[dallas]

Has anyone figured out why the coil spring fails in the first place?

If you’re asking about the doohickey spring it’s because the design is wrong. I’m not sure I can explain it well but the sloppy hole/shaft fit allows the tensioner arm that one end of the spring attaches to constantly jerk on it. Even though the original adjuster plate (doohickey) is clamped tight with a bolt the shaft with the tensioner arm attached to one end of the spring can pivot back and forth a lot. Using the torsion spring bypasses the tensioner arm and uses the tighter fitting Eagle Mike’s doohickey to rotate the shaft then lock it in place.
Please someone help clarify this...

 

[a4twenty]

Has anyone figured out why the coil spring fails in the first place?

the OEM spring is just too long from the factory, so it doesn't have any / enough tension and can't pull the lever back during adjustment.

when they beefed up the doohickey for second gen they just used the same spring. so now the doohickey doesn't grenade but still won't tension right.

the easiest solution for kawi is a shorter spring.

slightly tougher solution is to change the machining / punching of the doohickey mounting hole by rotating the locating flats so the spring applies the right tension.

my guess would be IF they do anything it would be the easiest / simplest but only time will tell.

 

[Tom Schmitz]

Has anyone figured out why the coil spring fails in the first place?

Cyclic motion. The balancer lever and spring lever both fit loosely on the shaft. If the engine is running at 4000rpm the spring sees 4000 cycles per minute. That's a quarter million cycles in an hour. Figuring that springs fail at 15,000 miles and stipulating that the 15,000 miles at an average RPM of 2500 rpm and 30 miles per hour, the spring fails at 75 million cycles. Completely made up numbers, but the point is illustrated. The movement is minute, of course, but it is happening. Failure often occurs at the sharp 90* bend that forms the hook.

The spring is also rubbing on the case, but that seems to simply burnish the case.

If the spring sat completely still, as the aftermarket torsion spring does, I doubt it would fail.

 

[Johnlharris]

I think the reason they have been quiet about the Doohickey is that it would admit to a design flaw. Manufacturers use the phrase “update”or revised. The one piece sub frame to frame update was said to improve the total frame strength not because it might be a week point bolting it together. Could you imagine Ford saying we “fixed” the gas tank so it won’t explode anymore. I like the new bike. After all it’s still a KLR. You can’t beat fuel injection. I live on Topsail Island NC and in 3 hours I can ride from sea level to 3500 feet to the NC mountains. There a plenty of other bikes to choose if you want a twin or lighter weight and apparently a 6 gear. Ive had bikes with 4,5 and 6 gears, tack and no tack, temp and no temp and still was able to enjoy my ride. I also have a 13 F6B but find myself riding my 09 KLR most of the time. Ride on 👍

 

[Damocles]

I think the reason they have been quiet about the Doohickey is that it would admit to a design flaw.

Maybe so, but . . . Kawasaki was NOT quiet about the larger-diameter front brake disk. Or, the thicker rear brake disk. Or, as you mention, the welded subframe.

Do not these updates/revisions suggest, "design flaws?"

Maybe not, but somehow, a revised/updated doohickey opens the gates to product liability.

 

[Damocles]

Cyclic motion. The balancer lever and spring lever both fit loosely on the shaft. If the engine is running at 4000rpm the spring sees 4000 cycles per minute. That's a quarter million cycles in an hour.

Trying to get my head around this mechanical description!

Seems to me, if the doohickey (eccentric shaft lever) is bolted down, the doohickey spring "sees" only the minor escalations of vibration. That is to say, once the doohickey is cinched down, the balancer chain tension slack does not impinge on the doohickey spring, because the ends of the spring are anchored, and no tension is exerted on the eccentric shaft by the spring--the hold-down bolt (not the spring) maintains balancer chain tension.

Hard to put into words, but . . . doohickey spring tension appears to me to be at issue ONLY when adjusting balancer chain tension; operationally, balancer chain tension is maintained by the snubbed-down doohickey, NOT the spring (whose elasticity plays a part ONLY during balancer chain tension adjustment.

Please disabuse me of this heretical and incorrect analysis/concept!

 

[Tom Schmitz]

Did you watch the video?

 

[pdwestman]

NOT the spring (whose elasticity plays a part ONLY during balancer chain tension adjustment.

The factory style spring system (extension spring & separate lever arm) (and early EM kits) are always in motion, be it ever so slightly after a proper adjustment (you've read my procedure), or excessively if never allowed to re-adjust or by the factory leaving the locking bolts loose (I've found too many, I think that Tom Schmitz '09 bike originally had a loose locking bolt, because of wear pattern in the paint behind the doo!).

The torsion spring system developed by EM is the only style that 'Locks' the spring out of any motion, unless people leave the locking bolt loose.

Damocles, have you watched Toms video by now?

 

[Norton 850]

Cyclic motion.

I think what you are describing is cyclic fatigue.

However, this is an unusual case because the spring appears to be very lightly stressed. In fact it appears to be stressed way below the endurance limit of the spring material.

Parts that are subjected to cyclic stresses should be able to do so indefinitely, so long as the stress levels are less than the endurance limit (0.5 of ultimate tensile) of the material. I think perhaps the spring failed owing more to an improper manufacturing technique rather than cyclic fatigue.

Jason

 

[Tom Schmitz]

But you're not saying that the very lightly stressed spring would, sitting under such a static load, fail after some time, are you? That is, lightly stress the spring by extending it between two fixed posts. Observe it and, at some point, it will simply fail.

 

[Norton 850]

lightly stress the spring by extending it between two fixed posts. Observe it and, at some point, it will simply fail.

It shouldn't fail regardless of the number of stress cycles so long as the stress is below the endurance limit of the material.

Jason

P.S.

Don't take my word for it; see attached from Mechanical Engineering Design by Shigley.

 

[dallas]

Trying to get my head around this mechanical description!

Seems to me, if the doohickey (eccentric shaft lever) is bolted down, the doohickey spring "sees" only the minor escalations of vibration. That is to say, once the doohickey is cinched down, the balancer chain tension slack does not impinge on the doohickey spring, because the ends of the spring are anchored, and no tension is exerted on the eccentric shaft by the spring--the hold-down bolt (not the spring) maintains balancer chain tension.

Hard to put into words, but . . . doohickey spring tension appears to me to be at issue ONLY when adjusting balancer chain tension; operationally, balancer chain tension is maintained by the snubbed-down doohickey, NOT the spring (whose elasticity plays a part ONLY during balancer chain tension adjustment.

Please disabuse me of this heretical and incorrect analysis/concept!

Respectfully, It seems you’re missing the fact that the pin is never locked in place because it fits so loosely in the ‘locking’ plate.

 

[Damocles]

Respectfully, It seems you’re missing the fact that the pin is never locked in place because it fits so loosely in the ‘locking’ plate.

"Pin?"

I'm trying sincerely to understand the cycling (excursion distance and cyclic rate) of the OEM doohickey spring; guess I have no valid view of the geometry involved.

TomSchmitz and pdwestman: I look forward to viewing the doohickey video!

 

[Norton 850]

"Pin?"

I'm trying sincerely to understand the cycling (excursion distance and cyclic rate) of the OEM doohickey spring; guess I have no valid view of the geometry involved.

TomSchmitz and pdwestman: I look forward to viewing the doohickey video!

I think dallas means "lever". The lever fits loosely and is free to move about even when the doohickey adjustment bolt is locked down. And since one end of the OEM spring is attached to the lever it, too, is free to move. As Tom pointed out, there is high frequency low amplitude movement of this lever and spring when the engine is running.

But as Paul mentioned, the EM doohickey and spring arrangement prevent movement of the spring because it is attached to the doohickey rather than the lever, hence a much better design.

Jason

 

[Tom Schmitz]

It shouldn't fail regardless of the number of stress cycles so long as the stress is below the endurance limit of the material.

Jason

P.S.

Don't take my word for it; see attached from Mechanical Engineering Design by Shigley. View attachment 28579

Yes to all that. I'm trying to pin down what you're saying the failure mode is. Trotting out Shigley does not answer the question as posed.

Extrapolating, you seem to be saying that it either simply fails from sitting there or that it does not fail and the failure is simply imagined. You do seem to be saying that cyclic motion, leading to cyclic fatigue, is not a factor.

What I have said is that the spring fails because it is subject to cyclic motion. I would maintain that if it were not subject to that, it wouldn't fail, even if there were a manufacturing defect (e.g., crack or stress riser) or design defect (too sharp a bend radius, under-tempered). I am not saying that the cyclic motion causes it to exceed an endurance limit. My original answer was not meant to be a RCA, it was simply an answer to the original question as posed.

But again, in the absence of motion the spring won't fail because, as is obvious, the spring is very lightly loaded.

 

[Norton 850]

I'm trying to pin down what you're saying the failure mode is. Trotting out Shigley does not answer the question as posed.

However, this is an unusual case because the spring appears to be very lightly stressed.

I think perhaps the spring failed owing more to an improper manufacturing technique rather than cyclic fatigue.

Jason

 

[Tom Schmitz]

Ah! I get it. You're being deliberately abstruse.

 

[PeteK]

My (not so) innocent question smoked out a pretty lively discussion!

Tom, thanks for the video link. I think I also saw it on your Souperdoo pages. And it’s a reasonable hypothesis. But I don’t think it’s the primary cause.

As Jason pointed out (and I appreciate his reference to Shigley, with whom I have more than a passing familiarity), on steels and certain other materials, there is a point in the stress/stain curve, called the fatigue limit, or endurance limit, below which that material will NEVER fail in fatigue, and it’s generally about .5 of ultimate tensile strength for steel. That’s why valve springs in diesel truck engines can go a million miles or more. Aluminum, BTW, does not have a fatigue limit. Even a very low stress, if repeated enough cycles, will eventually lead to cracking and fatigue failure. That’s why springs are not made of aluminum.

The question Jason referred to, but didn’t quite state clearly, is whether the minute stretch/relax motion of the spring, as the chain pushes against the chain tensioner and lever, stretches it enough to exceed the fatigue limit. Just looking at the spring and the minute amount of motion, I seriously doubt direct cyclic stress is the proximal cause.

Instead I suspect a different failure mode that I have observed in other applications of springs, and with which vehicle engineers are familiar.

In a word: “resonance.”

In particular, spring resonance is a major concern in the design of valve springs. Some of you may be familiar with “beehive” valve springs (coils are tapered to a small outside diameter at the top) and double valve springs. Engineers came up with these designs to quell spring resonance in racing engines, which operate at higher RPM, and whose engines produce more vibration which also gets transferred to, and excites resonances in, the valve springs. This resonance was sufficient to exceed the fatigue limit of the steel, and was known by engineers from the early 20th Century (IIRC, Sir Harry Ricardo noted this phenomenon on his 3-part treatise on internal combustion engines in the late 1920’s), but it was observed directly with high-speed photography in the 1940’s or 1950’s.

Back to the doohickey coil spring. I suspect that the vibrations within the engine, likely with contribution from the slight movement of the tensioning lever, excite a resonance in the spring, leading to its failure. So it does fail from cyclic stress, but not directly from the movement of the doohickey lever. The EM torsion spring, a completely different configuration, evidently doesn’t have this problem.

Anyway, I submit this hypothesis for your consideration.

 

[Tom Schmitz]

Yes, I agree. Especially about the lively part. ;^)

I am familiar with the resonance issues in valve springs, et al, but was not familiar with Sir Harry Ricardo. I am going to have to look that up. I thank you for that lead.

You asked a simple question and I gave an off-hand and simple answer. It wasn't a technical answer, it wasn't a root cause analysis, but it was taken to task as if it were. Perhaps it was because I used the term "cyclic motion" some deep analysis was inferred. I guess I should have said "cuz there be a whole lotta shakin' goin' on and the spring is moving and stuff".

For most people, Shigley's works, fatigue/endurance limit, tensile strengths, and such are passingly interesting, but understanding that the design is one that allows the spring to be in constant, albeit very small amplitude, motion provides a greater and intuitive understanding of the breakage. It also provides insight as to why the torsion spring is a better design.

I once asked a doctor about the leading causes of death. He laughed and said that there was really only one cause. "Their heart stops beating; that's the most immediate cause. If that didn't happen, they wouldn't be at room temperature!"

 

[PeteK]

Your doctor is a wise and practical man.

One more explanation of spring resonance (amd this discussion applies in principle to any other resonance, including wheel wobbles, high-speed weaves, electrical circuits, and software code loops): under a high-speed camera, you can see the coils of the spring slamming back and forth between the ends of the spring, like a Slinky, even though the ends don’t move very much. It’s my guess that is what causes the spring to break at the 90-degree bend to the hook.

Engineers also improved on simple hooked-end coil spring designs a century ago with what is called an “improved end” (you can also reference this in Shigley). In this design, the coils reduce in outer diameter for a few turns before the hooked end of the spring, much like the tops of Beehive valve springs. This reduces the concentration of stress at the bend to the hook. Making the bend to the hook more gradual also helps reduce stress concentration.

I’ll bet that if Kawasaki used that design, we wouldn’t even be having this discussion!

And with that, I’ve probably “beat” this topic to “death.”