| Author |
Message |
    
Steveshakeshaft
| | Posted on Wednesday, June 15, 2005 - 07:01 am: | 
|
Rocket the "easiest way" to get a high strength bolt is to use a Allen Cap Head Bolt (the black ones) they are (from memory) Gr 12.9 as standard.
However..... as with many things, that level of tensile stength may not be all that is needed in this application. A high degree of "toughness" is just as important as Tensile Strength on an application like this. And high tensile strength may not go hand in hand with toughness.
Furthermore, the (smallish) number of failures that I've seen have been Fatigue failures. Fatigue happens when a crack grows with stress cycles until the crack grows to the stage where the remaining (uncracked) metal can no longer take the load. Fatigue is a cyclical event not a steady state event and signifys that there is a flexing or other load/none load/load cycle happening. Whether that is flexing of the mount I don't know. Secondly, usually in a threaded fastener the loss of bolt tension in a bolt cyclically loaded is the prime suspect when looking for the root cause of a fatigue failure. It is vitally important therefore that the hardened washers are used and that the correct procure is used.
It is unlikely that a properly tensioned bolt that has enough "stretched length" will loosen through vibration. Hope that helps. I apologise in advance for not having the eloquence of Blake when I try to explain these things that I've studied for many years of investigating machinery failures.
My last word, I've got the NRHS mount and '04 Sportster heads on my '97 M2. And I'm much happier.
Regards
Steve. |
Al_lighton
| | Posted on Wednesday, June 15, 2005 - 08:55 am: |
|
Rocket,
I don't believe that the bolt is too short at all, though I would agree that the threaded hole is deeper than it should be based on how far the bolt needs to go in, and this weakens the head boss substantially. Based on where it fractures each time, I would conclude that the bolt length reach into the head is NOT the issue at all. There is at least 1.5 bolt diameters of thread engagement, any further does not add strength to the bolted interface. I don't believe that the head boss will break unless one of the bolts has already broken, or there is a latent defect in the head casting, even at the depth they have drilled/tapped it.
HOWEVER, the design IS suspect. A superior bolted joint does NOT put the shear interface across the threads, and the Buell design does. That is a significant mistake, one no structural engineer should ever make in a critical load application. Putting the threads at the shear interface results in very high stress concentrations in the bolt, rendering that Grade 9 bolt worth much less. Fatigue cracking in such a case is not unexpected, if there is substantial cyclical loading.
I have no doubt that the optimum solution is a longer unthreaded bolt length, coupled with a precision diameter/location counterbore on the head bosses, so that the full bolt diameter crosses the shear interface. With the head off the bike, it would be possible to implement this fix. But there is no way that one could implement this in situ, the requisite precision can not be obtained by hand/eye operated tools. The c-bore diameter has to be only a few thousandths larger than the bolt, and therefore the location of the two cbores relative to each other must be very precise. So the next best thing is reduce some of the flexing and oscillation amplitudes at the joint, and I believe the NRHS mount provides some of that. Is it enough? Or will my bike break a third set? Only time will tell. I think it will be, as the joint can't be grossly underdesigned just based on the number of miles that many a tube frame bike has with NO bolt failure whatsoever.
That's my theory and I'm sticking to it 
Al |
Doncasto
| | Posted on Wednesday, June 15, 2005 - 10:11 am: |
|
FYI
Link to my Bolt Failure Report
Don Casto
index page for "Buellstuff" |
Blake
| | Posted on Wednesday, June 15, 2005 - 10:22 pm: |
|
What Steve said.
Al,
You are right that threads in the shear plane degrade the shear loading capacity/strength of a bolt. But is perfectly acceptable and very much routine to have threads in the shear plane. In such cases, the loading must simply be kept below prescribed limits. Bolts have allowable loads for both cases, threads in and out of shear plane, also for single and double shear conditions. The threaded section is significantly weaker as you point out. Any idea of the single shear capacity of a 7/16" grade 9 bolt with threads in the shear plane? How about the design load for the engine mount? Interesting stuff.
Rocket,
When you reassemble the engine mount in question, ensure that the faying (mating) surfaces of cylinder head, engine mounting bracket are clean and bone dry and roughed up just a bit with some medium grit sand paper. And as Al points out following, be sure the bolt head and washer are clean and lightly lubricated to facilitate optimal torqueing/preloading. Ideally, that connection would be able to carry shear loading solely through friction action between faying surfaces, thus imparting no shear into the bolt and thus no flexure and no fatigue worries.
(Edited to correct mis-stated info. Thanks Al. )
(Message edited by blake on June 16, 2005) |
Al_lighton
| | Posted on Thursday, June 16, 2005 - 02:03 am: |
|
Well, OK, Blake, tis true that you can do it if the bolt is sized for the load. Surely a 1" diameter bolt with threads in the shear plane is stronger than a 1/4" bolt with a non threaded shear plane. And you're right, it is done all the time, as there are often other constraints that dictate that it must be done. But it isn't a superior joint design, and it is best avoided where practical, and in this case, it seemed quite practical to not do it, the geometry easily supports the superior design. The nature of the failures on this particular joint (almost always a fatigue fracture of the left hand bolt, within one thread of the shear plane) sure point to the bolt NOT being sized right for the loading it is seeing. I sure don't have the test data and load calculations to back anything up, I'm just armchair quarterbacking.
Interesting about the sand paper thing you're saying....Roughing the faying surfaces sure makes sense, but the bolt head and washers? I'm not so sure that makes sense for this application. The Buell service bulletin has you put loctite on the threads, and oil under the head of the bolt and washer. This is surely done so that the "torque to 60 foot-lb, loosen one turn, re-torque to 60 foot-lb" instruction yields a consistent bolt stretch. If you roughed up the bolt and washer, wouldn't you affect that torque to stretch relation? That seems bad.
Al |
Blake
| | Posted on Thursday, June 16, 2005 - 02:48 am: |
|
I misspoke egregiously improperly if I gave the impression that one should roughen up anything other than the two faying surfaces of the head and bracket. All surfaces clean, only the faying surfaces roughed up. Me no write good. Good call. |
Steveshakeshaft
| | Posted on Thursday, June 16, 2005 - 07:54 am: |
|
I refer the honourable gentlemen to what I consider the standard reference works on this (to me) very interesting topic. The topic is way more complex than many might imagine. But what I've learned from Bickford has without fail stood me in good stead in the world of industrial machinery failures investigations over many years.
Ref- "An Introduction to the Behaviour of Bolted Joints", JH Bickford, Marcel Decker
Inc.1981 ISBN 0-8247-1508-X
A search in Google finds it easily. Apologies upfront if I'm too far off topic.
Regards
Steve. |
Djkaplan
| | Posted on Thursday, June 16, 2005 - 08:08 am: |
|
I think Buellzebub's comments should be given more consideration.
"I showed the pieces to our resident mechanical engineer to get an expert opinion.
The failure looked very much like a standard fatigue failure. We came to the conclusion that thermal cycling over time may have contributed to the weakening of the thread locking compound and subsequently allowed the fastener to lose its clamping torque. Once the tensile load was removed the mechanism for the fatigue failure was in place and it was only a matter of time before the bolt would fail."
In all the posts in the Knowledge Vault about mount failure, I don't thing anyone ever considered this. It makes sense the left side bolt would fail first before the head breaks on the right.
When I install the Nallin mount, I'll take special care to mark the fasteners (just like the front rotor bolts are marked with yellow paint) so I can visually inspect them for loosening. |
Blake
| | Posted on Thursday, June 16, 2005 - 04:07 pm: |
|
Excellent suggestion. I had missed the B-bub's comments. I think he may well be onto the root cause of the failures. |
Bob_thompson
| | Posted on Thursday, June 16, 2005 - 10:56 pm: |
|
Gentlemen, I have a question concerning the vibration inherent in V-twins and bolt choice. Years ago as a structural ironworker and helping to build a few railroad bridges we used only grade 5 bolts or rivets because as I learned the stronger grade bolts were more prone to work hardening from vibration from the trains going over the bridge. Besides the mount bolts, going into aluminum, it seems, would not allow for sufficient torque to utilize the full potential of stronger grade bolts , such as grade 8. I understand that the reason for stronger grade bolts is to allow for more torque in tightening, therefore increasing shear strength. Am I going wrong here and if so where. Just maybe a grade 5 would be completely sufficient in strength and be less susceptible to vibration caused work hardening. One other thing is that I have safety wired the stock bolts to insure no loosning. Bob
(Message edited by bob_thompson on June 16, 2005) |
Rocketman
| | Posted on Friday, June 17, 2005 - 12:26 am: |
|
Steve, thanks for the input. I looked into using the Allen heads you suggested but the thread length is longer thus placing a partial amount of thread inside the holes in the mounting bracket.
There is no reference what so ever in the workshop manual as to what these bolts should be torqued to nor the correct procedure required. I missed the service bulletin. Yeah it's been a long time. I tightened them by hand and they feel good. Clean as a nuns nickers and Locktited to the max. No I am not going to worry. I am happy with the repair and modification though I may just run a torque wrench across the bolts to see how much muscle strength I have.
Just to put the cat among the pigeons...........................
I believe the head boss \ right side bolt cracked \ broke first. Why? Because the left side bolt scratched the rocker cover when it broke. If it had broke first the bolt would have fell away from where it lived whilst the right side bolt still offered support. Therefore it is not possible the left bolt snapping first could have scratched the rocker cover when the motor dropped. Unless of course gentlemen you know different?
Rocket
scratching head???? no pun intended LOL |
Steveshakeshaft
| | Posted on Friday, June 17, 2005 - 04:21 am: |
|
A point in Buellzebub's comments doesn't ring true to me though. (Mechanical Engineers never agree anyhow, that's nothing new!). He states-
"thermal cycling over time may have contributed to the weakening of the thread locking compound and subsequently allowed the fastener to lose its clamping torque"
Quite possible that the Loctite could degrade over time. However, an Engineer will never design a joint that relies on this element alone for its integrity. Too many variables involved that could go wrong. The Loctite is more a "belt and braces" approach. A properly designed bolted joint will have enough "stretched length" when the bolt is properly tensioned such that it will not loosen under cyclical loading. Now I'm not going to pull my motor apart to start measuring, but I would consider from memory that the fasteners specified on this joint have adequate "stretched length". Therefore, if torqued properly using Buell's procedure, they should not lose tension (Torquing is itself quite controversial but that's another story). It then follows that provided the degree of stretch given in the bolt is more than the amplitude of the stress cycle then the bolt will have adequate fatigue life. IE the load/unload/load stress cycle occurs entirely within the elasticity of the stretched bolt and does not allow the joint to undergo load/unload conditions. (Sorry if that doesn't read too well, Blake may help out here.)
If my logic holds true, then if loss of bolt tension isn't the root cause of the failure (we cannot be 100% sure on ALL failures here) then another mechanism is the root on at least some of the failures. There may well be different types of failures occuring with similar end results. Flexing of the mount being just one possibility. Poor castings and failure of the mount is another.
Steve.
(Message edited by steveshakeshaft on June 17, 2005) |
|
