Angle of suspension brackets

[Ceboe]

Hi,

I ve used the search function to get an answer to my question but I didn't find anything so I'll just ask it here.
Why do a lot of teams put the front suspension brackets an an angle, and not horizontal? I've seen this with teams like delft and other monocoque teams.
In the picture below it maybe clear what I mean.
Does it have something to do with anti-dive?
Why can't they just put the bracket horizontal or in the plane fromed by the top suspension A-arm?

http://i39.tinypic.com/10cui6s.jpg

[Tim.Wright]

Probably to increase the possible articulation angle of the arm. For example, a rod end might be limited to 14deg of misalignment about axes transverse to the bolt hole before it mechanically binds up. If your control arm needs 20deg you can rotate the rod end so a component of the rotation is about the bolt hole axis which is effectively unlimited in terms of the angle it can reach without binding.

As to why they don't rotate them completely 90deg, i have no idea.

[Markus]

Tim, rotating the bracket will increase the local bending load on the monocoque, so you don't want to rotate it more than you have to.

[Big Bird]

I think there is an equally interesting question - why do we assume such brackets should be horizontal??

Or vertical for that matter...

[Georgy Arutyunyan]

I think that spherical bearings work better when subjected to radial load. Axial maximum allowed load is only about 10% from allowed radial load, if my memory says me right.
So I believe that it is better to mount brackets horisontally, bacause this way the loads from braking and accelerating will act as radial forces.

[Thijs]

...because this way the loads from braking and accelerating will act as radial forces.

Are you sure that's different for rotated brackets? FBD FTW :)

Our reasons for rotating the brackets are simpler than you may think.
As a hint for Ceboe, here are some pictures of DUT09 ('horizontal' brackets) and DUT11 (rotated brackets):

DUT09:
http://dutracing.tudelft.nl/wp-content/gallery/dut09-fsg/dut9_fsg3.jpg

DUT11:
http://dutracing.tudelft.nl/wp-content/gallery/dut11-testing_1/first-meters-dut11-03.jpg
http://dutracing.tudelft.nl/wp-content/gallery/dut11-production/img_9457.jpg (wheels are in droop here)

Thijs

[Z]

I think there is an equally interesting question - why do we assume such brackets should be horizontal??

Or vertical for that matter...

Or that you NEED 16 of them.

Plus another 4 for the spring-dampers. Oh, and must not forget the other 4 for the rockers. Must have those rockers...

So that's 24 so far, each with 2 holes to mount them to the chassis. So that's 48 accurately positioned holes in the chassis that need jigging. Or maybe you should mount them with 4 bolts each, for more optimal bolt loading? Yeah..., 96 accurately positioned holes...

Hmmm..., now what else can you add, so that you can claim that your latest car is "optimised for simplicity and reliability...".

(And let's be honest, you will be saying that, won't you?)

Z

[Big Bird]

Z, you forgot front and rear toe control. And I guess that each of these 2 force members have outboard points too. And that the independent arcs that all these links swing through necessitate outrigger bearings for the steering rack (and we should all assume such a thing is a necessity...) , and plunge and articulation for the driveshafts.

So many holes, and relative motions, and components, and deviations from assumed component rigidity, and manufacturing processes, and potential sources of assembly error, and tolerance errors, and non-linearities, and failure points, that all have to line up to get this car thing to act like it does in the simulation...

[Sormaz]

Z, you forgot front and rear toe control. And I guess that each of these 2 force members have outboard points too. And that the independent arcs that all these links swing through necessitate outrigger bearings for the steering rack (and we should all assume such a thing is a necessity...) , and plunge and articulation for the driveshafts.

So many holes, and relative motions, and components, and deviations from assumed component rigidity, and manufacturing processes, and potential sources of assembly error, and tolerance errors, and non-linearities, and failure points, that all have to line up to get this car thing to act like it does in the simulation...
simulation? what simulation?

[Tim.Wright]

Z, you forgot front and rear toe control. And I guess that each of these 2 force members have outboard points too. And that the independent arcs that all these links swing through necessitate outrigger bearings for the steering rack (and we should all assume such a thing is a necessity...) , and plunge and articulation for the driveshafts.

So many holes, and relative motions, and components, and deviations from assumed component rigidity, and manufacturing processes, and potential sources of assembly error, and tolerance errors, and non-linearities, and failure points, that all have to line up to get this car thing to act like it does in the simulation...

Our 2008 car was built almost entirely out of shims to counter that problem.

[Ceboe]

Are you sure that's different for rotated brackets? FBD FTW :)

Our reasons for rotating the brackets are simpler than you may think.
As a hint for Ceboe, here are some pictures of DUT09 ('horizontal' brackets) and DUT11 (rotated brackets):

DUT09:
http://dutracing.tudelft.nl/wp-content/gallery/dut09-fsg/dut9_fsg3.jpg

DUT11:
http://dutracing.tudelft.nl/wp-content/gallery/dut11-testing_1/first-meters-dut11-03.jpg
http://dutracing.tudelft.nl/wp-content/gallery/dut11-production/img_9457.jpg (wheels are in droop here)

Thijs

The only difference I can see is a curved plane vs a straight plane

[Thijs]

The only difference I can see is a curved plane vs a straight plane
Alright then.

It's got nothing to do with anti dive, or axial loading of the rod ends, which should never occur either way, since the force should always be perfectly aligned with the rod. If there are any significant bending moments in your rods, you're doing something very wrong.

As you can see from the pictures, both setups do a fine job of keeping the angle of the rod ends close to zero despite the angle of the a-arms with the horizontal plane. This was the starting point in the design of the brackets.
The only difference is that the DUT11 brackets are smaller (lighter) and much easier to produce, their faces being perpendicular to the monocoque plane they bolt to. The DUT09 brackets angle up, and they had to be CNC'd, rather than milled by hand.

Tim, Markus, Georgy, have a second look. If you do it properly there are no unwanted force components and moments. In fact, since the bracket can be more compact, thereby moving the rod end as close as possible to the monocoque, if anything you'll be loading the bolts more evenly, and minimizing the moment arm in situations where the rod end is under a slight angle (bump, roll). But in either case, the force vector should be (close to) crossing the virtual line between the two bolts.
Rotating 90 degrees would of course introduce unwanted moments in most cases. If you want to use these simple brackets, the required angle of rotation is precisely determined: the bolt should be parallel to the intersection of the plane that is perpendicular to the rod, and the plane that the bracket bolts to.

Finally, I'd say a reason that you mostly see this on monocoque cars is that they're the only ones to use separate milled brackets, rather than welded on tabs, so the ease-of-production argument doesn't really apply for tubular frames, although there may still be marginal weight loss to be found.

[Adambomb]

Alright then.

It's got nothing to do with anti dive, or axial loading of the rod ends, which should never occur either way, since the force should always be perfectly aligned with the rod. If there are any significant bending moments in your rods, you're doing something very wrong.

As you can see from the pictures, both setups do a fine job of keeping the angle of the rod ends close to zero despite the angle of the a-arms with the horizontal plane. This was the starting point in the design of the brackets.
The only difference is that the DUT11 brackets are smaller (lighter) and much easier to produce, their faces being perpendicular to the monocoque plane they bolt to. The DUT09 brackets angle up, and they had to be CNC'd, rather than milled by hand.

Tim, Markus, Georgy, have a second look. If you do it properly there are no unwanted force components and moments. In fact, since the bracket can be more compact, thereby moving the rod end as close as possible to the monocoque, if anything you'll be loading the bolts more evenly, and minimizing the moment arm in situations where the rod end is under a slight angle (bump, roll). But in either case, the force vector should be (close to) crossing the virtual line between the two bolts.
Rotating 90 degrees would of course introduce unwanted moments in most cases. If you want to use these simple brackets, the required angle of rotation is precisely determined: the bolt should be parallel to the intersection of the plane that is perpendicular to the rod, and the plane that the bracket bolts to.
This. I like this. So many people talk about gew-gaws and trinkets when dealing with innovation in FSAE, but that's bunk. What I find brilliant about it is that you no longer hold the artificial constraint of placing them vertical or horizontal, but instead begin with a simple mounting geometry and place it so it finds a nice, cozy normal point based on kinematics and the shape of the monocoque where the inner node intersects.


Finally, I'd say a reason that you mostly see this on monocoque cars is that they're the only ones to use separate milled brackets, rather than welded on tabs, so the ease-of-production argument doesn't really apply for tubular frames, although there may still be marginal weight loss to be found.

ISU used this design for over a decade now on tubular frame cars, although the reasoning was more for a combination of ease of adjustment and structural efficiency. Although again the constraints are a bit different on a tube frame, where your anti's determine the angle of the tube connecting the nodes (hence the immediate assumption that the design shown had something to do with anti's), and the mounting points will run longitudinal to the tube (hence the now-shown-false assumption that the mountings should be horizontal). With a tube frame it's simple to get the mountings normal to the load path by simply rotating the mounting point on the tube itself, as there is no "envelope" it needs to conform to. This is a clever solution to doing the same with a monocoque.