17 replies to this topic

[Canuck]

I'm reasonably certain this is a bad idea but I want to make sure first.

I've swapped out the rear subframe and suspension in my old BMW for a slightly later model (as discussed elsewhere in here). One of the benefits of doing that is the greater availability of suspension upgrades in the aftermarket. However, I still have the 'old' front suspension, which means I'm drastically limited to what's available to me, if you can find anything.

So - it would be an unwise decision then install a beefier rear sway bar and springs if I can't install matching components in the front?

Is it possible to fab up your own swaybars?

[Greg Locock]

Roll your own:

Not easily, as a production sta bar is made out of some nice steel, which is heat treated after bending. However, when we get them made locally at the speed shop they only cost about $400 a pop, I'm sure the price you'd get is better than that.

Well, I suppose you could buy the bar, bend it up, and then get someone to heat treat it for you.


Stiff rear suspension, softer front suspension

bad bad bad bad bad. Depends a bit on exactly how soft the front is, for instance our sporty cars run a front spring that is 2.4 times as stiff as the rears, whereas the rates for the base car almost match front to rear, so it is possible to fit a somewhat softer front end to the sporty one without killing yourself.

[Canuck]

So...what you're saying is - it's not the best idea then

We do have heat-treat people we use daily at the machine shop. I'm sure I could build a jig to ensure I had the correct shape. Then I suppose one has to come up with new mounting bits for the larger diameter - not impossible. But...

What material to use? If I increase the dia by say 10%, what does that do?

This is perhaps another project best left for the future.

Edit - Whiteline in Aus makes swaybars for my car - go figure. What is the downside to sticking a heavy (27mm) bar on vs. a 'replacement' bar (23mm). (I noticed the rear was 18mm - significantly smaller)

[ray b]

take a tape messure to a junk yard and see if another cars bars will fit
swaping used parts sure is cheaper then custom made

btw BMW springs go soft over time

[blkirk]

The stiffness of a round bar in bending and torsion is a function of D^4. So, a 10% increase in diameter gives a 46% increase in stiffness. 1.1^4=1.4641

It might be easier to make a pair of adjustable drop links for the front suspension if you have the room. Making the radial arm at each end of the sway bar shorter by moving the drop link will also increase your roll resistance. The math on this one is a bit more complicated, but reducing the radius arm length by 10% should increase the stiffness by about 30%. Beware that this will also increase the stresses in the bar and therefore decrease its fatigue life.

[Greg Locock]

A 27 is a big step up from a 23. That would be our full tuning range and a bit more, for everything from taxis up to sporty cars.

It wouldn't necessarily hurt though. The front end will have more understeer, and will tend to hold the rear axle flatter, improving your traction out of corners.

I wouldn't get too frightened by that d^4 number. Hardly anybody designs the mounting system for their sta bar to extract its full performance - the D blocks and the links and the stiffness of the body to which the D blocks attach all actively reduce the effect of diameter changes.

If you work out the total roll stiffness of the car as is, then you'll see that most is supplied by the road springs, the a/r bars just provide the front to rear balance. I'd guess that even a 90% increase in a/r bar stiffness at the front wouldn't change the front to rear roll stiffness by more than 10%, so your ride will suffer, but hopefully the grip will compensate for that.

You need to use the same sort of steel as is used for coil springs, 4140 is a grade I recognise. I don't know what heat treat is used. You can work out the torsional stress, which is probably the killer. Figure on 50-60mm deflection across the axle.

[blkirk]

Originally posted by Greg Locock
[B]I wouldn't get too frightened by that d^4 number. Hardly anybody designs the mounting system for their sta bar to extract its full performance - the D blocks and the links and the stiffness of the body to which the D blocks attach all actively reduce the effect of diameter changes.

Quite true. The entire vehicle is a network of springs all connected in parallel and series. If you increase the stiffness of the stiffest spring, it will have almost no effect on the behavior of the system. If you manage to find the softest spring and make it stiffer, the change in the system will be nearly 1:1. So, if you have a stiff sway bar mounted with soft bushings, changing the sway bar will have almost no effect since most of the travel will be taken up by the bushings anyway.

If you work out the total roll stiffness of the car as is, then you'll see that most is supplied by the road springs, the a/r bars just provide the front to rear balance. I'd guess that even a 90% increase in a/r bar stiffness at the front wouldn't change the front to rear roll stiffness by more than 10%, so your ride will suffer, but hopefully the grip will compensate for that.

I have changed the front sway bar on two different vehicles. In both cases, the change in sway bar stiffness was quite large (~2x). The increase in understeer was obvious but manageable. It did not make the car undrivable. I agree with Greg that it might be a useful improvement as long as the car doesn't understeer horribly already.

Just to clarify, the 27mm vs. 23mm are your options on the front, or have we gotten lost along the way?

[Canuck]

Maybe I should just lay it out on the table instead of trying to figure it out by beating around the bush.

I have an older BMW ('81 E12) that I enjoy driving spiritedly. It is my daily driver and will have to suffice for both summer and winter conditions. It handles reasonably well at present but I'd like to reduce the body roll and be able to push it a little harder while enjoying the clover-leaf on and off ramps. Of course I'm leery of anything that will make winter driving any more 'exciting' than it already is. And while I want more sport, I don't want it to hammer the bondo out of the body either.

The 27 and 23mm specs are for the front bar, 18mm is the rear.

[McGuire]

For winter driving you definitely want to go light on the rear ARB.

[Goran Malmberg]

You need to use the same sort of steel as is used for coil springs, 4140 is a grade I recognise. I don't know what heat treat is used. You can work out the torsional stress, which is probably the killer. Figure on 50-60mm deflection across the axle. [/B]

If talking race cars using conventional sway bars, the wheeltravel is so short that any steel could be used as long as the "floating limit" (if so called) of the material in question is keept within 10 - 15 times the deformation limit.

What do you guys think?

Goran Malmberg

[Greg Locock]

Well, your terminology is unfamiliar to me. The problem is, if you yield the material then after each turn the car will hold a permanent roll on the next straight. I agree that on a race car you needn't worry about the fatigue life as such, so long as you keep careful track of how many races each bar has done, and cut them up when they are about to break.

But that sounds like more trouble and expense than making them properly out of the right material.

[Goran Malmberg]

If anyone is interested here is an swaybar test I did a few years ago. It show the difference in pring rate using different angle arm out to the wheel. (Hope the image will show up).

Goran Malmberg

[NTSOS]

[Greg Locock]

Goran , did you have a D block near the outboard end? I assume so from the first and last test.

The whole field of the shape of sta bars when you move away from the simple geometry is quite interesting for production cars. The other interesting thing you can do is use the sta bar to change the roll steer - Very handy.

[Goran Malmberg]

Greg, at the end where the word says "Fast" which means the same as "solid mount" in Sweden.
The bar is welded to a piece of iron that is put in a big vice. In the other end, close to the bend, the piece of iron has hole for the bar, still in a vice.

The reason for the test was that many books have formulas for sway bar that don’t take
arm angle in to the calculation. There are also all kinds of theories about how a swaybar work.

The test bar represent the half lengths of bar as mounted in the car. The middle of a bar is not moving-twisting and as fare as roll motion is concerned the bar could as well be welded solid to the frame in the middle of car. To picture this I made another version for the last example that has one arm each direction for the same torsion bar length. At the same load of 185 kg the deflection range is -120 to -430 mm.

I putting these "bending" motion in, in a calculation we do not get exactley the same numbers, but as allways I end up with messuring errors, but n this case not to big ones.

Sorry for using bad terminology, but I dont have the strenght to go to the dictionary and spend time. I just have a few minutes once in a while to spend on "forum activities".

Goran Malmberg

[McGuire]

Of course there is ARC, Active Roll Control as used on BMW 7 series and others... the ARB is split at the center and compensating inputs are fed into the frame end of the bar with hydraulics in response to various inputs. I have driven it and it does not feel as creepy as one might imagine.

Then there is SARC, Semi-Active Roll Control. This is a really simple setup that uses an on/off actuator in the ARB's vertical link. There are inputs to a controller: road speed, g lat, steering wheel angle, stop lamp switch. The actuator decouples the ARB in straightline drving, recouples in cornering etc. TRW is shopping one around but I don't know who is running it.

[scooperman]

There was an idea a few years back for linking sway bars together to get anti-dive, one just had a mechanical link. The other one put another brake cylinder in the middle of each sway bar, pushed downwards on the front suspension, pulled up on the rear. I think Ferrari tried a version of this and quickly dumped it.

[Greg Locock]

http://www.kinetic.au.com/