John, I learned a good amount from your explanation my first time through it, and I
will read it again.  Thank you.

"BSFC"?

By the way I have been lurking here for 10 days or so.  I have had an '82 Rabbit
diesel pickup since '90, and have had, but no longer own, a '71 and '72 bus.  My
brother just rebuilt the engine in his '83 diesel Vanagon, and I forward items to
him that he might find interesting.
-Dana-

>
> Oh fun, another chance for me to fail in communicating a concept.  Out of 700+
> List members, my pleas for the merits of gearing steady state highway cruising
> to match the engine's lowest BSFC rpm range has not generated EVEN ONE voice
> in agreement.  But despite that silence, I'm still confident that I'm right,
> and to fail to follow that design path is comparatively inefficient.  Let's
> see if I can fail to communicate so poorly again.
>
> This time I'll describe the simple 90 degree link Universal Joint, often
> simply called a U joint.  They are NOT CV joints, and to understand why CV
> joints were made, you need to understand how these tough, old fashioned,
> reliable flex couplings fail to maintain constant angular velocity agreement
> between the driving and the driven shaft.
>
> To see how the angular rotation velocity of a driven shaft is affected by a
> "universal joint," it is very helpful to exaggerate these transmission
> irregularities by examining extreme angles of transmission.  So gain access to
> a U joint that you can hold in your hands if you don't have an unusually good
> dynamic 3D imagination.  An old unit, temporarily saved from a mechanic's
> trash barrel, which has been degreased would be a great help for instruction.
> But a socket set U joint will do if you must stay clean. <G>
>
> Now, define one end as the driving end, and the other as the driven end.
> Misalign the driving and the driven rotation centers so they are pointing
> widely apart.  This forces the center coupling to flex through an extreme
> angle range as the two ends are rotated,.  Next, imagine affixing an analog
> clock's minute hand to each U joint end, with a clock face for each.  At the
> beginning of our demonstration, align both minute hands parallel with each
> other and with one of the coupling link's two pivoting axis directions.  Then
> spin the clock faces so zero seconds on each aligns with its minute hand.
>
> Analog clock time seconds always correspond with 6 degrees of rotation.  There
> are 360 degrees in a circle, but only 60 seconds per minute, so 360 / 60 = 6
> degrees per second.  So any analog face clock can serve as your visual
> protractor which you can mentally apply anywhere you go.  This enables you to
> estimate observed angles in degrees quite well.  Teach your kids how to
> mentally apply this and it will serve them well through their whole life.  I
> helped a friend's six year old girl to become able to look at angles and
> describe them within 6 degrees after just a few days of "calibrating her eye
> ball."  Work on it.  If you're caught staring off into space, you can always
> explain that you're calibrating your eye balls.
>
> Now we know that when the driving shaft second hand is pointing at 0 seconds,
> the driven shaft will be pointing at 0 seconds, because we just set them that
> way.  I can tell you that at an indicated 30 seconds, they will both be
> aligned again.  But between these two points of rotation, they will not
> agree.  I just got done writing out a word description of the angular velocity
> changes of the driven shaft compared to the driving shaft, broken out by 1/8
> turns of rotation for the first 1/2 turn (30 seconds or 180 degrees).  There
> was no need to repeat that explanation, because it is exactly repeated in the
> second half turn.  But I deleted it because it got too messy for an audience
> this wide.  I suspect few who do not already understand this would track it
> well.  The main thing to remember is that when periodicity of rotation is
> transferred through a simple 90 degree link U joint, the driving and the
> driven shaft do not have the same velocity most of the time, even though after
> any total number of complete rotations, they have rotated the same number of
> total degrees..  It's just that they took different paths in terms of
> rotational velocities.
>
> Matched U joints pairs can compensate for each other:
> But this single 90 degree link U joint effect can be revered by correctly
> aligning a second U joint so it does the same thing, BUT IN THE OPPOSITE
> DIRECTION.  I have seen some incorrectly assembled drive systems where these
> effects were not mutually canceling.  I remember a shortened drive shaft for a
> project car (blush) which got welded back together (don't ask) with the yoke
> spun 90 degrees from how it should have been.  On plenty of vehicles, just
> sliding them in the splined shaft coupling out of sync by 90 degrees will do
> the same thing. Rather than canceling each other, the U joint angular velocity
> swings were additive!   Talk about a nasty shaking dog putting pulses through
> the drive system when ever it would flex!  If you look at the drive systems in
> some old carnival rides, a slower version of this same effect is sometimes
> intentionally employed to cause riders to experience  continuous cycles of
> acceleration and deceleration.  But that's another story.
>
> If there is any flexing done by correctly assembled 90 degree link U joint
> link drive system pairs, the center coupling shaft will accelerate and
> decelerate many times per second as it goes flexing down the road, but the
> input and output shaft will always be in angular velocity agreement.  But you
> MUST align those U joints to correctly to cancel each other's non-CV effects.
> If you align them so these effects are additive, the inertia loads will hammer
> the drive train when they flex.  Even correctly assembled, the connecting
> link's inertia induces some minor pulsing loads, but they don't cause much
> trouble. (That last sentence clause may cause some disagreement, as this is a
> controversial judgment call.)  Remember, when there is no drive system
> flexing, and in a well designed drive system, that's true most of the time, U
> joints cause no non-CV effects.  If a drive system is designed to require
> flexing with every rotation, some guy with big black rubber ears named Mickey
> was an influential member of the design team.  Just my opinion.
>
> Having described the effects of the simple, comparatively reliable 90 degree
> link U joint design, I will leave it to others description of how nicely CV
> joints handle all this rotation velocity agreement.  As far as I'm concerned,
> that's the only thing they do well in real applications.  I'll also let them
> explain why they are not nearly as reliable.  CV joints are, on paper, more
> elegant than the simple 90 degree link U joint system I just described.  But
> your vehicle is not just a paper drawing.  You won't find CV joints in drive
> systems where maximum service life is the top design goal.  If they were as
> reliable, they wouldn't be a constant topic of maintenance discussion.
>
> Has anyone on the List but me ever owned a 600 cc BMW Isetta car?  They had
> rubber rings with four holes cut through them as their flexing drive
> coupling.  Two across would attach to each drive or driven shaft.  I have a
> book showing the mechanics of early cars.  In one they had a paddle wheel with
> center mounted blades driving rubber blocks captured inside a bowl with blades
> coming in from the outside.  The more rubber blocks, the closer it would be to
> a CV joint. Even a CV joint doesn't have an infinite number of loaded ball
> bearings, so it's not quite CV is it?.
> John
>
>  Maybe we should try cement tires and rubber roads.