cymtriks
|
| posted on 17/7/03 at 10:05 PM |
|
|
Deon Analysis
The picture of my analysis is in the photo section. The picture shows the stress due to acornering load on the tyre contact patch.
This analysis supports the table of tube stresses that I posted earlier. This is the result of a finite element model of a design similar to a
Caterham Deon axle.
The tube stress table is the result of basic beam analysis (secant formula).
The cornering load on the tyres is taken as 3 times the corner weight assuming equal weight distribution. The total car weight is given in the left
column and includes everything: driver, fuel, luggage, etc. This does give highish stresses but they are still useful to compare different tube
sizes.
[Edited on 17/7/03 by cymtriks]
[Edited on 17/7/03 by cymtriks]
|
|
|
|
|
craig1410
|
| posted on 19/7/03 at 07:33 PM |
|
|
Cymtriks,
I've been thinking about this and although I'm not sure that I understand your stress picture in the photo archive I did have this
thought:
Under hard cornering the outer wheel will experience both a lateral load on the wheel/tyre at the contact patch and a vertical load due to the weight
of the car added to the weight transfer you will get under cornering. In terms of the load on the de-dion axle, won't these two forces cancel
out to a degree, leaving a single force vector of diminished magnitude?
To explain a different way, the lateral cornering force will cause the bottom of the wheel to be forced inboard as if trying to attain more positive
camber but the vertical component of mass and weight transfer will cause the opposite bending load on the de-dion axle as if trying to attain more
negative camber.
Now depending on the distance between the center of the contact patch and the point where the "ears" attach to the de-dion tube it should
be possible to "tune" the axle design to effectively cancel out these force components at the expected maximum lateral G load. The bigger
the distance between the contact patch and the ear-axle joint, the more lateral G will be required to overcome the vertical mass induced force. I
would expect(intuitively only!) for road use with road tyres the best compromise will be a distance between the ear-axle joint and the contact patch
roughly equal to the rolling radius of the tyre. That way a 1G vertical component and a 1G lateral component should (I think) produce a near zero
bending force on the axle (but this is all just in my head at present as I have not drawn out the force vectors or anything yet.)
When at rest the only force on the de-dion tube will be caused by the pure vertical weight and under this condition a large distance between the
contact patch and "ear" would cause a large bending moment on the axle. This force would be reduced during cornering on the outer wheel
but increased on the inner wheel but since the weight will transfer away from the inner wheel it shouldn't be a problem.
Does this make any sense? If not then I'll try to draw up a sketch to clarify what I'm on about.
Hopefully this idea may explain why your calculations seem to disagree with empirical evidence.
Cheers,
Craig.
|
|
|
craig1410
|
| posted on 20/7/03 at 04:33 PM |
|
|
Has nobody else got anything to say on the subject?
Very unusual..... 
Craig.
|
|
|
blueshift
|
| posted on 20/7/03 at 08:50 PM |
|
|
I hear you craig. makes sense to me.
|
|
|
craig1410
|
| posted on 20/7/03 at 10:00 PM |
|
|
Thanks Blueshift, maybe you can explain it to me sometime... 
Seriously though I think this may well explain the apparent descrepancy between theory and practise. I'll be interested to see what Cymtriks has
to say. He should be able to modify his FEA model to take account of the opposing forces I described. Hopefully it will then show that the stress on
the de-dion tube is much lower than originally thought and I'll be able to sleep better at night again... 
Cheers,
Craig.
|
|
|
blueshift
|
| posted on 20/7/03 at 10:45 PM |
|
|
I don't know if they're exactly opposing.. but the forces would seem to be trying to deform the bar into an S-shape rather than a simple
bend. if you see what I mean. I've never done any engineering so I couldn't say if the bar could withstand this any easier. Think
I'll shut up and let those with the clues debate it
|
|
|
craig1410
|
| posted on 21/7/03 at 12:09 AM |
|
|
Blueshift,
I think I know what you mean but at each component interface there will tend to be a combination of tension, compression and bending forces.
Essentially bending is just a combination of tension and compression in itself. My feeling is that Cymtriks has created a model which is too simple as
I believe that the force interactions in a de-dion axle are much more complex and I think that you need to take account of all the components (eg.
axle, ears, hub carrier stubs etc) to properly predict the forces at work. I'm sure that a simplified model can be developed which will
represent "most" cases but I think the model is too simple at present.
I did electronic engineering at Glasgow Uni between 1990 and 1995 so I am not 100% up to speed with mech eng but many of the classes were common to
both elec and mech engineering and with the application of some common sense and a bit of guesswork I can usually get by.
Cheers,
Craig.
|
|
|
Rorty
|
| posted on 21/7/03 at 02:54 AM |
|
|
Sorry Craig, I was too lazy to scroll across and read your post, and so, passed it by. My brain hurts. I think you're looking far too deeply
into a relatively simple device, of which there are innumerable working examples to draw on.
One more coal on the fire though; how about building some neg camber into the de Dion axle?
[Edited on 21/7/03 by Rorty]
Cheers, Rorty.
"Faster than a speeding Pullet".
PLEASE DON'T U2U ME IF YOU WANT A QUICK RESPONSE. TRY EMAILING ME INSTEAD!
|
|
|
Noodle
|
| posted on 21/7/03 at 05:59 AM |
|
|
quote:
How about building some neg camber into the de Dion axle?
Spendid idea. Independant tuners used to offer various camber-configured tubes for the Alfa 75/Milano in the days before <cynicism>all
Alfa's were made out of Fiat Punto's or somesuch</cynicism>
Different driving conditions track/road/Grandma to hairdressers can be setup with pre-configured accuracy.
Cheers,
Neil.
|
|
|
