Suspension Geometry Discussion - Roll Centers, Rake, and other sweet stuff!

crystalhelix · 07-01-2009, 09:14 PM

Hey all,

All my measurements are from my 2004 STi, prepped for BSP, so it's all stock lateral links and LCA components. Car is sitting on 17" Rota Boost with 245 Azenis on it.

I was researching roll centers and I was going back and forth with BIGSKYWRX in the Brake and Suspension forum and for me I would have more value from the feedback I'd get from Subaru autocrossers. I know ButtDyno has documented his ride heights quite a bit and I am trying to put science behind the mad-ness or at least talk about it. Rake also comes into play with the discussion. It would be nice if anyone has any CG info on the Subies as well.

Firstly - I am working on some simple Solidworks 2D models of the front and rear suspension. Measurements taken from the front of the front LCA and the rear most lateral links.
14" Front Ride Height

13.5" Rear Ride Height

Firstly I worked out the rear, I still think it needs some tweaks but here's a vid of RC and IC migration. 9.56" on the inner pivot is my rear ride height set to 13.5" hub to fender. I have rolled fenders by the way.

YouTube - 2004 Subaru Impreza Front Suspension

YouTube - 2004 Subaru Impreza Rear Suspension

Then I explored the suggested ride heights and this is how I got the RC and ICs to work out. Interesting that the RC's are so close to each other. This is assuming you have camber plates and they are both slammed inward.

Front RC Equation where Ride Height in Inches = X
y = 2.4791x - 31.583

Rear RC Equation where Ride Height in Inches = X
y = 2.1596x - 26.071

Without camber plates slammed I find that you can loose anywhere from 1.2-1.96" of your roll center height depending on your ride height in the front and 0.6-1.4" in the rear dependent on your ride height. With a CG of ~21" this represents up to a 10% increase in distance from the CG to the respective RC.

I am going to try playing around with different ride heights (maybe from ButtDyno's adventures) but if anyone would like to see something specific let me know. I am working on some charts and what not but I want to make sure the models make sense first.

Thanks ahead of time for any feedback, I am hoping to start a good discussion over this. If there's already an awesome thread I didn't find it when I searched. Scattered info was all I found.

Thanks,
Justin

* Registered users of the site do not see these ads.

Mykl · 07-01-2009, 09:29 PM

I found this today...

http://www.neohio-scca.org/comp_clin...namics2007.pdf

I've been putting some thought into this with the somewhat recent popularity of roll center adjustment kits. Perhaps I can be corrected for being wrong, but my understanding is that raising the force application point in relation to the center of gravity has the effect of stiffening roll resistance due to the reduced length of the moment arm.

Seeing that the rear roll center and the front roll center on these cars is actually relatively close is great information for me to see. Because it potentially confirms my thoughts that only raising the roll center of the front, without changing the rear or increasing rear spring/bar stiffness would make the car more prone to understeer, even if the increased rate in which weight is transfered helps the car to change directions more briskly. However, the effect of slammed camber plates on the FAP may very well mean that raising it via a roll center kit could even things out again. But I don't have the ability to model this like you do.

I'm not great at math, so if my intuition on the subject is incorrect, I welcome a correction.

crystalhelix · 07-01-2009, 09:32 PM

I've read the Mitchell stuff but without the program I am helpless and that's beyond me I think,lol. I have Solidworks so I gave it a shot.

I think the big question is how far are the respective roll centers from the center of gravity on the front of the car vs. the rear. Many people have found that they like matched ride heights in some cases. Say 13.5 all around or 13.6 all around. Most stick to a 0.5" ride height difference front to rear.

13.5 Front and Rear

Mykl · 07-01-2009, 09:44 PM

Quote:

Originally Posted by crystalhelix

I've read the Mitchell stuff but without the program I am helpless and that's beyond me I think,lol. I have Solidworks so I gave it a shot.

I think the big question is how far are the respective roll centers from the center of gravity on the front of the car vs. the rear. Many people have found that they like matched ride heights in some cases. Say 13.5 all around or 13.6 all around. Most stick to a 0.5" ride height difference front to rear.

Yeah, that does seem to be the big mystery that needs to be solved before we can go very far with this. The other variable that comes into play that's hard to get around is the amount of cornering force a given tire can generate.

I think as far as ride heights go we have to know what this change has on the length of the moment arm. In my very unscientific "testing" it felt as though lowering the front in relation to the rear increased the car's desire to rotate. Which makes me think that due to the front geometry of the car, the force application point moves down at a faster rate than the center of gravity. Thus having a softening effect on the suspension, making the rear more willing to step out under steady state cornering since it is now relatively more stiff than the front.

*edit* well now I see a picture,

and that confirms my thought that the moment arm up front gets longer the more you lower it. But I think this is already common knowledge.

AndyRoo · 07-01-2009, 09:56 PM

some of the better diagrams i've seen on the subject for STI. good job

As you displayed with your last diagram, the even ride height front and rear means a lower roll center up front in relation the back and a raked roll axis (at 13.5 inches f + r). More oversteer/less understeer would be the result, which is what most Subaru driver's are looking for.

This is I think what ButtDyno noticed with his ride height experiements, and IIRC he noticed handling balance differences with minor ride height changes.

Still, as you know there are advantages to having a roll center that isn't too low up front and spring rates, sway bars, the complete picture should be designed accordingly.

Roll center migration is the "next step" in analysis, and not a fun one. Have you seen stretch's thoughts on slamming the camber plates in vs. running more camber at the lower mount instead?

- andrew

crystalhelix · 07-01-2009, 10:08 PM

on stretch's thoughts, I don't think I saw that specific thread..

I had read one thread talking about the subject but I wasn't getting the concept, is the idea to keep the struts at similar angles or maintaining motion ratios? I know my front struts are leaned into the car almost twice as much as the rear.

I actually think there's a lot to the "minimum front ride height of about XX having to do with your roll stiffness and needing enough room in your geometry to handle 3-4 degrees of roll without your roll center traversing the ground plane and encounter strange weight jacking..

AndyRoo · 07-01-2009, 11:17 PM

Quote:

Originally Posted by crystalhelix

on stretch's thoughts, I don't think I saw that specific thread..

http://www.iwsti.com/forums/gd-suspe...-analysis.html

Quote:

Originally Posted by crystalhelix

I actually think there's a lot to the "minimum front ride height of about XX having to do with your roll stiffness and needing enough room in your geometry to handle 3-4 degrees of roll without your roll center traversing the ground plane and encounter strange weight jacking..

Agreed.

- drew

Mind · 07-01-2009, 11:30 PM

Fantastic diagrams.

It looks like for the two scenarios you had, dropping the front from 14" to 13.5" (or a 0.5" drop) moved the front RC height from 3.12" to 1.89"... so a 0.5" drop in ride height resulted in a 1.23" drop in RC height. Confirms the non-linearity.

I think that stretch inferred some CG data from NHTSA rollover data - every car is rated a metric for likelihood to rollover, and I think this is based primarily on track widths and CG heights. We know the track widths of the car, and if we so if you know the equation of the metric you can guesstimate the cg heights. Time to go searching...

Mind · 07-02-2009, 08:43 AM

Found the CG post - it looks like a stock height impreza has a CG height of ~21 inches:

Quote:

Originally Posted by stretch

...The Static Stability Factor (SSF) listed by safecar.gov for Imprezas (not STI specific, which has a wider track, but this should clue us in on Cg heights):

2006: 1.37
2008: 1.44

SSF = track width / (2 * Cg height)

Thus, if we put in the track width of each year (averaged front and rear):
2006 WRX: 1.37 = 58.4 / (2 * Cg), thus Cg = 21.31
2008 WRX: 1.44 = 58.5 / (2 * Cg), thus Cg = 20.31

Thus, comparing non-STI's, the 2008 chassis has a one inch lower Cg. Obviously the STI's will be better than the standard models due to lower ride heights, but the point is the 2008 chassis is far better in this regard. Reportedly the biggest change is a lower engine mount, but clearly there's more working here than just that.

If the 2008 STI has the same Cg, then its SSF = 60.4 / (2 * 20.31) = 1.48 due to its much wider track- but as I said, the STI's Cg will be lower due to its lower ride height plus any weight saving measures (thinner glass, aluminum body work, etc).

For reference, the 2005 Toyota Camry has a SSF of 1.40. That should put things into perspective. 2005 Camry's may be sofas, but will transfer less weight around the car when pushed than a 2006 WRX.

Original post:
http://www.iwsti.com/forums/1769917-post13.html

Mykl · 07-02-2009, 11:15 AM

Here's a fun question that I can't figure out the answer to...

Since the length of the moment arm has an effect on the roll stiffness of the front suspension... is there any way to convert changes in length to chanes in effective spring rate?

Basically what I'm asking is... for every inch you decrease it's length, how much spring rate do you need to lose to maintain the same exact roll resistance?

The idea is this... if instead of running forward rake, you raise the front of the car to bring the roll center up to roughly even with the rear.... or if you're not bound to autocross rules you use roll center adjustment ball joints.

By doing this you can run less bar up front and maintain the same roll stiffness you had with a bigger bar and forward rake.

crystalhelix · 07-02-2009, 11:53 AM

Mind...that's good info, I will try to add that to my models just so we can see how it works out.

Mykl - It would be nice to relate optimal roll stiffness and moment arms, I am looking at natural frequency relations as well...

Quote:

Originally Posted by roo

It looks like for the two scenarios you had, dropping the front from 14" to 13.5" (or a 0.5" drop) moved the front RC height from 3.12" to 1.89"... so a 0.5" drop in ride height resulted in a 1.23" drop in RC height. Confirms the non-linearity

I am working on this, I have a excel spreadsheet that I have input all the data into and plotted the curve. There is a relationship between the ride height and the static RC height so that you can just use a simple equation to determine your RC based on ride height. I may even be able to come up with a fudge factor for how much of the camber plate a person is using.

mccanixx · 07-02-2009, 12:24 PM

I feel like towlie from southpark.

"I have no idea what's goin' on."

I'm trying to understand........

CamaroFS34 · 07-02-2009, 12:27 PM

Quote:

Originally Posted by mccanixx

I feel like towlie from southpark.

"I have no idea what's goin' on."

I'm trying to understand........

You're not the only one.

Splash · 07-02-2009, 12:33 PM

The whole using the adjustment at the knuckle more than the plates thing is an attempt to keep more of the camber you dial in. The SAI and included angles are less when done that way. The larger the included angle is, the more camber you lose as you turn the wheel since the caster is never as high as the included angle. If it wasn't for the jacking effect, I'd say there is no such thing as too much caster due to that.

Mine is set so I'm as close to the strut as possible with the bolts, then I set the plates to the target setting (I could get -5 with it all maxed). It's pretty much the best I can do there. I also run similar heights front/rear, and right now I think the front is maybe a tic lower than even, but that's just due to my rear helper springs being a tiny bit too short.

I'm not too worried about the lowness of the front and the strut's camber curve because, at 600lb/in (+ the bar if we are in a corner), I doubt it moves enough to be that big of an issue. I spent the past winter converting from roll stiffness via bars to roll stiffness via springs...

crystalhelix · 07-02-2009, 01:15 PM

mccanixx and Camaro FS34 - I actually blame you two for making me thing about these things, lol, I was thinking damn...those ESP cars are LOW and FAST...how can it be that my car has the suck, it turns out my ride heights front to rear were ****ed on top of other issues..

crystalhelix · 07-02-2009, 01:17 PM

Quote:

Originally Posted by Splash

The whole using the adjustment at the knuckle more than the plates thing is an attempt to keep more of the camber you dial in. The SAI and included angles are less when done that way. The larger the included angle is, the more camber you lose as you turn the wheel since the caster is never as high as the included angle. If it wasn't for the jacking effect, I'd say there is no such thing as too much caster due to that.

I'd agree to a point but I'd also say it would limit the static caster you would run, which is good and bad, it all depends on how much you are turning the wheel to generate the camber needed. I'd rather have most of it there all the time since I spend most of my time turning in auto cross.

mccanixx · 07-02-2009, 01:42 PM

Quote:

Originally Posted by crystalhelix

mccanixx and Camaro FS34 - I actually blame you two for making me thing about these things, lol, I was thinking damn...those ESP cars are LOW and FAST...how can it be that my car has the suck, it turns out my ride heights front to rear were ****ed on top of other issues..

Damn you, don't ever accuse me of making people think.

I can only tell you that all my car's settings are a 8. It seems best there.

Actually I don't know if there's alot of difference between mine (wrx) and yours (sti) in regards to width and geometry. But I wondered what mine modeled like, my ride heights are currently 13.25 rear and 13.125 front.

I guess these models also don't account for suspension bushings and the like, not that it'd drastically change it.

blueVTsubi · 07-02-2009, 01:45 PM

this stuff is the amazing, finally a tread about something important. instead of my car looks more fly then yours bs that usually comes out of "hella flush" diarea mouthed ricers.

crystalhelix · 07-02-2009, 01:50 PM

Quote:

Originally Posted by blueVTsubi

this stuff is the amazing, finally a tread about something important. instead of my car looks more fly then yours bs that usually comes out of "hella flush" diarea mouthed ricers.

ROFL - that was my goal.

Greg - I will post a image of your ride heights with your permission when I get home, being at work sucks...I'd much rather not be working..

mccanixx · 07-02-2009, 01:56 PM

Quote:

Originally Posted by crystalhelix

Greg - I will post a image of your ride heights with your permission when I get home, being at work sucks...I'd much rather not be working..

Yep yep my permissions ye have.

Work do sucketh so. But hey I assume most of us are looking at a 3-dayer

crystalhelix · 07-02-2009, 03:38 PM

Quote:

Originally Posted by mccanixx

Yep yep my permissions ye have.

Work do sucketh so. But hey I assume most of us are looking at a 3-dayer

Here's a close representation. I need to work out a few things so I can set it closer to your actual numbers.

Your front RC is ~1.6 inches below your rear RC.

crystalhelix · 07-02-2009, 03:50 PM

Here's a cool image, did a simple extrusion of the geometry triangles.

boost junkie · 07-02-2009, 03:56 PM

This is why having a separate auto-x section is stupid...a great discussion like this could be going on and most of us would never even know about it.

crystalhelix · 07-02-2009, 03:56 PM

Front RC Equation where Ride Height in Inches = X
y = 2.4791x - 31.583

Rear RC Equation where Ride Height in Inches = X
y = 2.1596x - 26.071

So much info...I don't know what to do....must leave computer..or brain will explode.

crystalhelix · 07-02-2009, 03:56 PM

Quote:

Originally Posted by boost junkie

This is why having a separate auto-x section is stupid...a great discussion like this could be going on and most of us would never even know about it.

Sorry I am an auto crosser, lol. Welcome!

07-02-2009, 12:33 PM	#14
Splash Scooby Specialist Member#: 55936 Join Date: Feb 2004 Chapter/Region: MWSOC Location: Columbus, OH Vehicle: 04 WRX STi PSM	The whole using the adjustment at the knuckle more than the plates thing is an attempt to keep more of the camber you dial in. The SAI and included angles are less when done that way. The larger the included angle is, the more camber you lose as you turn the wheel since the caster is never as high as the included angle. If it wasn't for the jacking effect, I'd say there is no such thing as too much caster due to that. Mine is set so I'm as close to the strut as possible with the bolts, then I set the plates to the target setting (I could get -5 with it all maxed). It's pretty much the best I can do there. I also run similar heights front/rear, and right now I think the front is maybe a tic lower than even, but that's just due to my rear helper springs being a tiny bit too short. I'm not too worried about the lowness of the front and the strut's camber curve because, at 600lb/in (+ the bar if we are in a corner), I doubt it moves enough to be that big of an issue. I spent the past winter converting from roll stiffness via bars to roll stiffness via springs... Last edited by Splash; 07-02-2009 at 12:42 PM.

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
1G DSM Head and other DSM stuff!!! Also Civic Stuff!!!	KassanovaSTi	Non-Subaru Stuff	0	06-13-2007 10:57 PM
-=C=- discusses his OT habits and other crap over email as Katrina raged...	-=C=-	Off-Topic	5	11-03-2005 02:28 PM
Suspension questions about, roll centers, and pollor movment of inertia?	wrxsubaru	Brakes, Steering & Suspension	3	02-13-2005 12:54 AM
Roll Center, questions and answers	Arnie	Brakes, Steering & Suspension	5	12-21-2004 03:26 PM
Hood Scoop and Vents backing plates and other RS stuff	2.5 Blue Roo	Private 'For Sale' Classifieds	27	07-10-2004 05:14 PM

07-01-2009, 09:29 PM	#2
Mykl Scooby Specialist Member#: 14311 Join Date: Jan 2002 Vehicle: 2010 GTI White	I found this today... http://www.neohio-scca.org/comp_clin...namics2007.pdf I've been putting some thought into this with the somewhat recent popularity of roll center adjustment kits. Perhaps I can be corrected for being wrong, but my understanding is that raising the force application point in relation to the center of gravity has the effect of stiffening roll resistance due to the reduced length of the moment arm. Seeing that the rear roll center and the front roll center on these cars is actually relatively close is great information for me to see. Because it potentially confirms my thoughts that only raising the roll center of the front, without changing the rear or increasing rear spring/bar stiffness would make the car more prone to understeer, even if the increased rate in which weight is transfered helps the car to change directions more briskly. However, the effect of slammed camber plates on the FAP may very well mean that raising it via a roll center kit could even things out again. But I don't have the ability to model this like you do. I'm not great at math, so if my intuition on the subject is incorrect, I welcome a correction.

07-01-2009, 09:32 PM	#3
crystalhelix Scooby Guru Member#: 35313 Join Date: Apr 2003 Chapter/Region: MAIC Location: SpryRacing.com AgileAuto.com Vehicle: Spry Forester Silver Spark	I've read the Mitchell stuff but without the program I am helpless and that's beyond me I think,lol. I have Solidworks so I gave it a shot. I think the big question is how far are the respective roll centers from the center of gravity on the front of the car vs. the rear. Many people have found that they like matched ride heights in some cases. Say 13.5 all around or 13.6 all around. Most stick to a 0.5" ride height difference front to rear. 13.5 Front and Rear

07-01-2009, 09:56 PM	#5
AndyRoo Scooby Specialist Member#: 20952 Join Date: Jul 2002 Chapter/Region: MAIC Location: Baltimore, MD Vehicle: RCE Mini Cooper S BRG LOL BRZ	some of the better diagrams i've seen on the subject for STI. good job As you displayed with your last diagram, the even ride height front and rear means a lower roll center up front in relation the back and a raked roll axis (at 13.5 inches f + r). More oversteer/less understeer would be the result, which is what most Subaru driver's are looking for. This is I think what ButtDyno noticed with his ride height experiements, and IIRC he noticed handling balance differences with minor ride height changes. Still, as you know there are advantages to having a roll center that isn't too low up front and spring rates, sway bars, the complete picture should be designed accordingly. Roll center migration is the "next step" in analysis, and not a fun one. Have you seen stretch's thoughts on slamming the camber plates in vs. running more camber at the lower mount instead? - andrew

07-01-2009, 10:08 PM	#6
crystalhelix Scooby Guru Member#: 35313 Join Date: Apr 2003 Chapter/Region: MAIC Location: SpryRacing.com AgileAuto.com Vehicle: Spry Forester Silver Spark	on stretch's thoughts, I don't think I saw that specific thread.. I had read one thread talking about the subject but I wasn't getting the concept, is the idea to keep the struts at similar angles or maintaining motion ratios? I know my front struts are leaned into the car almost twice as much as the rear. I actually think there's a lot to the "minimum front ride height of about XX having to do with your roll stiffness and needing enough room in your geometry to handle 3-4 degrees of roll without your roll center traversing the ground plane and encounter strange weight jacking..

07-01-2009, 11:30 PM	#8
Mind Scooby Specialist Member#: 109216 Join Date: Mar 2006 Chapter/Region: MAIC Location: PA Vehicle: 08 STI DGM 76-77 Lancia Scorpion(s)	Fantastic diagrams. It looks like for the two scenarios you had, dropping the front from 14" to 13.5" (or a 0.5" drop) moved the front RC height from 3.12" to 1.89"... so a 0.5" drop in ride height resulted in a 1.23" drop in RC height. Confirms the non-linearity. I think that stretch inferred some CG data from NHTSA rollover data - every car is rated a metric for likelihood to rollover, and I think this is based primarily on track widths and CG heights. We know the track widths of the car, and if we so if you know the equation of the metric you can guesstimate the cg heights. Time to go searching...

07-02-2009, 11:15 AM	#10
Mykl Scooby Specialist Member#: 14311 Join Date: Jan 2002 Vehicle: 2010 GTI White	Here's a fun question that I can't figure out the answer to... Since the length of the moment arm has an effect on the roll stiffness of the front suspension... is there any way to convert changes in length to chanes in effective spring rate? Basically what I'm asking is... for every inch you decrease it's length, how much spring rate do you need to lose to maintain the same exact roll resistance? The idea is this... if instead of running forward rake, you raise the front of the car to bring the roll center up to roughly even with the rear.... or if you're not bound to autocross rules you use roll center adjustment ball joints. By doing this you can run less bar up front and maintain the same roll stiffness you had with a bigger bar and forward rake.

07-02-2009, 12:24 PM	#12
mccanixx Scooby Specialist Member#: 60566 Join Date: Apr 2004 Chapter/Region: SCIC Location: San Diego, Ca. Vehicle: 2002 WRX Unicorn 2001 Legacy L 2005 XT 5sp	I feel like towlie from southpark. "I have no idea what's goin' on." I'm trying to understand........

07-02-2009, 01:15 PM	#15
crystalhelix Scooby Guru Member#: 35313 Join Date: Apr 2003 Chapter/Region: MAIC Location: SpryRacing.com AgileAuto.com Vehicle: Spry Forester Silver Spark	mccanixx and Camaro FS34 - I actually blame you two for making me thing about these things, lol, I was thinking damn...those ESP cars are LOW and FAST...how can it be that my car has the suck, it turns out my ride heights front to rear were ****ed on top of other issues..

07-02-2009, 01:45 PM	#18
blueVTsubi Scooby Newbie Member#: 181891 Join Date: Jun 2008 Chapter/Region: NESIC Location: Subiville USA Vehicle: 2006 Impreza WRX srr	this stuff is the amazing, finally a tread about something important. instead of my car looks more fly then yours bs that usually comes out of "hella flush" diarea mouthed ricers.

07-02-2009, 03:50 PM	#22
crystalhelix Scooby Guru Member#: 35313 Join Date: Apr 2003 Chapter/Region: MAIC Location: SpryRacing.com AgileAuto.com Vehicle: Spry Forester Silver Spark	Here's a cool image, did a simple extrusion of the geometry triangles.

07-02-2009, 03:56 PM	#23
boost junkie Top Scoob 009 Member#: 68273 Join Date: Aug 2004 Chapter/Region: MAIC Location: Silver Spring, MD Vehicle: 2012 WRX GD Race Car in progress	This is why having a separate auto-x section is stupid...a great discussion like this could be going on and most of us would never even know about it.