Can you field test a sway bar to find the "rate"?

So I have an idea for a REALLY cheap rear sway bar improvement for my challenge car. Problem is, I would like some way to get some "numbers" on the finished product. Sure I can go driving and get qualitative data, but the engineer in me wants a way to get quantitative data too.

Here is what I am thinking:

1. Jack up rear of car and place on stands, both wheels on.
2. Place floor scale in between bottom of one rear wheel and jack.
3. Jack one wheel up 1 inch. Read scale.
4. Measure distance between other rear wheel and ground.
5. Continue jacking wheel up until the other wheel moves up 1 inch.
6. Make sure chassis has not lifted off stand.
7. Read scale.
8. Maybe repeat on other side and average measurements.

Have I found the 'effective rate' (lbs/inch) of the sway bar? (lbs to move non-jack side 1 inch - lbs to move jack side 1 inch divided by 1 inch). Or should I just lose the floor scale altogether and just focus on distance compressed on one side vs distance compressed on other?

I would like to do this test on the stock bar as well as on my idea for a fabricated bar. I could then approximate the '35% stiffer than stock' type numbers as well.

care to divulge into you master plan? if its the same material then just measurement would suffice but it sounds like you maybe trying to cook up something a bit different? but if i were you and wanted to actually measure id make a jig to hold the swaybar and measure just it (jack on one side, scale on the other sounds good) its just the added suspension and stuff throws too many variables in.

I believe you can calculate the rate based on some measurements and some maths.

http://www.gtsparkplugs.com/Sway-Bar-Calculator.html

Couldn't you just fix one end and the mounting points to something rigid, hang a specified weight off the other, and measure deflection?

You need to hold the other endlink from moving, assuming IRS, having the opposite side suspension under normal compression should work. Also remove the spring from the corner under test.

In reply to Stefan (Not Bruce):

Doesn't that make assumptions about sway bar design, alloy, etc.?

Unhook one link, have a tape measure to record the deflection with a known weight hanging. Measure the distance from the link connection point to the mount location in the long axis only.

Sway bar stiffness goes up as the 4th power of diameter.

Two factors will interfere with your measurements with this idea: The stiffness of the rear springs and the unsprung weight of the rear suspension.

You should remove the entire rear suspension (or disconnect it entirely from the sway bar, if you can get enough room), connect a known weight to one side of the rear sway bar and the jack & scale arrangement to the other. That would give you some measurement of the stiffness of the bar itself.

Knurled wrote: Sway bar stiffness goes up as the 4th power of diameter.

Is this for solid, hollow, or both?

In order to get a wheel rate you should lock out the far side of the suspension (or just the far side swaybar arm). Then do your jack/scale/measurement process only on the near side. That gives you a true wheel travel to load measurement. Also others mentioned to remove the spring form the near side, which is a good idea and removes the spring wheel rate from your data. Make sure to collect more than a single data point and plot it up to get a good idea of the geometry change with droplink motion, which may not be linear. Dont forget to keep checking the ground to chassis measurements as you progress with loading, to get accurate wheel travel numbers and note when the jack stand gets free.
To lock out the far side suspension, you could place a collar over the shock shaft or build and end link from the swaybar that bolts to something solid nearby (chassis point).

Some good ideas here, thanks for them so far - removing near side spring, etc. Let me tell you guys a bit more about my idea and try to use some stolen (ok google images search) pictures to explain. All the suggestions to calculate based on diameter and lever arm length and such will not work, which is why I wanted to test what the actual rate on the car ends up being.

Ok, first, the car is a 99 saab 9-3. The rear sway bar is an interesting and super simple design. Here is a stock one and an aftermarket one next to each other. They just bolt on both ends, there are no center supports or bushings like you normally see on a sway bar.

A popular upgrade is to just double up the stock rear sway with another stock rear sway bolted to it, like this: (The added one was painted green)

My idea is to simply take a large length of angle iron and bolt it on bottom of the stock sway. Straight across, no 'u' shape at all. I would essentially be adding 4 holes to the horizontal plane of the angle iron, and then the vertical plane of angle could be trimmed down if the resulting 'bar' was too stiff.

Test it by hanging weight off of it, meaning pulling it down towards the ground, with one end disconnected.

Robbie wrote: All the suggestions to calculate based on diameter and lever arm length and such will not work

Yes, they will. If you put angle iron on a straight bar the lever arm length is from the bolt holes in the angle to the center of the bar. If you use a hollow bar or solid you just need to do the right math to calculate the force it takes to twist it given that lever.

You can field test it by bolting one end to a fixed object and putting a giant torque wrench on the other.

Huckleberry wrote:

Robbie wrote: All the suggestions to calculate based on diameter and lever arm length and such will not work

Yes, they will. If you put angle iron on a straight bar the lever arm length is from the bolt holes in the angle to the center of the bar. If you use a hollow bar or solid you just need to do the right math to calculate the force it takes to twist it given that lever. You can field test it by bolting the other end to a fixed object and putting a giant torque wrench on the other.

I'm sure you can and this stuff has already been modeled by someone somewhere, but I feel quite certain that getting any sort of math that turns into a realistic estimation of reality in this situation will be much harder than just testing the finished product. For example, where is the 'center' of a piece of angle iron? What is the 'diameter'? What if the lever arm length drops to zero? The roll stiffness surely doesn't go to infinity, but suddenly the bar starts bending differently. At some point you probably start seeing real effects from both twist and bend, and how much does each affect the result?

Again, I'm sure it can be done (I have a degree in physics), I'm just not sure I want to or that anything about it would be accurate enough to be very realistic.

EDIT: the torque wrench is a good idea!

I have a home made sway bar that consists of a stock bar with another piece of steel rod welded to the long center section. It doesn't extend the full length. Instead it short cuts and triangulates down and joins up with the ends. Forms an open triangle at both ends. I think the main resign it does this is because there are front mounts that still bolt around the stock part of the sway bar near the ends. It works exceptionally well, but no way of knowing how much stiffer than stock it is. It does allow me to run softer springs than what I would need with a standard uprated sway bars commercially available on the race car in my sig. Really like it on bumpy tracks where you still want to corner flat, but need softer springs to keep the wheels in contact with the bumpy surface. This thing came with a bunch of spare parts that came with another TR8 race car I bought years ago. Think the guy that originally made it was strictly an autocrosser.

In reply to Robbie:

The bushing mounts are there to limit bend - and in doing so introduce stiction that won't be part of a calculation but - you will get very close to the real rate of the bar alone by measuring carefully. You need diameter of the bar, thickness of the tube if hollow, length and lever and the constant for the steel alloy you choose.

If the lever goes to zero then you are left with only the force to twist the bar - like the blade type bars used in karts that adjust my changing the relationship of the flat to the direction of force. The lever you will be using is the distance from the center of the point it mounts to the car to the center line of the straight bar. So, if you use angle iron - it's the distance from the center of the bolt hole to the center of the bar measured perpendicular to the center of the tube.

So, do the math, then field test. If you measure carefully and calculate correctly... next time you will trust the math. :)

Robbie,
Since you are trying to come up with a non-typical bar, you should just use the empirical methods described (measurements with parts fitted to the car). This will give you the best real world information for what is supposed to be a simple and cheap project. You can make the approximations for the standard bar calculators, but they are just approximations.
One of the difficulties of using angle (or any other open cross-section) is the center of twist is located outside the shape while the center of mass is inside the shape. This puts the highest stresses in odd locations compared to a round section.
The idea you are describing is the same idea behind the twist beam rear axles on alot of front drive cars, so there is some history behind making the system work as you describe it. Keep in mind alot of those twist beams were too stiff, but you are trying to figure out what works and how to adjust it, so you are on the right track.
Make several versions and test them all at the same time so you can swap them out during a test event if you find you need more/less rear stiffness.