How is the engine/gearbox assembly mounted to the tub?

[ringo]

It is a good example becuase those parts are under designed. I am sure most factory manifolds do not fail after decades of service.

Another perfect example is a disc rotor. It is restrained by 4 or 5 bolts. Those things glow red hod when in use. Do the bolts holding the center lead to ultimate failure at the rotor center?

To tell the truth I am actually shocked, hearing that a part cannot be restrained from expanding.

A simpler example is a bimetalic strip. Does it fail?

Arrogance can be very dangerous. I wouldn't want to have a brake rotor on my car where the bolts have to move to allow the rotor to expand.

[747heavy]

I see where you coming from Ringo.
I think the "problem" with the tub/engine interface is, that it has to transmitt
a lot of torsional load. It is probabla one of the most limiting factors in achieving a good axle-axle torsional stiffness of the whole structure.

I don´t know if you have any specific manifold in mind, but almost any of the ones I have seen, are "loosly" fit. Which means something like an M6 stud/bold in a 6.2 mm bore. It relies on the clamping load, but there is very little additional stress to the manifold, apart from it´s own weight perhaps. IMHO

As an order of magnitude, there are some figures in the public domain, claiming the axle to axle stiffness of an (2000 F1 Ferrari) with ~ 6.5 kNm/°.
The monocoque/tub in it selfs (front to rear bulkhead stiffness) is said to be up to 40 kNm/°

A good Touringcar (Racecar) has values ~ 40 kNm/° from damper to damper point as well.

I can´t garantee that the figures are 100% acurate, or how well they relate to an current F1 car.
DaveW maybe able to but a ball park figure for axle to axle stiffness out.

In any case a F1 car is not very stiff in torsion (at least not when compared to a touring or sports car) The lower cross section, especially at the engine/gearbox interface are the main reason for it.

[747heavy]

I hope, you don´t consider me arrogant Ringo,
I don´t really have a dog in this fight, just try to learn something.

I´m not all that up to date with oad car braking technology, but in allmost every
racing car, I have worked on, the brake disc´s are mounted, so that they can float.
It´s the case with steel disc´s as well, as with carbon disc´s

[autogyro]

There is that surface fretting problem again riff_raff.

[xpensive]

...
To tell the truth I am actually shocked, hearing that a part cannot be restrained from expanding.

A simpler example is a bimetalic strip. Does it fail?
...

Sigh...please come back WB, all is forgiven!

[ringo]

I hope, you don´t consider me arrogant Ringo,
I don´t really have a dog in this fight, just try to learn something.

I´m not all that up to date with oad car braking technology, but in allmost every
racing car, I have worked on, the brake disc´s are mounted, so that they can float.
It´s the case with steel disc´s as well, as with carbon disc´s

Well i'm not familiar with race car tech.
I think these are rigidly fixed at the fasteners, but what i think happens here is that It was designed to have a low contact area to the centre piece, amongst other things it may be the best method to fix 2 disimilar materials such as carbon rotor and magnesium? center.

Here is my example where a part can be designed to take thermal expansion. I know this as fact.


we have a beam of length L and area A fixed at the left end and free to roll at the right end.
Under free thermal expansion with a change in temperature dT, it changes in length by an extension, e

e = alpha*L*dT, alpha being the coefficient of linear thermal expansion.

This beam will not be under any internal stresses since the beam is expanding freely.

In the second case, the right end of the beam is restrained by a spring and wall.
The wall is immovable. Spring constant is k and the spring extends by e.

Force in spring, P = e * K [1]

amount retrained = free expansion - spring extension = alpha*L*dT - e [2]

Agree so far?

for elastic materials:

strain* young's modulus = stress
extension e / length L * E = force P/ area A

P = e*A*E/L [3]

P = (alpha * L * dT - e)* A * E/L substituting [2] in [3]

e * K = (alpha * L * dT - e)* A * E/L substituting [1] in above

solve for e.

e = alpha*dT*A*E/ ( K + AE/L)

a sample, K= 60000N/cm, dT = 100, A = 1cm^2, E = 2x107 N/cm2, L = 100cm,alpha 10^-6
I went extreme on the K and T.

e = 10^-6 * 100 * 1 * 2* 10^7 / (60000 + 2*10^7/100)
e = 0.0076 cm

force = ke = 461N, stress = 461N/cm2 , way bellow yeild point yet it was restrained

It would have extended to 0.01cm without the spring. That's a factor of 1.3 times.

restraint yes, failure no.

Secondly take note of the order of extension, this is a 1m long rod that only moves out less than a mm.
These guys are over blowing the thermal expansion issue. A motor vehicle engine is child's play in terms of temperature... 100 degrees The difference with bulk head and engine wot even be that high.

[ringo]

...
To tell the truth I am actually shocked, hearing that a part cannot be restrained from expanding.

A simpler example is a bimetalic strip. Does it fail?
...

Sigh...please come back WB, all is forgiven!

Engineers of jargon and swagger, with a fear of numbers; it exposes the fallacies.

[xpensive]

I guess you tried to prove that a design subjected to thermal xpansion needs a built-in flexible element in order to cope?

Congratulations, I think that's xactly what the rest of this fine forum has tried to tell you all along, isn't it?

Need to work a little on attention to detail though, try to stick to consistent terminology and units, easier to follow.

[ringo]

No don't try to flip things around. Stop pretending like you had it figured.
Your jumping through loop holes more than ferrari.

Nothing is infinitely stiff, everything has a spring rate of some sort, I wouldn't consider 0.0076 cm "compensation" either.
You failed to provide any evidence,now i proved to you what is normal practice in engineering.

A spring rate of 60,000N/cm is also pretty much a brick wall,more stiffness than that carbon tub could provide.
I could have done that exercise with a spring that would actually compress the the rod as well, That's worse than a stationary wall.
The maths work out either way.

I don't work with inches, lb etc, like the yanks.
If a part is small i rather use cm. m and ft don't correlate well with bolts and other small stuff.

[Richard]

Ringo - The point is that the expansion of an engine block could not be restrained by a bolt. The loadpath needs to have some part that allows room to breath, ie some sort of ductility. Bending in the bolts is my hunch, combined with the spherical mounting xpensive has mentioned.

Here are the numbers ...

- Imagine you have a 400 mm long, 50 mm diameter Alu-bar, you heat it up 110 K to watch it grow one full mm.
- Put the bar in a hydraulic vice to compress it to its original length, you succeed with that, but at what force?
- Hooke will tell you that stress is epsilon times modulus, epsilon is 1/400 and E is 70 GPa, resulting in 175 MPa.
- 175 MPa over the area of a 50 mm dia bar equals 344 kN or about 35 metric tons, which would shear off an M24.

[ringo]

Ringo - The point is that the expansion of an engine block could not be restrained by a bolt. The loadpath needs to have some part that allows room to breath, ie some sort of ductility. Bending in the bolts is my hunch, combined with the spherical mounting xpensive has mentioned.

You cannot think like that. This breathing is what you want to happen or rather think should happen, but it doesn't have to happen. The designer can be sadistic and do the opposite; as long as the part is bellow the yield stress.
My example was what i think the tub would do without any built in devices for "compensation". The spring may have well be the slower expanding tub wall.

I posted a picture that you all overlooked, the engine was breathing very freely in all directions, only that the bolt pins were restraining it in certain areas.

Red line is the engine's new dimensions after thermal expansion, blue is original.
So you cannot say that i restrained the engine completely, i made that clear pages ago.
The restrictions will strain the engine in the zones near the bolts, but beyond the bolts, the cylinder heads will freely expand.

side view:


The engine will simply expand were it can. In the lengthwise direction the engine and gearbox will expand.

And when i say restrain, it could mean by reducing the expansion, it could mean completely eliminating it, and it could even mean reversing it.

You know, a FEA could end this debate, but since most of you detest it, I'll hold out for the sake of argument.
Little did they know, there are people out there who may have been taught actual engineering out of a book before they learned that dreaded FEA CAD.

[ringo]

Some food for thought:

a. Pipe of unity area is at an original temperature of 80*, no stress here.

b. pipe is restrained between 2 walls, of finite stiffness, pipe manages to expand 0.0008in after being heated to 480*

c. Pipe is restrain between 2 walls, of the starwars garbage compactor type .
Pipe decreases in dimension after being heated to 480* while walls apply garbage compacting force.

For free expansion:

delta L = adT = 7x10^-6 x (480-80) x 10in = 0.028 in

In case b. since the measured expansion was only 0.008in, the constraints must apply forces sufficient to produce a deflection of 0.020in (0.028-0.008).

delta L = PL/AE

0.02 = P(10)/(1)(30x10^6) P = 60,000 lb

area is unity so stress = 60,000 psi = 60ksi

In case c. where the length has been reduced despite a temp increase,

delta L is (0.028 - (-)0.008*) = 0.036 * negative since extension is reversed.

0.036 = P(10)/(1)(30x10^6) P = 108,000 lb

stress sigma = 108,000 psi = 108ksi

These are extreme cases, I can adjust my restraint if I feel that the bar was stressed to much in case b and worse in case c.
108ksi is pretty much super alloy territory, i could increase my bar area to 2 in^2
and thus reduce the stresses by half. If i have safety factor of 1.5 i could further increase my bar thickness, or loosen up the trash compactor.


, I plead to you guys again!! Know what you speak of, before you unwittingly belittle others.

i often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind.

Lord Kelvin

[747heavy]

I don´t want to take this thread of onto a tangent.
just in case you or any other reader is interested in brake disc to brake bell fixing. YOu can have a look here.

http://www.apracing.com/drawings/CP2872_1000CD.pdf

and yes, there is surface fretting in this part, but seeing that it´s service life is restricted by other constrains, it is not much of an concern, as far as the brake bell goes. There are other, perhaps more elegant methods, for this fixing, which some teams use, but all allow for a degree of expansion/movement between the disc and the bell. But a brake disc fixing is, IMHO, not a good comparsion for the problem at hand.

Nevertheless it usually does cater for the difference in the ETC of the materials involved, at least in racing brakes, where delta T is high (in case of the carbon brakes up to 1000K at peak load).
Which is may not the case for mountain bikes, scooters or light FWD car rear brakes.

[747heavy]

I think your first example, one end contrained, one end free floating describes almost any bridge construction in existence. ( which have to account for large values of l ergo a very large expansion length for a given CTE)

Your last example (outer space garbage compactor) is maybe a rough describtion of a hot forging process, which will lead to an increase of density in the material in question.

[endless]

...

Just 2 more pics of cosworth.