Frame wobble - custom built bike

On Sun, 27 Feb 2005 17:20:55 GMT, Jose Rizal wrote:

dianne_1234:

On Sat, 26 Feb 2005 04:36:11 GMT, Jose Rizal wrote:

Tom Sherman:


How about wrapping the tubes in the head tube/down tube/top tube are
with carbon fiber composite to add some stiffness to the frame?


It's not the stiffness that will change the frequency of shimmy onset,
but the change in mass of the system. You can hang a brick on the frame
and eliminate shimmy at the usual speed you encounter it. Not
practical, but neither is carbon fibre wrapping, for costs' sake.

Not sure I follow your logic, Jose. Isn't natural frequency a result
of both stiffness and mass (with some constants thrown in)?

f = 1/[2(pi)]*sqrt(k/m)

From http://personal.cityu.edu.hk/~bsapplec/natural.htm, Eq. 12

Yes,

Thanks. I think we agree stiffer structures have higher natural
frequency.

but to effectively change the stiffness only of the frame with
carbon "wrapping",

I'm not suggesting the gent wrap his new bike with carbon; my earlier
post on that was simply to confirm that doing so could makes the
substrate stiffer.

I would propose to the op that his frame builder could improve the
situation by making the frame stiffer. The builder will know well
enough how to do that: replace the top or down tube (or both) with
stiffer ones, or maybe switch to a different (stiffer) fork, or give
the op an entirely different (stiffer) frame, or some other structural
change, or change the steering geometry, etc.

the carbon matrix needs to be rigidly bonded to the
tubes.

Yep. Laminating carbon/epoxy onto an existing structure works great
with ordinary surface prep.

The loads placed on a frame are in the head tube ends, bb,
dropouts and seatpost.

Sure, we agree here.

With the exception of the latter, at all the
other points the load is placed on the tube metal, not the exterior of
the tubes where the carbon wrapping will be. Therefore the loads will
have to be transmitted from the tube metal to the carbon via a rigid
bond if the effective stiffness of the frame is to be significantly
changed.

Right, bonding the carbon to the tube during lamination is easy to do
and is the ordinary expected result of (properly) wrapping anything
with carbon/epoxy.

More likely it will be the change in mass of the system that
will change the resonant frequency, and not any additional "stiffness".

Both mass and stiffness play affect frequency.

On Sun, 27 Feb 2005 17:26:26 GMT, Jose Rizal wrote:

It's doubtful that a carbon fibre matrix will significantly change
stiffness of the frame by itself,

Right, the matrix (epoxy) by itself won't make the tube significantly
stiffer; it needs the properties of the carbon fibres to effect a
significant increase in stiffness over the existing tube.

On Sun, 27 Feb 2005 11:43:04 -0700, wrote:

Dear Dave, Dianne, and Ron,

Here are the results of searching the archives for shimmy,
frequency, and hz:

http://groups.google.co.uk/groups?q=...sa=G&scoring=d

Bob Mitke's post mentioned these shimmy frequency figures:

[begin Bob]

"I have seen data that identifies the modal frequencies for
frames, forks, and bars. A gentleman by the name of Dr.
Eric Gross wrote his dissertation about mountain bikes while
at the University of Hamburg. My german is horrible
(anybody have german translating friends??) but I did find
plots/tables that identify these natural frequencies to be:

Frame: 5 Hz to 110 hz (there were 7 modes within that range)
= therefore we can assume that road bikes are excited over
this frequency range since Jobst Brandt has identified high
speed shimmy is caused by vibrating frame tubes.

Fork: 7 Hz to 85 hz

Bars: 80 Hz up to 350 Hz depending on style and material

http://groups.google.co.uk/groups?se...&output=gplain

[end Bob]

Carl Fogel

Thanks Carl, great sleuthing!

I wonder if it's possible to find the conditions under which those
numbers are obtained? As Jose points out, changes in the mass of a
system have a large effect on frequency. Components measured alone may
not have the same frequency as when they are part of the bike+rider
system.

In Velonews during the last Tour there was an interview with a
composites engineer from Giant. He said something like the frame has a
fairly high natural frequency, but when mass increased by adding the
rider it dropped to around 10 Hz. I don't know if that might be near a
shimmy frequency though.

On Sun, 27 Feb 2005 21:11:10 GMT, Jose Rizal wrote:

Shimmy is most commonly experienced on bikes cruising on a straight
path. Therefore "lateral loads" on wheels are non-existent

.... until the shimmy gets started. Then lateral loads exist. They
oscillate as the shimmy reverses direction. But I don't know how large
the lateral loads get. Guessing, I'd say 5-10 kgf?

Jose Rizal wrote:
jim beam:


well, with respect, a couple of millimeters on bracing angle /does/ make
a difference. the smaller the bracing angle, the greater elasticity
will be experienced by the wheel. 1/sin theta & all that stuff.


You're making things up. Show your workings and "all that stuff".

no. you go use google. on a related note, bracing angle's effect on
tension ratio is also in damon rinard's spoke calculator if you care to
check it out.



harmonics, sure, that's complex, especially for two angled tubes, but
you can still make rough guesstimates based on tube torsions.


Tube torsions? How is this related? Show your guesstimates then.

google is your friend. stuff for single tube torsion is easy to find.



bigger
diameter tubes with a reciprocating torsional load have different
[higher] resonant frequencies to a smaller diameter tube of same wall
thickness.


It's the mass and shape of the system that's different, which is
affected by geometry. Now show how torsional load exists on a bike
frame in motion, and how this affects resonant frequencies at speed.


The shimmy described, being worse no-hands or one-handed, is what Jobst
has described in his FAQ on the subject.

that's the same faq that mistakenly claims gyro forces are relevant,
right? "bikes" without any wheels at all still shimmy.


At what speed did you experience shimmy with these bikes without wheels?


i'd guess somewhere between 5 & 50 knots - my friend had the speedo, i
was behind the boat trying not to drown so never thought to ask.

wrote:
On Sat, 26 Feb 2005 04:48:02 GMT, Jose Rizal wrote:


jim beam:


[snip]



"bikes" without any wheels at all still shimmy.

At what speed did you experience shimmy with these bikes without wheels?


Dear Jose,

Jim is probably referring to ski "bikes" again, vaguely
bicycle-like things with skis where the wheels would be:

http://groups.google.co.uk/groups?q=...eranews&rnum=1
or http://tinyurl.com/5glw9

I don't think that anyone has ever posted speeds or details
of shimmy on such contraptions.

I'm not even sure if the riders ever put their feet up on
these things, or just drag them on short little skis.

Jobst has also used these no-wheels ski "bikes" to
illustrate points, such as how steering works.

This would be a good time for some as-yet-unheard-from
poster to educate the rest of us about these
pseudo-bikes--how fast they go, how much they weigh, what
kind of frames they have, where the feet are placed, what
sort of shimmy (if any) occurs and whether it seems related
to the cold and body shivering, how much steering is lean
versus edging, and so on.

Carl Fogel

for examples, go to http://www.ski-bike.org/

dianne_1234 wrote:
On Sun, 27 Feb 2005 11:43:04 -0700, wrote:


Dear Dave, Dianne, and Ron,

Here are the results of searching the archives for shimmy,
frequency, and hz:

http://groups.google.co.uk/groups?q=...sa=G&scoring=d

Bob Mitke's post mentioned these shimmy frequency figures:

[begin Bob]

"I have seen data that identifies the modal frequencies for
frames, forks, and bars. A gentleman by the name of Dr.
Eric Gross wrote his dissertation about mountain bikes while
at the University of Hamburg. My german is horrible
(anybody have german translating friends??) but I did find
plots/tables that identify these natural frequencies to be:

Frame: 5 Hz to 110 hz (there were 7 modes within that range)
= therefore we can assume that road bikes are excited over
this frequency range since Jobst Brandt has identified high
speed shimmy is caused by vibrating frame tubes.

Fork: 7 Hz to 85 hz

Bars: 80 Hz up to 350 Hz depending on style and material

http://groups.google.co.uk/groups?se...&output=gplain

[end Bob]

Carl Fogel


Thanks Carl, great sleuthing!

I wonder if it's possible to find the conditions under which those
numbers are obtained? As Jose points out, changes in the mass of a
system have a large effect on frequency. Components measured alone may
not have the same frequency as when they are part of the bike+rider
system.

In Velonews during the last Tour there was an interview with a
composites engineer from Giant. He said something like the frame has a
fairly high natural frequency, but when mass increased by adding the
rider it dropped to around 10 Hz. I don't know if that might be near a
shimmy frequency though.

most definitely. people make first-shot estimates by orders of
magnitude on this stuff anyway, so being as it's definitely not 1 Hz, &
not 100Hz, 10Hz is order of magnitude right on the money.

On Sun, 27 Feb 2005 15:52:07 -0800, jim beam
wrote:

wrote:
On Sat, 26 Feb 2005 04:48:02 GMT, Jose Rizal wrote:


jim beam:


[snip]



"bikes" without any wheels at all still shimmy.

At what speed did you experience shimmy with these bikes without wheels?


Dear Jose,

Jim is probably referring to ski "bikes" again, vaguely
bicycle-like things with skis where the wheels would be:

http://groups.google.co.uk/groups?q=...eranews&rnum=1
or http://tinyurl.com/5glw9

I don't think that anyone has ever posted speeds or details
of shimmy on such contraptions.

I'm not even sure if the riders ever put their feet up on
these things, or just drag them on short little skis.

Jobst has also used these no-wheels ski "bikes" to
illustrate points, such as how steering works.

This would be a good time for some as-yet-unheard-from
poster to educate the rest of us about these
pseudo-bikes--how fast they go, how much they weigh, what
kind of frames they have, where the feet are placed, what
sort of shimmy (if any) occurs and whether it seems related
to the cold and body shivering, how much steering is lean
versus edging, and so on.

Carl Fogel

for examples, go to http://www.ski-bike.org/

Dear Jim,

Aha, they do use stubby little foot skis. (Wonder if that
affects shimmy?)

Sixteen to twenty-five pounds. (Oddly heavy, about the same
as a bicycle, despite lack of drive train, wheels, brakes,
and so forth. But since they are essentially coasting
downhill, maybe heavier is better?)

Up to 125 mph. (Gulp! I bet they worry about shimmy at that
speed!)

"Where can I buy one? Wal-Mart." (Is there a Fury
Snowmaster?)

"No, not yet." (Curses! They were teasing me.)

Carl Fogel

On Sun, 27 Feb 2005 14:45:53 -0800, jim beam
wrote:

[snip]

1. if you want a simple logical test for this, ask yourself why wheels
have two sets of spokes, not just one purely radial set.

[snip]

Dear Jim,

While opposing sets of spokes do brace the rim laterally,
aren't they also necessary just to true the rim?

Hmmm . . . How wide were the hubs on the big wheel of a
penny-farthing? That is, were they getting awfully close to
being effectively just one purely radial set?

Come to think of it, how would you squeeze the spokes on one
of those monsters? Did they tend to break spokes?

A glance through "King of the Road" finds a few pictures
from the right angle that suggest that the hubs were
scaled-up in width to match the height of the huge wheels.

The double-page photograph of an 1885 "ordinary" on page 93
shows a 72 spoke main wheel.

Carl Fogel

On Sun, 27 Feb 2005 14:45:53 -0800, jim beam wrote:

sorry, jose, when i use "stiff" wheels, they shimmy less. hence
elasticity /is/ relevant.

Take the tires off of the "stiff" wheels, and the non-stiff ones, and
measure the deflection, particularly vertically. You won't find any
significant difference.

bracing angle /is/ a component in the
stiffness equation. do a google search for why you can never have a
perfectly horizontal wire drawn between two posts.

Which has nothing to do with this.

then figure out why
structures like the golden gate bridge are built with "slack" rather
than "tight" suspension wire. 1/sin theta.

It's not, by the way, 1/sin of anything. The curve is a catenoid, as
described in
http://server1.fandm.edu/departments.../solution.html

Not, perhaps, the best book on the subject, but it gives the gist of the
problem.

--

David L. Johnson

__o | Accept risk. Accept responsibility. Put a lawyer out of
_`\(,_ | business.
(_)/ (_) |