How to dyno your anti-roll bars at home

By Andy Hollis
Mar 5, 2023 | DIY, how-to, anti-roll bar | Posted in Shop Work , Suspension & Handling | From the Dec. 2014 issue | Never miss an article

Photograph Courtesy Mazda

[Editor's Note: This article originally appeared in the December 2014 issue of Grassroots Motorsports.]

Anti-roll bars are a common and inexpensive way to tune the suspension of your car. Generally speaking, adding a larger front bar means less oversteer, while adding a larger rear bar means less understeer. But that just tells us the direction of the change, not the amount.

Sure, there …

This content is available for GRM+ members and Grassroots Motorsports magazine subscribers only.

You can read it for free in 31 days or subscribe to GRM+ to read right now.

Subscribe now

Already a member?

Login to read

Join Free Join our community to easily find more DIY, how-to and anti-roll bar articles.
More like this
View comments on the GRM forums
Cactus Reader
6/18/18 4:16 p.m.

So obvious, yet not something I've thought to do. I'll be sure to give this a try some day.

malibuguy GRM+ Memberand Reader
6/19/18 11:14 p.m.


te72 Reader
6/25/18 11:32 p.m.

Had some sway bar shenanigans pop up recently. When the weight would settle on the front left corner, a popping sound occurred. Not a good sign when you're prepping for an autocross. So, my first thought was to remove the quick disconnect end link from the front bar and retest. Sure enough... no popping. So, sway bar related, most likely.


Notice I didn't say definitely.


Greased up the bushings, removed the end link on one side and greased up the ball sockets, in case they were somehow causing the binding and subsequent popping noise. Turns out, the problem was the lower end link stud that connects to the control arm, on the PASSENGER side. The retaining nut had somehow torqued itself to the point it was preventing the little bit of movement needed.


Weirdest part to me, was that the sound and the symptoms both appeared in the driver side wheel well, yet the issue was actually located on the passenger side. That was a fun one to learn!

Andy Hollis
Andy Hollis
7/5/21 4:20 p.m.

The original picture...which somehow got lost in this online publication.


Pete. (l33t FS)
Pete. (l33t FS) GRM+ Memberand MegaDork
7/5/21 6:41 p.m.

In reply to Andy Hollis :

Oh that's brilliant.

noddaz GRM+ Memberand UberDork
7/5/21 6:43 p.m.

Great idea!

BTW, the first picture is not a CRX.  

Another thought that is actually relevant to the story.

Could some sort of bracket be welded up to test something like this on a bench?  Some sway bars do not come on and off of cars easily.


Keith Tanner
Keith Tanner GRM+ Memberand MegaDork
7/5/21 7:07 p.m.

You'd want a fairly strong bracket, but sure :)

One thing to note: it usually takes a fairly significant change in bar stiffness to change handling since they are not the only source of roll stiffness. A 100% increase is not unknown.

Another thing to note: you don't really need to know the absolute stiffness of the bar if you're keeping the same geometry. You can calculate the change in stiffness by simply calculating the change in stiffness of the material, which is usually just the change in diameter. Remember that stiffness increases with the diameter to the power of 4.

If you're trying to figure out what the wall thickness is of a mysterious hollow bar, you can sometimes see the wall thickness in the flattened end where the bar has been squashed flat for a mounting tab.

frenchyd GRM+ Memberand UltimaDork
7/5/21 8:09 p.m.

In reply to Keith Tanner :

Racers know to build hollow bars and sliding links. The center of a bar does nothing to add stiffness except add weight. The sliding links allow me to adjust the bars to the track in minutes.  After I bend the bar to my shape I take to the local truck spring shop  for heat treating. 
    I use aluminum bushings/ brackets ,  But learned a lesson.  Always  put a radius where the  tube enters and leaves the bushing/bracket. Failure to do so increases the chance of binding  caused by chassis flexing.  Oh and lube the brackets/bushing every race. 
  PS  learn the trick of preloading the swaybars.  You can adjust the suspension to do one thing  on right hand corners and another on the right hand  ones. ( don't do this until you really understand  what and why )   It's always safe to set it up square  and extremely tricky not too. 

Keith Tanner
Keith Tanner GRM+ Memberand MegaDork
7/5/21 9:09 p.m.

The center of a bar does indeed contribute to stiffness, but the outer skin is far more important. Basically, to get the stiffness of a hollow bar you take the number for a solid bar with the same OD and subtract the number for a solid bar with the ID. That 4th power means it's not a significant contribution, but it's also not zero. 

Sliding links are a way of adjusting bar stiffness, yes. But they're harder to calculate on a spaghetti bar. Note that I did specify the same geometry. Life is much simpler when your bar is straight with arms that stick out at 90 degrees, but the need to do things like turn your front wheels can prevent that. Which is basically the point of this article - if you're messing with arm geometry, it can get really tricky to calculate in a lot of cases.

Preloading bars is a way to get handling asymmetry. Whether that's on purpose or by accident is another story. 

I've found that absolute bar stiffness is something people think they need to know but don't know what to do with when you give it to them. That's why we don't publish the numbers for our bars but we'll tell you if you ask. 

frenchyd GRM+ Memberand UltimaDork
7/5/21 9:27 p.m.

Again as a racer I can adjust for those last decimal places by adjusting the sliding links.  What I love is the speed and ease I can change the stuff so if the track should suddenly get a deluge of rain  moments before the race  a few minutes and a wrench I can make the required changes. 
      The art and skill of preloading is one of those skills not to be undertaken without massive experience  and trials  to know what you're getting into.  

Keith Tanner
Keith Tanner GRM+ Memberand MegaDork
7/5/21 11:12 p.m.

I think it's fair to characterize most people here as "racers". Most of us know how to adjust adjustable sway bars. 

The article is not discussing how to easily adjust sways, but to calculate/measure their stiffness. 


frenchyd GRM+ Memberand UltimaDork
7/6/21 7:50 a.m.

In reply to Keith Tanner :

Kieth.  I'm sorry to be on opposite sides from you but we both know that there is no one correct answer. 
     Too many variables. Hot/cold track, dry, wet, damp and drying, damp and getting wetter, new tires/ old tires, greasy track,  sand blown on the track,  etc. 

Then there are different tracks that call for a different answer. 
     Easy adjustability takes some of that away and allows the driver to  use the car rather than drive around a handling problem.  

Keith Tanner
Keith Tanner GRM+ Memberand MegaDork
7/6/21 8:59 a.m.

I don't understand what we're on different sides of.

I'm talking about how to quantify sway bar stiffness, as that's what the article was about. Not about figuring out what the One True Sway Bar might be, but being able to compare the actual stiffness of a couple of different bars.

You're talking about how to adjust sway bars at the track. You're basically having a different conversation and you think I'm disagreeing with you.

frenchyd GRM+ Memberand UltimaDork
7/6/21 9:31 a.m.

In reply to Keith Tanner :

My point is the number doesn't tell you anything except the number. Not how the car will handle turn 5 or the carousel  If it will understeer or oversteer  on corner entry/exit.  How a greasy track will affect it, or any of a dozen variants. 
        Now if you're asking if  one bar is stiffer than another, sure measure it. 

Stampie GRM+ Memberand MegaDork
7/6/21 10:32 a.m.

In reply to frenchyd :

Hmmm yeah ... that's what they were asking sure enough.

burgermeister New Reader
1/22/22 2:05 p.m.

It is possible to calculate non-ideal anti roll bars.  The most significant deviation from ideal is when the location of the bushings is considerably inboard of the end links.  I recently had this issue, and dug out my old engineering books...this won't solve every bar, but it will get many  of them.  The exact shape of the arm "e" is not very critical - the action happens with "a", "f", and "d".  As a bonus, this method also ends up showing where the bar is inefficient.  The formula is the top of the page on the right.


I keep a spreadheet of my car and its Lateral Load Transfer Distribution.  That way I know how significant, or not, of a change I am making.  I seem to like the LLTD to be pretty close to the weight distribution.  It's a SLA front, 4-link rear, happily the geometry is simple enough where roll centers and camber curves are easy to calculate.

Not a racer, just a track day enthusiast.

RustBeltSherpa New Reader
1/24/22 10:26 p.m.

In reply to frenchyd :

"Young people speaking their minds.

Getting so much resistance from behind."

"For What It's Worth" by

Buffalo Springfield

clshore Reader
2/13/22 1:03 p.m.

Could also set up the rod to the link end as a pushrod.
Threaded rod, nut, washer, bushing into the link end 'eye'.
Set up a floor jack with a digital scale between the pushrod and the jack saddle.

As the jack is raised, record deflection distance vs scale force reading.

Bluebayou22 New Reader
2/8/23 1:59 p.m.

I would like to do a test that evaluates the roll center/lowering car theories.  My car GR86 is ready to do the lowering part.  Now just need an OEM delivered GR86 as the "standard" for OEM roll center performance.  We just need a skid pad, slalom and race track.... thats all.  Test, test, test.  Beer.  

Our Preferred Partners