Learn me on roll axis inclination?

Anyone knowledgeable about the subject?

Basically....I am not (but need to be ) I know what it is, and that the roll axis should generally rise towards the rear of the car, but I am rather lost beyond that.

I know there is likely no magic formula, and that it will vary from car to car, and with intended use... But anyone have any good advice? recommended reading(internet is HIGHLY perfected,scheduling is telling me to hurry up and know what I am talking about already, and my bank account isn't too interested in new purchases right now...)? Rules of thumb? Things you heard smarter friends say.... Anything?

Vehicle in question is a tube frame open wheel car with a ride high of approximately 2 inches, weight of ~1000lb with a 40/60 weight distribution (front/rear). Track width of 50in and wheel base of 60in. Total chassis length of 102in... Unfortunately no info on CG height as drive train components/packaging is still being worked out, but front roll center is located at ground height with the car sitting still.
The car will be used on an autocross style course, and much less importantly for drag racing.

Formula SAE? I've heard good things about http://www.amazon.com/Race-Car-Vehicle-Dynamics-R146/dp/1560915269/ref=pd_sim_b_1

The roll axis is controlled not by the CG, but rather by the suspension geometry. For an overview without a lot of numbers and mumbo jumbo, the old standby 'How To Make Your Car Handle' will cover the basics.

The short version: design your car with the lower control arms horizontal with the ground. Now the roll center will be controlled by the position of the upper inner control arm pivot. As a general rule, if you raise the inner pivot it will lower the roll center. Yeah, this seems bass ackwards, but again going back to Puhn's book the measurements and diagrams will help explain why. The roll centers change with body roll and suspension position so be careful how short you go with the upper C/A, as the shorter the arm the quicker/bigger the RC change.

Short version, again: the upper C/A should be approx 70% of the length of the lower C/A in most cases. If the suspension is real stiff and has very short travel (like a Formula Ford, etc) the C/A's can be the same length and not move the roll axis too much. Since you say you will run 2" of ground clearance, I'm thinking you will probably wind up using nearly equal length C/A's.

Again as a general rule: the 'tall' end of the roll axis will be the 'loose' end of the car. There's other things which go into this like weight distribution, etc. So if you go up on the roll axis toward the rear, the car could ultimately oversteer. Go the other way, it understeers.

Jensenman wrote: The short version: design your car with the lower control arms horizontal with the ground.

An expansion of that. It's not that your control arms want to be horizontal. It's that the line between the inner and outer pivot points wants to be horizontal. That's an important distinction. My Locost uses a ball joint that has a pivot about 1" above the mounting point, so the control arm actually wants to be off level. There's that pesky difference between the clean geometric diagrams in the books and actual fabricated parts :)

Keith is exactly right. Thanks for clearing up my booboo.

Thanks for the help so far...

Interestingly enough, I have both books mentioned so far. I don't think Phun talks much about roll axis inclination, just roll centers. If anyone has pages of Race Car Vehicle Dynamics to referenced feel free, chapter 17 is amazing at giving a step by step procedure to design the suspension... but I haven't found anything about recommendations for specifics of inclination angles... Just the same higher in the rear.

Close to FSAE, and we are building off an 06 FSAE chassis... but unfortunately we are about double the weight, its for the Formula Hybrid competition... very much the same rule book with an extra 27 pages of modifications.

The lower control arm, at least for front points down(very slightly)... The front roll center is actually located ground level, its what the chassis was designed for and with out making some drastic track width changes it only moves like tenths of an inch. We are mostly leaving the front of the chassis alone, so the only real influence we have there is were we put the outer ends of the control arms. The rear end of the chassis was built around a 600cc engine, which we are not using (limited to a 250cc engine) so we are hacking of everything past the main roll hoop and starting over so this is where I mainly get to set the roll axis inclination...

I know its purely a geometric concern as far as the actual locations of the roll centers go, but the height from roll axis to CG determines how much weight is transferred through t the roll center vs through the suspension... So I threw that in about the CG height before someone else brought it up. I also get how relative lengths of upper and lower control arms effect swing arm length/camber curve...I plan to go back and ensure nothing too wild happens there after we get the basic geometry down

I do understand roll axis inclination to the point that it dictates the rate of weight transfer in transient condition, and why that means the rear should be higher.... I just have no idea how much higher, I guess that was mainly what I was looking for advice on.

Puhn's book does talk some about roll axis inclination, he basically just describes it and says to not let it get too out of whack.

Hmmm. I have no idea of how much higher one end of the roll axis 'should' be than the other. Back to my project (the Abomination): the rear roll center is, IIRC, about 1 1/2" higher than the rear (all my measurements etc are at home). I really wanted it to be level at rest since it will change as the front/rear rise/fall/turn around their roll centers, but unfortunately things like diffs and axles wanted to occupy the same space as suspension/frame parts and to get the control arms longer to allow for changes without huge arm angularity I would have had to basically chop the rear of the frame off and start fresh. Not to mention flare the crap out of the rear fenders. I am lazy.

EDIT: I just thought of something: is there any reason you couldn't make your upper inner pivot adjustable for height so you could experiment? Maybe use shims or something?

You need to get in touch with Mark Ortiz, "mortiz49 at earthlink.net" and get a copy of his "Minding Your Anti's" video tape. I think it's $50.00. It will explain more than a dozen books on the subject. He shows the right way to calculate the roll center, and explains why and how several books show the wrong way. Also I would sign up for his monthly e-mail newsletter they have a lot of info in them. You can also search and find his old newsletters posted in various places. Archives I found

Jensenman wrote: EDIT: I just thought of something: is there any reason you couldn't make your upper inner pivot adjustable for height so you could experiment? Maybe use shims or something?

I was actually thinking about that, I like adjustable stuff I know they make adjustable lower control arms for the 240sx, I suppose I should go take a look back at those and see if it isn't something we could implement... Money is kinda a factor though, as at this point we don't really have any.

Anyway you could let me know either the difference in heights, and horizontal distance between roll centers, or the angle from horizontal?

If nothing else I suppose I could just look at what a bunch of different cars have used, then just pick something reasonable....

I'll have to look into that video. A video would be nice so I could sit the rest of the suspension team down and we could have movie time...

I used Miata uprights or 'knuckles', the LCA's were 11 1/12" center to center and the UCA's were 7 5/8" center to center. I wound up with the upper inner pivot about 1/4" higher than the upper outer. I couldn't go any higher without FUBARing the camber curve completely, it would have gained positive instead of negative camber. Small amounts of change make a big difference when you are working with short CA's.

The Miata upper knuckle pivot point is a good ways inboard of the lower knuckle pivot point (approx 2" IIRC), so that set of measurements probably won't do you any good.

It's been a little while, but why does the lower arm have to be level to the ground? Our FSAE car had both arms declined (we were running a high roll center to eliminate the weight of a ARB), my Midget is going to have the lower arm up also.

The way I did the FSAE suspension was 1. I figured out the lowest farthest out possible location for the Lower ball joint 2. Selected the camber curve I desired for that end of the car (Camber curve is determined by initial Instant center location) 3. Decided on my roll center height. 4. Drew a line from the contact center through my roll center height, determined the IC location along this line, and then drew the line of the lower A-arm through the IC to the LBJ. 5. Figured out how wide I wanted/needed the lower chassis rails to be, and therefore determined the lower inner pivot location. 6. Decided on a KPI/offset and drew a coresponding line out at the tire through the lower ball joint. 7. found the tallest possible point for the upper ball joint along that line (allows the chassis to be taller, and therefore stiffer) 8. Drew a line form the UBJ to the IC, and again decide how wide the chassis needed to be.

For step 8, I finally started to manipulate the suspension in a suspension software to see How the IC length changed and where the RC went in bump and roll. If I didn't like what it did, I tweeked the location of the Upper inner pivot.

For the front suspension I then independently considered Caster and bumpsteer.

For the Midget since I used production uprights in the front and rear, I had to work with a fixed LBJ in the front and fixed upper and Lower in the rear. In the rear this didn't really pose a challenge as I went with a short IC and am just going to build a stiff rear subframe to handle the load, and weight isn't as important as it was on the FSAE car. In the front I am using a strut type bottom of the upright so I have freedom to locate the upper ball joint, and went with a longer IC. I like the Long front, short rear (140+" front, ~80" rear) IC because this works well with what you need the wheels to do. In the front you want a less agressive camber curve in bump/drop so the tires stay upright in braking, I handle the roll camber curve by running more caster. In the back you get NO roll camber help from steering so I use the short IC to achieve my goal.
As to the Origional question, I try to keep the roll axis fairly close to the CG axis per Carrol Smith's recomendation. So this keeps weight jacking, and roll angles fairly constant as the car corners.

Sorry about the long post. Just thought I'd share my thoughts.

Daniel Cummings

Did a quick search and found a dimensioned picture of a miata upright... says 1.6in horizontal and 8.4in vertical... Steering axis inclination of 11 degrees? Is that right? I know that results > theory... but most of my reading has said to keep SAI under 8deg... seems extreme?

The lower CA does not necessarily have to be horizontal. It can be inclined up or down, but as you noted this changes the upper inner CA pivot location to achive a desired RC height. It's also important to consider how this changes the roll center as the suspension moves through its arc. What may start out as a reasonable roll center location can quickly become less than optimal. It's hard to picture, but yes it can happen. It can also fark with the camber curve. Given that the car's ride height will be 2", there won't be much suspension travel (at least not in bump) so maybe this wouldn't matter.

It just makes more sense to me to start with the lower CA (as measured by Keith's method) level or as nearly so as possible and design around that, since in that range of movement roll center changes would be minimized. I would think that's probably why just about every car out there has the suspension designed in that manner. No flame, just stating.

The place to go is.

http://www.auto-ware.com/index.htm

Front end geometery pro is GREAT!

RedS13Coupe wrote: Did a quick search and found a dimensioned picture of a miata upright... says 1.6in horizontal and 8.4in vertical... Steering axis inclination of 11 degrees? Is that right? I know that results > theory... but most of my reading has said to keep SAI under 8deg... seems extreme?

I can't see the pic at work, are you looking at the rear or front upright? Reason I ask, I was referring to the rear upright in my post about dimensions.

ahhh, I believe that pic/those dimensions are of the front.

That's definitely the front.

Hey, Keith: if you have 1.8 front uprights, will 1.6 calipers and rotors fit (to clear 13" wheels)?

All 1990-97 front uprights are identical. The 1999-05 ones will interchange, but the steering arm is about 7mm higher.

For your application, you need new caliper brackets, pads and rotors to move to the 1.6 parts. Keep the same calipers. I think there are a number of 13" wheels that will fit over the 1.8 brake setup though.

Cool. Thanks for the info.

Jensenman wrote: The lower CA does not necessarily have to be horizontal. It can be inclined up or down, but as you noted this changes the upper inner CA pivot location to achive a desired RC height. It's also important to consider how this changes the roll center as the suspension moves through its arc. What may start out as a reasonable roll center location can quickly become less than optimal. It's hard to picture, but yes it can happen. It can also fark with the camber curve. Given that the car's ride height will be 2", there won't be much suspension travel (at least not in bump) so maybe this wouldn't matter. It just makes more sense to me to start with the lower CA (as measured by Keith's method) level or as nearly so as possible and design around that, since in that range of movement roll center changes would be minimized. I would think that's probably why just about every car out there has the suspension designed in that manner. No flame, just stating.

Yeah, I had been planing on setting everything at static ride hight, getting it all worked out, then checking against with the suspension compressed. I have been messing around using solid works to do all the geometry work for me, setting a few points and defining them, then connecting everything and leaving the stuff that up for design undefined so I can drag it all around and see what happens. The upper front control arm is shorter by a few inches, so the instant centers should move inward, increasing the rate of camber change with travel... and I think down... Haven't really drawn it out in compression though so thats just sitting here thinking about it.

What was your reasoning wanting the roll axis level Jensenman?

http://www.performancetrends.com/SuspAnzr.htm

RedS13Coupe wrote:

Jensenman wrote: The lower CA does not necessarily have to be horizontal. It can be inclined up or down, but as you noted this changes the upper inner CA pivot location to achive a desired RC height. It's also important to consider how this changes the roll center as the suspension moves through its arc. What may start out as a reasonable roll center location can quickly become less than optimal. It's hard to picture, but yes it can happen. It can also fark with the camber curve. Given that the car's ride height will be 2", there won't be much suspension travel (at least not in bump) so maybe this wouldn't matter. It just makes more sense to me to start with the lower CA (as measured by Keith's method) level or as nearly so as possible and design around that, since in that range of movement roll center changes would be minimized. I would think that's probably why just about every car out there has the suspension designed in that manner. No flame, just stating.

Yeah, I had been planing on setting everything at static ride hight, getting it all worked out, then checking against with the suspension compressed. I have been messing around using solid works to do all the geometry work for me, setting a few points and defining them, then connecting everything and leaving the stuff that up for design undefined so I can drag it all around and see what happens. The upper front control arm is shorter by a few inches, so the instant centers should move inward, increasing the rate of camber change with travel... and I think down... Haven't really drawn it out in compression though so thats just sitting here thinking about it. What was your reasoning wanting the roll axis level Jensenman?

Since the axis inclination indicates which end may be 'loose' and I was wanting to do a car as near 'neutral' as possible, it seemed that if I kept the roll axis as near horizontal as possible I'd have a much better chance of a neutral handling car. Due to the difference in control arm lengths from front to rear, the roll axis does change more in the rear than the front. The car actually is not bad. It does have ultimate oversteer but with a lot of warning.

jg did an article a while back about the Mustang in which he delved into the infamous Mustang push which turns into snap oversteer, and it appears a lot of that came from a roll axis which started high in the front but low in the back (IIRC) which caused the car to first push. Then the rear suspension 'bind' came into play, raising the rear spring rate and roll center which led to the snap oversteer. Not a very predictable car! Predictability is the key to consistent and fast lap times. The less drama, the faster you go.

FWIW, I used the lessons from that article and from a couple of other places on our LeMons car (also a Fox chassis), with very good results. Everyone on our team had nothing but good things to say about the car's handling, I just wish the drivetrain had been up to the task. Oh, well: there's always next year.

Here's a pic of a Miata rear upright in a Locost. You can see the upper 'offset' of the outer upper C/A mount.

Yeah, higher roll center in the front definitely sounds like it would be no good... from what I have read its mainly a concern as the car is transitioning, since the hight from the axis to the CG dictates how much of the weight transfer acts through the springs and shocks (and is therefore damped) and how much is transfered straight through the suspension linkages (transferring weight immediately) leaving faster weight transfer and less grip on what ever end has higher roll center.

Or at least that is the understanding I have thus far... (even if I did a bad job explaining rolling vs non-rolling moments)

The article I got that from also pointed out that a higher rear roll center would make the car want to yaw towards the outside of the turn, increasing stablitly, while a higher front roll center would do the oposite, making the car want to turn in more, increasing the forces, making the car want to turn in even more, ect ect... I don't really understand that though, it seems opposite to me? If the rear is allowed to roll more.... oh wait never mind, I think I just got it....