fractured), nor like the necked down before breaking (or if caught just before that point) of many rod bolt failures.

You engineers would have a better idea if the preload was removed (allowing the bolts to rotate and loosen) due to elastic stretch allowing the flywheel face to be periodically unclamped from the flywheel just enough that it could rotate the small amount loose bolts allow, over and over. With the bolts backing out a little each time. Or something else. The other characteristic of this problem is metal transfer from the flywheel to the crank, and crank to flywheel. I have one flywheel which looks lunar in that area. The crank, which looked equally bad, was tossed long ago.

Porsche tried to deal with this with a crank pulley damper on the 2.8 RSR.

But perhaps this was torsional? The harmonic added just enough extra what - angular accelleration? - that the clamping force was overcome, allowing the small back and forth motions, which tended to loosen the bolts when moving one way, but not to retighten them when moving the other way?

As to 4th or other orders, I am just repeating what was said in an impressive technical article in Pano long ago (memorialized in one of the Upfixins - collections of Porsche Club of America Panorama magazine technical articles).

[chris_seven]

Walt,

If the flywheel moves in relation to the crank then the bolts are quite likely to shear. If the vibration were torsional you would expect to see the effects of fretting on the interface between the flywheel and crank . I saw an Historic 240Z with a new re-designed crank that had tried to weld the damper to the end of the crank - clealry there was a torsional and insufficient clamping force.

The classic beech marking associated with fatigue is generally associated with low to medium strength steels.

Grade 12.9 fasteners may well suffer from fatigue failure without showing any beach marking.

If the fastener is at the top end of its hardness range (43 - 44RC) it is likely to suffer from fatigue failue after the propogation of a very short crack and this could easily be mistaken for a brittle fracture. It is good practice to select the fasteners with this level of hardness and not use them in very critical applications. These comments wouldn't apply to fasteners made of very sophisticated alloys with improved fatigue properties.

In a normal configuration the preload in the bolt exceeds any axial forces that are generated between the flywheel and the crank.

If there are second order longitudinal vibrations that result is stresses that are lower than the preload then no fatigue loading occurs and you can simply design the bolt to suit its maximum strength - say 80% of yield. This approach assume that the joint is much stiffer than the bolt.

If the longitiudinal vibration reached a critical resonance then the intertia forces acting on the flywheel on the flywheel could become high enough to cause the faces to seperate.

This wouldn't unload the bolt - it would increase the load in the bolt but allow a gap and the possibility of rotation as the joint friction would reduce. (Elastic deformation is caused by stress)

This would add a shear stress to an already very high tensile stress and could break the bolts.

If there were still sufficient friction to prevent rotation you could still have axial forces high enough to cause fatigue.

I guess the answer could be somewhere between the two possibilities.

If the engine survives being driven at 7500rom (assuming this speed produces the resonant frequency) for a reasonable time period before the flywheel drops off then I would tend to favour fatigue. If if failed as soon as it hit the resonance it is much more likely to be shear.

Improved clamping helps reduce both problems.


Max,

With regard to Flat 6 engines there is some work by Lycoming that defines 2nd order vibrations in Flat Sixes being mainly due to inertia and the dead weight of a propellor.

The 4th order vibrations in these engines were mainly due to gas pressures and can be absorbed by using the correct counterweight.


There is also mention of 4th order vibrations being absorbed by suitably designed counterweights. (Difference between 2.0 litre L/E /S Cranks and 2.0 Litre T cranks?)

[Walt Fricke]

Chris

Just to be clear: the flywheel issues the 2.8s had do not involve failure of the six flywheel bolts. They don't shear off, nor obviously elongate or in any way break. I've even reused a few in other applications, so the threads are not distorted.

So either a loss of clamping force due to harmonic vibrational axial loads leads to the bolts unscrewing and the mating faces to fret. Or torsional effects from harmonic vibrations exceed the clamping force, leading to bolts unscrewing and mating faces fretting. But I don't think fatigue is at work.

Increasing the clamping force would seem to mitigate either of these possible causes of bolts backing out, I'd think. In any event it solves the problem.

Perhaps the 9 bolt cranks don't have this problem, even with much smaller bolts (M10 instead of M12) at lower fastening torque because the increase in the diameter of the mating faces produces a sufficient increase in whatever one calls the ability of a joint to resist slipping under torque to overcome the forces trying to loosen things or otherwise induce the bolts to back out.

You'd think a flywheel with loose bolts would show up pretty clearly whilst driving on a track, but having driven perhaps as much as 20 minutes in this condition I can tell you that isn't necessarily the case. Shifting on track does not really require a complete release of the clutch if timing is decent. Where the issue shows up unmistakably is in the paddock. Can't engage first gear, because of the bolt heads rubbing on the disk's steel hub.

The first time this happened to me I assumed I'm goofed up torquing the bolts. I replaced the crank and flywheel. Next time I had the car out I recognized the problem more quickly. And I was able to poke at the flywheel through the drain hole on the bottom of the bell housing. I could move it. Time for another rebuild, though this time flywheel and crank end were salvageable.

[HawgRyder]

There may or may not be enough room to get a couple of hardened pins into the face of the flywheel (to stabilize the crank/flywheel from rotating).
VW did this on most of their cranks...the pin is a force fit in both the crank and the flywheel.
Bob

[Walt Fricke]

There is room for one, I think. But no one does this. It isn't needed if you add the extra pre-load (and add red Loctite as suspenders).