Pick up point loading...

[Christopher Crowe]

Does anyone happen to know the approximate fore-aft load an upper inboard pick-up point sees on a typical FB car? I know it can't be astonomical, but I'd love to have some kind of ball-park number if anyone has one.

Thanks,

Chris

[S Lathrop]

Your question is too vague.

I can say that I use and recommend for all my cars 5/16 mild steel rod ends for the front upper a-arms, Aurora MM5TY. These have never failed in 28 eight years of racing. You can look up the strength ratings for those rod ends.

The rod ends are aligned with the legs of the a-arm and never see any sheer loading.

These rodends do brake in crashes but, frequently, the upper a-arm will not be damages beyond braking the rod end. I almost never have to repair the upper a-arm pickups after a crash either.

Is that the information you want?

[R. Pare]

Chris:

Look up "free body diagrams" and "vector analysis". With those, and knowing your loads at the tire contact patch, you can calculate the forces at the pickups.

[nulrich]

As Richard says, you really need to calculate the forces seen at the ends of each wishbone. Unless you have a strange suspension geometry, you don't have to worry about the vertical component of the forces, so it's a very simple 2D statics problem.

You do need to consider braking and acceleration forces.

Nathan

[Christopher Crowe]

And thanks.

C

[rmstringham]

As Richard says, you really need to calculate the forces seen at the ends of each wishbone. Unless you have a strange suspension geometry, you don't have to worry about the vertical component of the forces, so it's a very simple 2D statics problem.

You do need to consider braking and acceleration forces.

Nathan

Actually its statically indeterminate.

[JJLudemann]

In case any other people check this thread, my results are here:

Calculating the Loads on the Front Suspension


-Jim

[john f]

When looking at loading on a suspension joint or frame member for that fact, do you look at just the stresses created or do you look at the deflection of the joint? Even though a joint or member may have a stress that does not excede the yeild or fatigue strength, can not that joint or member deflect enough to have an effect on alignment or ??? I'm thinking of toe control or caster/camber problems. When people say that a lb or 2 of corner weight can make a difference in car handling or if your scales are out of alignment by a 64th of an inch, I would think that the added deflection of a 1/4" rod end vs. a 3/8" rod end in the same position could have a hell of a lot more influence. Comments???

john f

[rmstringham]

In case any other people check this thread, my results are here:

Calculating the Loads on the Front Suspension


-Jim

Is there a reason you are only analyzing the braking load? Is a 4G brake scenario the worst loading possible? Doesn't a suspension experience forces in all three directions? You might want to add the push rod and tie rods to complete the model.

That being said, the fact that you are using a FEA program to solve this problem means you are on the right track for properly determining the suspension loads.

[rmstringham]

When looking at loading on a suspension joint or frame member for that fact, do you look at just the stresses created or do you look at the deflection of the joint? Even though a joint or member may have a stress that does not excede the yeild or fatigue strength, can not that joint or member deflect enough to have an effect on alignment or ??? I'm thinking of toe control or caster/camber problems. When people say that a lb or 2 of corner weight can make a difference in car handling or if your scales are out of alignment by a 64th of an inch, I would think that the added deflection of a 1/4" rod end vs. a 3/8" rod end in the same position could have a hell of a lot more influence. Comments???

john f

Short answer, yes. Suspension deflection can clearly influence tire orientation and handling behavior. How much is acceptable is dependent on your weight budget, tire sensitivity among other things.

[JJLudemann]

When looking at loading on a suspension joint or frame member for that fact, do you look at just the stresses created or do you look at the deflection of the joint? Even though a joint or member may have a stress that does not excede the yeild or fatigue strength, can not that joint or member deflect enough to have an effect on alignment or ??? I'm thinking of toe control or caster/camber problems. When people say that a lb or 2 of corner weight can make a difference in car handling or if your scales are out of alignment by a 64th of an inch, I would think that the added deflection of a 1/4" rod end vs. a 3/8" rod end in the same position could have a hell of a lot more influence. Comments???

john f

That blog post is a little out of date. Based on the feedback I received here, I did more research and found a derating factor must be used to turn rod-end max loads into allowable operating loads, and the factor can be high. I ended up upgrading my 1/4" rod ends to 8mm (approximately equal to a 5/16"). Unfortunately, the ship had already sailed with my big order from the US, and only metric rod ends can be sourced locally.

-Jim

[Billy Wight]

the fact that you are using a FEA program to solve this problem means you are on the right track for properly determining the suspension loads.

FEA is not magic, it's math. It's only as good as the parameters you put into it, and I would argue the contrary: The fact that you're using an FEA program to get these results throws up a big red flag to me that they're probably wrong.

What were the assumptions used? What were the boundary conditions on the model and are they realistic? Were any hand calculations done to verify the FE result? How did you get the loading numbers you input into the model? What is your failure criteria? Is a static approximation good enough? What might another faiulure mode be (buckling, fatigue, etc.) and was it evaluated as well? How is the element quality? Was a mesh convergence study done (and would it matter for these results)?

So many questions to consider, yet one colorful plot often puts peoples minds at ease so easily...

[Wright D]

FEA is not magic, it's math. It's only as good as the parameters you put into it, and I would argue the contrary: The fact that you're using an FEA program to get these results throws up a big red flag to me that they're probably wrong.

What were the assumptions used? What were the boundary conditions on the model and are they realistic? Were any hand calculations done to verify the FE result? How did you get the loading numbers you input into the model? What is your failure criteria? Is a static approximation good enough? What might another faiulure mode be (buckling, fatigue, etc.) and was it evaluated as well? How is the element quality? Was a mesh convergence study done (and would it matter for these results)?

So many questions to consider, yet one colorful plot often puts peoples minds at ease so easily...

I totally agree with Billy. It is very easy to setup a FEA simulation incorrectly. I strongly recommend doing lots of hand calcs and some sort of real life correlations. Even us engineering "experts" always verify that our simulations and models are producing realistic and useable data.

[Christopher Crowe]

Over years I have seen the fruits of the pure theoretical/mathmatic/computational approach to engineering meet with some disasterous results.

It's nice to hear two gifted engineers talk about the phenomenon. Computers have great authority. But always question the hell out of it...

[john f]

Over years I have seen the fruits of the pure theoretical/mathmatic/computational approach to engineering meet with some disasterous results.

It's nice to hear two gifted engineers talk about the phenomenon. Computers have great authority. But always question the hell out of it...

As they say:

"Sh#t in, Sh#t out"

john f

[R. Pare]

Some FEA programs are pure sh!t. One I used - once - told me that my 20-year-old caliper piston design would burst the back wall at less than 500 psi !

Strange, that, considering that I would destruct test the new caliper designs at 1500 psi for maybe 30000 cycles!

[rmstringham]

FEA is not magic, it's math. It's only as good as the parameters you put into it, and I would argue the contrary: The fact that you're using an FEA program to get these results throws up a big red flag to me that they're probably wrong.

So many questions to consider, yet one colorful plot often puts peoples minds at ease so easily...

Please dont put words in my mouth. i never said that FEA was magic. I am fully aware that behind all the pretty colors, it is only math and material mechanics. I am also aware that the validity of the analysis is dependent on the input of the user. The point i was making is that the finite element method can be a very accurate means of determining suspension loads, if done correctly (JJLudemann was on the right path, but his analysis was still incorrect).

You're absolutely right. FEA in the wrong hands can be extremely erroneous, and dangerous. The reason that using the finite element method (by hand or computer software) is the more correct method over a 2-d hand calc is because the suspension system, with multiple parallel load paths, is statically indeterminate. You cannot solve for the forces in ALL the members by simple statics. Period.

Scoff all your guys want, the old school will be taken over by new school before you know it. And your cars will be left in the dust. If they ever get built...

[rmstringham]

Over years I have seen the fruits of the pure theoretical/mathmatic/computational approach to engineering meet with some disasterous results.

I've seen the same thing happen when the blind lead the blind.

It's nice to hear two gifted engineers talk about the phenomenon. Computers have great authority. But always question the hell out of it...

What am I Chris, chopped liver?
Let's not forget to grab a beer before I ship out.

[rmstringham]

Some FEA programs are pure sh!t. One I used - once - told me that my 20-year-old caliper piston design would burst the back wall at less than 500 psi !

Strange, that, considering that I would destruct test the new caliper designs at 1500 psi for maybe 30000 cycles!

As they say:

"Sh#t in, Sh#t out"

john f

QED

[Wright D]

You cannot solve for the forces in ALL the members by simple statics. Period.

Scoff all your guys want, the old school will be taken over by new school before you know it. And your cars will be left in the dust. If they ever get built...

[FONT=Verdana]You can calculate the loads in each member. [/FONT]

[FONT=Verdana]There are typically three pick up points on an upright. The lower typically holds the upright constrained in 3 degrees of freedom; the upper point in two degrees, and of course, one last point controlling the upright in one degree. This is not a statically indeterminate system. Since the loads form the uprights get fed back into the A-arms, it is quite possible to calculate the loads in theses. [/FONT]

[rmstringham]

[FONT=Verdana]You can calculate the loads in each member. [/FONT]

[FONT=Verdana]There are typically three pick up points on an upright. The lower typically holds the upright constrained in 3 degrees of freedom; the upper point in two degrees, and of course, one last point controlling the upright in one degree. This is not a statically indeterminate system. Since the loads form the uprights get fed back into the A-arms, it is quite possible to calculate the loads in theses. [/FONT]

The force at the contact patch is a vector with 3 components in space, though I will concede that there may be times when one or more of those components is zero, which would simplify the analysis to an extent.

While the lower upright pickup typically does constrain the vertical motion of the upright, the load is reacted by both the lower a-arm, and the push rod. Also, while the tie rod does react moments about the steer axis, it is also a load path for reacting the internal forces of the upright.

I totally agree with Billy. It is very easy to setup a FEA simulation incorrectly. I strongly recommend doing lots of hand calcs and some sort of real life correlations. Even us engineering "experts" always verify that our simulations and models are producing realistic and useable data.

When its all said and done, this^^^^

The ideal design process would not start and end with FEA. I hope my input was not interpreted this way. Before starting any advanced analysis, you have to start from first principles, in this case a free body diagram to identify the externally applied forces and the constraints/reactions. Then if required, using an engineering tool (thats all it is, a tool, you gotta know how to use it though) to help solve a complex problem. You wouldn't try to determine chassis stiffness by hand would you? Then to validate the model/simulation, physical testing and correlation are necessary to show that you aren't just chasing rainbows.