Sway bar attachment methods

[tdzanic]

Hello, I'm Tarik from Princeton's Formula Hybrid team. I was wondering what are most teams' methods of attaching the sway bar to the lever? Welding seems like an obvious choice, but since our design calls for the bearing housings to be on the inside of the levers, it would prevent it from being disassembled.

[JT A.]

Your bearing housings could split apart at the centerline, so you unbolt them and the sway bar just drops out with the bearings and half of the housing. Do you think a split journal bearing would work for your application? Or you could have the ends of the sway bar machined with a male hex shape and the "levers" machined with a female hex shape so it just slides over the end of the sway bar, with a screw/cap thing on the end of the bar to hold it on. That's all the thinking I'm going to do for you.

[tdzanic]

Your bearing housings could split apart at the centerline, so you unbolt them and the sway bar just drops out with the bearings and half of the housing. Do you think a split journal bearing would work for your application? Or you could have the ends of the sway bar machined with a male hex shape and the "levers" machined with a female hex shape so it just slides over the end of the sway bar, with a screw/cap thing on the end of the bar to hold it on. That's all the thinking I'm going to do for you.

I was thinking about doing a spline type connection, but more so with just a keyway than a hex spline. Interesting that you mentioned the split journal bearings, I didn't think of that. That's probably the method I'll go with. Thank you!

[Claude Rouelle]

Just a perspective; on the many many FSAE / FS cars I have observed in more than 15 years: I would say that 80 % of ARBs are not working; too much play, too much compliance, major design errors (such as droop link not in the plane of the rocker - that is you use inboard suspension, single shear rod end attachment, huge distance between the ARB arm and the ARB housing etc...) The installation stiffness (I should say smoothness) is such that the real TLTDD (or as I call it "magic umber" of front/total weight transfer distribution) must be far, far away from the calculated Excel spreadsheet.

[JT A.]

Very true, Claude. We had a front ARB on a fairly recent car that didn't engage until the car was at about 2/3 of its total roll travel. I wanted to just take it off the car, but "Noooo, we have to have our fancy titanium blade electronically adjustable ARB so we look smart in design". But they wouldn't let me hang some titanium blade wind chimes off the roll bar either, leaving me thoroughly confused about my team's decision making criteria.

[Claude Rouelle]

JT A,

Good for you and your team I did not judge your car. :)

There is a simple test a judge or any student can do. Grab the car by the roll hoop and push or pull sideways to simulate some lateral force and roll moment. Or ask somebody else to do it and carefully watch the ARB movement. Sometimes the play is very, very visible.

Want an illustration? On an Indy car on oval the ARB motion ratio is slightly different on the left and right; you can see it when you look at the LF and RF rockers; the ARB droop link attachment point is slightly different. The goal? Even if the suspension movement is the same on the LF and RF (let's say same LF and RF spring and same LF and FR aero downforce - remember in Indianapolis the straight line is flat, no banking) then the front ARB will still be slightly pre-loaded before the driver enter in the corner. If this preload wasn't there it will take about 30 meters (30 meters at 370 km/'h is a short time....but a short time without any ARB weight transfer control) for the ARB movement to "beat" the play.
Driver subjective feedback, pushrod strain gauge data as well and steering aggle, steering torque, gyro and slip angle, lateral acceleration sensors will show you the difference.

Of course that does work on a oval because you only turn left but this story, I hope, tells you the importance of decreasing the play ...and by the way the other way around: the compliance

Cheers

Claude

[Jay Lawrence]

Claude,

Admittedly I don't follow and have no interest in oval racing, but I'm confused about why you would need an ARB? I wouldn't have thought there'd be much pitch/heave/single wheel bump/rebound going on at these tracks, so what's the contribution of the ARB that can't be provided by corner springs alone? Is it just that you can quickly access higher resolution adjustments (compared to changing/modifying springs)?

edit: silly me... I'm guessing the (obvious) answer is load sensitivity

[Claude Rouelle]

Jay,

That topic has been covered many times in this forum but here we go again.

The ARBs are there for 2 reasons
1. Control the roll
2. Adjust the anti roll stiffness distribution therefore the front and rear weight transfer or if you want the dynamic cross-weight and therefore (and because, as you said, of the tire load sensitivity) therefore the understeer/ oversteer characteristic of the car

Yes you can control the roll with springs but then a) to get the anti-roll stiffness distribution you want you will need such small spring stiffness increments that does not exist. b) you will change the ride frequency and the attitude of the car and that will be an issue even more if your car is aerodynamically ride height sensitive.

Good driver can feel a difference of 0.05 % of anti-roll stiffness distribution. Try to do that with springs or bump rubber only. There is a reason why such small adjustments are possible (with adjustable inboard ARB) on professional race cars and there is no reason that wouldn't work on a Formula Student car. hat is of course if there isn't too much play and/or compliance , which was my initial comment.

Claude

[DougMilliken]

... Yes you can control the roll with springs but then a) to get the anti-roll stiffness distribution you want you will need such small spring stiffness increments that does not exist. b) you will change the ride frequency and the attitude of the car and that will be an issue even more if your car is aerodynamically ride height sensitive. ...

Claude, why do you limit your thinking to coil springs? Torsion bar primary ride springs have been used on many cars. By changing the active length, the spring rate can be adjusted in very fine increments.

See for example, Bill Milliken's "Equations of Motion" page 498. Dad's MX-1 "Camber Car" has square torsion bar springs with a clamping mechanism (on the chassis end) that slides along the length of the spring for rate adjustment. The clamp also incorporates an angular adjustment for ride height.

Many other configurations are possible.

[Claude Rouelle]

Hello Doug (we miss you as a judge at FSAE events!) ,

Good thought. I do like Torsion Bars too; possibility of different Motion Ratio than dampers (compared to the coil over), they work both ways (clockwise and anticlockwise) while i never saw a suspension spring working in tension, often better for packaging (at least on race car, often in the center of the rocker. A bit of challenges for some teams when it is about manufacturing splines.

All that is in page 567 of our 2015 fall seminar :)

[Kevin Hayward]

Don't need to make it a splined end. You can start with stock that is a convenient shape for retaining, and then turn down to the torsion rod.

Kev

[Claude Rouelle]

Kevin,

HI! What is a "stock" ?

Claude

[Pat Clarke]

Stock is 'the part of the gun you make soup from' Claude (old silly definition of 'Stock' in karting).

Kevin is using the term as used in Australia when describing steel bar, ie, 'round stock', 'square stock', 'flat stock', 'hexagonal stock' etc

Pat

[Claude Rouelle]

Pat, Kevin,

OK, thank you for te Australian translation.... then what? How do you connect one end of the torsion bar with the suspension (rocker for example) and the other end with the chassis? Because if we are speaking about an hexagon or square section end, I am afraid we could go back to the square one story about play and clearance,

Claude

[Adam Farabaugh]

Claude,
Do you think it's impossible to preload the joint enough to eliminate free slop?

[Claude Rouelle]

Adam, unless I misunderstand your question: The pre-load in one rotation direction will become a play in the other,

[DougMilliken]

Claude,
I don't think there is any play in the MX-1 torsion bar ride-spring installation. I don't have the drawings handy to check, but from memory the square bar is clamped like it might be in a bench vise (or across opposing corners between little "V-blocks"). Plenty of clamping load trying to crush the bar.
-- Doug

[Z]

Claude,


...the square bar is clamped like it might be in a bench vise (or across opposing corners between little "V-blocks").

Not obvious???
~o0o~

And thinking about the "prior art". What was that funny little car that sold, ooh..., more than 20+ million times...? The, err, ... something... "Beetle"?

As I recall (it was many decades ago that I last looked, although plenty of them still around...), they have front springing via lateral torsion bars, with said "bars" made from a stack of rectangular leaves, which together make up a roughly ~20 x 20 mm "square bar". Inner leaves of stack about 20 x 4 mm, and outermost leaves ~10 x 4 mm.

This arrangement of multiple leaves gives a lower rate and more maximum twist angle for a given length of bar, hence more strain-energy capacity (ie. more efficient because less unstressed "dead area" in the centre of the bar/s). And no problems if one leaf breaks, just a slightly sagging suspension...

These top and bottom, lateral, "square torsion bars" are fixed to the twin-trailing-suspension-arms (at each side of car) with a ... pointed grub screw! The point of the grub-screw comes in "sideways" to the stack, thus wedging apart the inner leaves of the stack, and thus "clamping" the bar-stack inside the squarish hole in the trailing-arm. This then allows the bar to carry the torsional forces without "fretting", and also the wheel's lateral (Fy) forces that try to pull the trailing-arm out of its housing.

Methinks both students, AND Design Judges (!), should spend more time climbing around car wreckers yards... :)

Z

[Kevin Hayward]

Claude,

First off I am sorry if the word "stock" is not used outside of Australia. I am often caught by surprise how often the Australian use of the English language is just better. I'm fairly sure that in 1000 years or so we will have eliminated most of the syllables from the language and be able to communicate complicated engineering concepts with little more than a grunt.

I think the play and clearance issue has been dealt with after my post.

I would add that in practice splines can be more difficult to reduce slop than profiles with defined flat surfaces or angles. This is one of the reasons that some of the very good steering quick releases do not use spline profiles.

Kev

[Adam Farabaugh]

Yes we simply machine an external hex on the end of our ARB torsion springs, and an internal hex in the ARB arms with a large slit, and we simply clamp the s*** out of it. I guess this is not what you would normally think of when you consider a preloaded joint, but you can design the clamp arms to undergo a rather large elastic deformation around the hex, resisting rotation in both directions.

If the slit is not large enough or the internal hex is machined too large then yes it does not work, but this is just an exercise in GD&T.

I would love to try polygonal splines as a method to reduce stress risers at the clamped locations. Traditional splines are great but difficult to machine in a typical student shop, although we are having good results with our jerry-rigged 4th axis and some clever work in our post processor.

[Claude Rouelle]

I used torsion bar in 1983, 1984 and 1985 on the front suspension of the Alfa Romeo GTV6 si Iam familiar with such device and yes we never had any play. We had splines but I believe it is possible to do it with an hexagon or even a square end section but honestly I have of experience of such torsion bar end. However waht I shoud have said is that I have seen sevral of such devices in FS / FSAE with tons of play. I even remember the patch on a patch of 4 screws on the female square section to put pressure on each face of the male square section. So Mea Culpa for the lack of experience on that topic. What I saw was simply a good concept badly executed.

Now, Kevin: ".... the Australian use of the English language is just better" You guys down under are so full of it!

Claude