Looking for a good base Camber gain curve for rear IRS

I have decided to make my own unequal length setup for the rear of my E30. I have a pretty good idea on how to control the roll centers, can calculate all the rest of the parameters and so on but... I don't have a target to calculate for.

For a race car - a sedan in the 1 ton range with a TBD final width... how does one determine what a good starting point is for total travel and camber gain across the range? Surely there is a pretty well established set of values that will work without exhaustive testing with my particular chassis - I mean - race teams have been making this sort of thing for decades. Does a typical World Challenge car start with 4" of travel and 2 deg negative across it? 7? 5?

I guess what I'm looking for is a reasonable starting point to work the rest of the values from that is typical of a road race application on an average tin top. I can adjust it for clearances. Should I just be crawling around the paddock at a pro racing venue with a tape measure or is there a "rule of thumb" to reference?

First you need to know how much suspension travel you'll have. A long travel suspension needs a completely different camber curve than does a short travel suspension.

How much body roll will you be getting, that factors into it as well.

carguy123 wrote: First you need to know how much suspension travel you'll have. A long travel suspension needs a completely different camber curve than does a short travel suspension. How much body roll will you be getting, that factors into it as well.

The question I'm asking is how much travel do I need and how much gain to apply. Every book will tell you how to calculate all of the numbers but surely - after decades of racing - there has to be a "4 inches of travel and -1.5 degrees ought to be fine for that application" type rule of thumb. I'm designing from scratch - what should I be shooting for?

Giant Purple Snorklewacker wrote:

carguy123 wrote: First you need to know how much suspension travel you'll have. A long travel suspension needs a completely different camber curve than does a short travel suspension. How much body roll will you be getting, that factors into it as well.

The question I'm asking is how much travel do I need and how much gain to apply. Every book will tell you how to calculate all of the numbers but surely - after decades of racing - there has to be a "4 inches of travel and -1.5 degrees ought to be fine for that application" type rule of thumb. I'm designing from scratch - what should I be shooting for?

If there is a rule of thumb like that then someone share secret because I have never heard of it and I have read Milliken's Race Vehicle Dynamics, all the Carroll Smith books, etc. One thing I can point out is the rear will need less camber compensation then on the front. You are going to want the front and rear to work to together so what I would suggest for a starting point is finding out about as much information about the front as possible such as the roll center movement and the camber curve. Camber curves are great but if the roll center is moving in such a way as to make the car uncontrollable then the best camber curves in the world aren't going to matter. So you are going to want to make sure that the front and rear roll centers are moving in the same directions in roll and also keeping the roll center movement close to the same as the CG movement in bump will result in keeping a fairly constant roll gradient.

I feel like from the answers given so far I'm asking the wrong questions. The front is a bit irrelevant to me - its not that good either - I hoped to adapt it to better match an idealistic rear design. I chose to start in the rear because its already broken and is simpler - no steering curves to deal with - just straight-up wheel control.

I have read all of those books and few more as well. I guess I need to find a successful car in my basic weight/size category and copy it's base dimensions as a starting point. I can alter it to make the roll centers work for my dimensional constraints and adapt it as necessary but if I don't have a place to start... you know - you read all the books but until you suffer thru it once the hard way you have no base from which to make educated guesses. It's like being well read on boxing and then jumping in the ring.

I just find it difficult to believe that if I went to PTG or Flying Lizard or someone they would not have a clue as to what the ballpark of a chassis should be in without modeling the whole thing. I am looking for that experience to say "start with this and work from there" but... maybe that is just the wrong question to ask.

Have you looked around at Corner carvers?

How much different would the suspension curves need to be from a Locost?

Why not start with travel/gain curve of a Miata?

z31maniac wrote: Have you looked around at Corner carvers?

They really don't tolerate noob development like this. I need better questions to go there with or the whole thread will be about how stupid I am.

How much different would the suspension curves need to be from a Locost?

I have not built one yet...

Why not start with travel/gain curve of a Miata?

Because that is a street car that comes with the compromises that go along with that. I am trying to shoot for a pure racing application so while I'm fine with plagiarism I'd rather plagiarize a tube chassis something or other from Grand Am.

1. True
2. Figured the info would be easily available from multiple sources.
3. Just throwing it out there as a starting point?

I ain't an automotive engineer. If anyone is an automotive engineer and wishes to contradict me, please do so. Other type of engineers don't count.

Your camber curve should match the body roll. Remember the important thing is to keep the tire on the ground and on the ground at the optimal angle for grip. This can be different for different tires, but a camber of -1 degree was the optimal grip angle for a race tire published in one of Smith's books.

Also, consider that if you squat with acceleration, that will load up (bump) the suspension too, and if you have built in a lot of body roll and compensated for that with camber gain, that camber gain isn't going to help you on a launch, as the rear suspension compresses down and the camber increases from your static to whatever. You could get less traction for the launch, depending on the curve. Also, ideally, in droop (the inside wheel), it would be great if the camber angle went positive, which would really be keeping it negative with respect to the road surface. I strived for that when I designed my suspension on the Locost. Afterwards, I found a camber curve for the new (at the time) super hot Caterham and that's what they did too. Then, you want to minimize roll center movement and scrub. It's hard to do it all.

I recently did the camber curve for a stock Europa S2 rear suspension, measuring the wheel angle and moving the wheel through it's range with no spring/shock connected. I could post the numbers and one of you Excel (Open Office?) experts could graph it and hotlink the graph if anyone is interested. Basically, the camber curve was a straight line.

I AM and automotive engineer

And I do wish to contradict you

The problem is that I have no idea, so I can't...

I have thought this question many times, but being an emissions guy doesn't really help in suspension design and tuning.

Dr. Hess wrote: Your camber curve should match the body roll. Remember the important thing is to keep the tire on the ground and on the ground at the optimal angle for grip. This can be different for different tires, but a camber of -1 degree was the optimal grip angle for a race tire published in one of Smith's books.

This. I would try to find some tire data if possible but this can be difficult.

Giant Purple Snorklewacker wrote: I feel like from the answers given so far I'm asking the wrong questions. The front is a bit irrelevant to me - its not that good either - I hoped to adapt it to better match an idealistic rear design. I chose to start in the rear because its already broken and is simpler - no steering curves to deal with - just straight-up wheel control.

It is going to be difficult to adapt the front to match the rear without redesigning the front.

You're asking a very open-ended question. You haven't said what the car weighs, where its CG is, how hard it's going to corner/lean, what type of surface it'll be driven on, that sort of stuff.

I'm no Carroll Smith, but I did stay at a....well, I do have some dynamics experience. Anyways.

I would be less concerned about camber gain/loss with respect to wheel travel (degrees per inch), and pay more attention to how it changes in roll (degrees per degree of roll). It's two sides of the same coin, but I find it easier to think in those terms.

Ideally, you'd probably want something as close to -1 degree of camber per degree of roll as possible (for the outside tire's benefit). The actual camber of the tire is a superposition of body roll and suspension camber change. Imagine replacing your shocks with solid rods. If you lean the body 2 degree's to the outside, the tires now have 2 more degrees of positive camber (bad), but if you let the suspension move in such a way that the tires gain -1 deg for each deg of roll, you now have a tire that's still square to the road (good).

In practice though, it's hard to get that much negative camber gain. Usually packaging or some other suspension movement (roll center, lack of travel, toe changes, etc) will limit you. Gotta love tradeoffs. Generally, the closer you can get to -1 the better.

That's part of the reason you dial in a lot of negative static camber. If you have -2 degrees to start with, the body rolls 2 deg, and you only gain -0.5 per deg of roll, then you're still left with -1 degree at full roll. [-2 +2 + (2*-0.5)]. There's also a thing called camber thrust that helps a tire generate force (think of how a bicycle turns), which is why a tire doesn't alway make the most force when it's totally square to the road. Most tires make the most grip with a little bit of negative camber.

When it comes time to redo the front, shoot for just a little bit less then what you did in the rear. That means that the front will loose grip due to positive camber gain a little quicker than the rear, dialing in some nice safe understeer, which you could always tune out later using a larger rear bar if needed.

Good luck, I'd be interested to hear about what trade offs you end up having to make.

kb58 wrote: You're asking a very open-ended question. You haven't said what the car weighs, where its CG is, how hard it's going to corner/lean, what type of surface it'll be driven on, that sort of stuff.

It is a front engined, rwd BMW E30 with front E36 M3 hubs, spindles and corrected geometry to retain E30 wheelbase, lower castor to workable levels and minimize bump steer due to lowering. No other corrections to fix the roll centers or shrink the scrub radius has been done. It will be driven on road racing tracks across north america. It will corner as hard as humanly possible given the tire - probably Hoosier 245/40-17s but all out radial slicks are a possibility if I get a raise :)

It has an M Coupe rear trailing arm suspension right now but assume that I plan to cut the entire structure out and start with a cassette to be tied into the rear of the cage.

2140lbs, wet, 1/2 tank no driver. It is 2335 at 53% front biased with a 50% cross weight with driver (me).

Assuming I can modify the front suspension to whatever the rear requires... and the rear can be built to be fairly adjustable within a certain range what is your ideal rear end?

My first thought is to copy the rough measurements from a Grand Am or SWC car and use that as a starting point and correct it for my target ride height and then correct the front to match.

Based on this, it'll be very stiffly sprung, so it's not going to lean much. That means you can have less camber compensation in order to keep the tires more upright under braking and acceleration. What you need to figure out/guess is just how much it'll lean, squat, and dive, then design back from there. What makes it much tougher is not having tire data (but no one else does either.) Some race tires want more camber than others, so often the best design ends up nothing better than a guess due to lack of tire data. Sorry for the lack of hard answers, but you can see how all sorts of stuff come into play when doing something like this. However, it's not that bad since everyone else is guessing also!

I always just design for a front VSAL of ~140" and a rear VSAL of ~80". Unless you are running some crazy different length A-arms your VSAL length will not change much over your Suspension travel range.

The thinking is thus,

Your camber curve is simply a function of your VSAL at any given time. The shorter the length of VSAL the larger your camber gain will be for a given wheel travel. The actual Camber curve is of low importance because you will only ever get to 1:1 if you have your VSAL = 1/2 your track width. Just know that at lengths shorter than 1/2 track width it's greater than 1 (which would be awesome, but numerous other issues pop up in this situation), and at infinite length it is 0.
In the front, you can use Caster to gain camber while turning the wheels, so you want to tend towards a longer VSAL so under braking you minimize camber gain (pure bump, long VSAL, lower camber gain). During cornering you can count on the steering angle to contribute to the Camber angle through caster. I've used the number ~2-2.2 x track width with good success.
In the rear you have no steering angle to contribute camber so you rely only on the suspensions natural camber gain through suspension travel. For normal cars you need to balance the desire to keep the tire upright for accelration, to your desire to keep it perpendicular to the ground during cornering. Through testing with our FSAE car, we felt a VSAL of ~1.4 x Track Width was optimum. On a 55" track width this puts you in the 80" range.

Once you chose your VSAL, and know your desired RC position, you simply start at the lower ball joint and your suspension designs it self. If you have a fixed upper ball joint, you just need to pick A-arm lengths. The relative lengths will impact your camber curve somewhat, however if I remember right even a 2:1 lower arm to upper arm length difference only caused ~10% change in camber curve over 2" of suspension travel, but tended to impact the RC migration greatly.

This is at least my theory on suspension development, however if you find someones theory you like better please post it up. I like learning about other ideas of suspension theory.

Daniel

Herb Adams has some lengths of front view swing arm lengths he thinks work well. I don't have the book on me right now but it also depends somewhat on the travel. A Formula SAE car is going to have a much shorter swing arm length then a Baja SAE vehicle. Really I would treat the tire as if it is happiest at -1 degree of camber unless you can find data otherwise.

nocones.... thanks for the write-up. Its good food for thought. I'm re-reading Staniforth's Competition Car Suspension book right now as a warm-up to get my head around the all the variables. I love this guy's construction paper mock-up techniques (since I have no Solidworks skills or license). I need to visualize some of this and I have lots of paper and thumb tacks ;)

I am in full research mode right now but I hope to make it an interactive build thread for this winter once I get my brain tuned up.

Ah yes, Stantiforth's "Paper Dolls." I found them helpful for visualizing how it all works out, but not too helpful for modeling what I wanted to do.

If you can get a free CAD software of any kind, you can do a front view sketch and if you constrain everything correctly, you can then push the suspension thru its travel and record the Roll center movement and camber change.

The problem with doing it that way is that you draw up a suspension model in CAD, bump it, see what it did, move your pivots around, bump it, see what it did, etc. It will tell you what your model does. There's a free suspension program that does exactly that. I think it's suspen.exe. What you need, and what isn't out there, is something that will find where you can put your pivots to provide you with the camber, roll center, scrub you want given a particular set of limitations (spindle, frame, etc.) That isn't out here, as far as I've seen, at least not comercially. Even the professional automotive suspension CAD programs don't do that, or didn't when I was doing my Locost. That's why I wrote my own. I have heard that the F1 teams have their own inhouse stuff for it.