Car Physics: Suspension setups

Camber, it is such an interesting term, thrown about in conversation willy nilly but what EXACTLY is it a product of? For example, if you have an SLA setup with the upper and lower arms actually the same length and the vertical distance between the inner and outer ball-joints the same. This will give a KPI of 0 degrees and assuming that the inner ball-joints are fixed at the same X offset from the CG (assuming that +Z points to the front of the car) this will give you 0 degrees camber. If we then wanted to add some camber we would either have to change the length of one of the arms or change the vertical distance between either the inner or outer ball-joints. Right? Take another example, if you had a more traditional SLA setup, long lower arm, short upper arm, upper arm has a larger X offset from CG, and about 5 degrees of KPI. Surely the camber for this system could possibly be different for when the tires are touching the ground as when the car is up on jacks? Also, wouldn't changing the spring rates or dampers change the camber? And now for the question, is camber merely a variable used to represent the angle between the wheel and the contact surface for a stationary vehicle? Stating that the front wheels on a vehicle have say 5 degrees of camber is only a valid statement when the vehicle is not in motion? I really need to get this straight in my head before I code it all up. TIA, Colin Reed ----------------------------------------------------- My homepage : http://www.cersoft.com -----------------------------------------------------

On 9 Jan 2002 16:08:18 -0800, [email hidden] (Colin Reed) wrote: ... >If we then wanted to add some camber we would either have to change >the length of one of the arms or change the vertical distance between >either the inner or outer ball-joints. Right? Camber is quite dependent on I think most everything you change in the suspension links. Just like I read about bump steer for example, which is the change in toe-in as a function of suspension deflection. It's how you attach your wheel to the chassis. (too many parameters for me at times) >Take another example, if you had a more traditional SLA setup, long >lower arm, short upper arm, upper arm has a larger X offset from CG, >and about 5 degrees of KPI. Surely the camber for this system could >possibly be different for when the tires are touching the ground as >when the car is up on jacks? Right. There is something like 'camber change per cm deflection', so if you're in bump, the camber will/can change, depending on how the links are done. What I intend to do is define things like 'camber change per m deflection' instead of really going through all the links and constraint force stuff (which Gregor for example is doing). Although ODE (www.q12.org for those interested) might make things more possible, I don't think it will do much good for performance, and I still doubt that the links will do more than the resulting parameters (camber change/m) can simulate. It depends on your input data; do you know the rod lengths, the angles and such? Then just entering the links directly might be more suitable. But I generally have no knowledge about these things, so I'll use the other, resulting parameters. > Also, wouldn't changing the spring rates >or dampers change the camber? Dampers only change timing, so they only change the camber change per time (which is implicitly calculated anyway in most sims). But a spring rate that would cause a different range of deflection, that would case a different range of camber (notice I leave out steering effects here, which are also there ofcourse, just like camber changes, as if camber & steering weren't related, lol). >And now for the question, is camber merely a variable used to >represent the angle between the wheel and the contact surface for a >stationary vehicle? Right. So it changes according to the track, the wheel and suspension deflection. > Stating that the front wheels on a vehicle have >say 5 degrees of camber is only a valid statement when the vehicle is >not in motion? Right (or at least I'd think that's the only workable definition when you're standing in a pits). And that's also on a flat surface. Ruud van Gaal Free car sim : http://www.marketgraph.nl/gallery/racer/ Pencil art : http://www.marketgraph.nl/gallery/

Ruud van Gaal <[email hidden]> wrote in message news:[email hidden]... > Right. There is something like 'camber change per cm deflection', so > if you're in bump, the camber will/can change, depending on how the > links are done. > What I intend to do is define things like 'camber change per m > deflection' instead of really going through all the links and > constraint force stuff (which Gregor for example is doing). Although That's the way I've seen it done. The handling model I played with for my FYP had 'static camber' and 'camber gradient', as well as 'static toe' and 'toe gradient' - units as you're suggesting. You can probably assume linearity for road/race cars; if you're going off-road, you might want to consider at least a quadratic approximation or even something more accurate. Depends how camber- -sensitive your tyres are, I suppose ;-) Jonny

A Q from a non-physics guy; I understand how camber changes with suspension deflection, but that's relative to the tub/frame, right? If you're modeling in this detail, do you not then have to also look at roll to determine the actual angle of the tire to the track surface at a given moment? Or is there not enough camber variation from roll vs. suspension deflection to warrant doing this? TIA, SB "Ruud van Gaal" <[email hidden]> wrote in message news:[email hidden]...

Quoted text (62 lines)

> On 9 Jan 2002 16:08:18 -0800, [email hidden] (Colin Reed) wrote: > > ... > >If we then wanted to add some camber we would either have to change > >the length of one of the arms or change the vertical distance between > >either the inner or outer ball-joints. Right? > > Camber is quite dependent on I think most everything you change in the > suspension links. > Just like I read about bump steer for example, which is the change in > toe-in as a function of suspension deflection. It's how you attach > your wheel to the chassis. (too many parameters for me at times) > > >Take another example, if you had a more traditional SLA setup, long > >lower arm, short upper arm, upper arm has a larger X offset from CG, > >and about 5 degrees of KPI. Surely the camber for this system could > >possibly be different for when the tires are touching the ground as > >when the car is up on jacks? > > Right. There is something like 'camber change per cm deflection', so > if you're in bump, the camber will/can change, depending on how the > links are done. > What I intend to do is define things like 'camber change per m > deflection' instead of really going through all the links and > constraint force stuff (which Gregor for example is doing). Although > ODE (www.q12.org for those interested) might make things more > possible, I don't think it will do much good for performance, and I > still doubt that the links will do more than the resulting parameters > (camber change/m) can simulate. > It depends on your input data; do you know the rod lengths, the angles > and such? Then just entering the links directly might be more > suitable. But I generally have no knowledge about these things, so > I'll use the other, resulting parameters. > > > Also, wouldn't changing the spring rates > >or dampers change the camber? > > Dampers only change timing, so they only change the camber change per > time (which is implicitly calculated anyway in most sims). > But a spring rate that would cause a different range of deflection, > that would case a different range of camber (notice I leave out > steering effects here, which are also there ofcourse, just like camber > changes, as if camber & steering weren't related, lol). > > >And now for the question, is camber merely a variable used to > >represent the angle between the wheel and the contact surface for a > >stationary vehicle? > > Right. So it changes according to the track, the wheel and suspension > deflection. > > > Stating that the front wheels on a vehicle have > >say 5 degrees of camber is only a valid statement when the vehicle is > >not in motion? > > Right (or at least I'd think that's the only workable definition when > you're standing in a pits). And that's also on a flat surface. > > > Ruud van Gaal > Free car sim : http://www.marketgraph.nl/gallery/racer/ > Pencil art : http://www.marketgraph.nl/gallery/

On Fri, 11 Jan 2002 14:57:24 GMT, "Steve Blankenship" <[email hidden]> wrote: >A Q from a non-physics guy; I understand how camber changes with suspension deflection, but that's relative to the tub/frame, right? >If you're modeling in this detail, do you not then have to also look at roll to determine the actual angle of the tire to the track >surface at a given moment? Roll and suspension deflection are often the same thing. It all boils down to getting the roll angle of the tire (which you can see as camber) relative to the *track surface*. So that means including: - rest camber angle - orientation of the car chassis - orientation of the track surface. And then do some angle stuff (haven't investigated this in detail yet, so no formulae). > Or is there not enough camber variation from roll vs. suspension deflection to warrant doing this? As a body in a 10 degree angle roll has 10 degree 'camber' with respect to the track, it certainly does make a difference. If you're with 2 wheels of the ground, the wheels touching the ground will have an added 10 degrees of camber, which is a lot ofcourse. At that point though, you also start to wonder, 'hmm, shouldn't I have 3 points at which I look at the tire-track collisions?', since one end of the tire is probably in the air, the middle is so-so, and the outer part firmly touches the ground. Nice for those temperature calculations. </rambling> Ruud van Gaal Free car sim : http://www.marketgraph.nl/gallery/racer/ Pencil art : http://www.marketgraph.nl/gallery/

"Ruud van Gaal" <[email hidden]> wrote... > Roll and suspension deflection are often the same thing. DUH! ;-) Guess that makes sense, even to ME; hard to get roll without deflection, assuming all your wheels are on the track.

Quoted text (9 lines)

> As a body in a 10 degree angle roll has 10 degree 'camber' with > respect to the track, it certainly does make a difference. If you're > with 2 wheels of the ground, the wheels touching the ground will have > an added 10 degrees of camber, which is a lot ofcourse. > At that point though, you also start to wonder, 'hmm, shouldn't I have > 3 points at which I look at the tire-track collisions?', since one end > of the tire is probably in the air, the middle is so-so, and the outer > part firmly touches the ground. > Nice for those temperature calculations.

Interesting (read: baffling to me). Is anyone modeling tire casing/contact patch deformation currently, or will that not get you anything useful? Thks, SB

On Fri, 11 Jan 2002 14:40:51 -0000, "Jonny Hodgson" <[email hidden]> wrote:

Quoted text (13 lines)

> >Ruud van Gaal <[email hidden]> wrote in message news:[email hidden]... > >> Right. There is something like 'camber change per cm deflection', so >> if you're in bump, the camber will/can change, depending on how the >> links are done. >> What I intend to do is define things like 'camber change per m >> deflection' instead of really going through all the links and >> constraint force stuff (which Gregor for example is doing). Although > >That's the way I've seen it done. The handling model I played with >for my FYP had 'static camber' and 'camber gradient', as well as >'static toe' and 'toe gradient' - units as you're suggesting.

It seems a lot faster than going through restrained links. >You can probably assume linearity for road/race cars; if you're >going off-road, you might want to consider at least a quadratic >approximation or even something more accurate. Good to know, thanks for that piece of info. Ruud van Gaal Free car sim : http://www.marketgraph.nl/gallery/racer/ Pencil art : http://www.marketgraph.nl/gallery/

Ruud van Gaal <[email hidden]> wrote in message news:[email hidden]...

Quoted text (8 lines)

> On Fri, 11 Jan 2002 14:40:51 -0000, "Jonny Hodgson" > <[email hidden]> wrote: > >You can probably assume linearity for road/race cars; if you're > >going off-road, you might want to consider at least a quadratic > >approximation or even something more accurate. > > Good to know, thanks for that piece of info.

Do bear in mind that unless I cite actual evidence / experience, I'm only talking in approximate terms so don't take everything I type as gospel ;-) Someone like a Milliken would be a useful source of confirmation, for instance! Jonny

It seems my reader missed the original post by Collin.. Could someone re-post it please? I think I've got answers to all this, been at suspension setup stuff for awhile now.. Todd Wasson --- Performance Simulations Drag Racing and Top Speed Prediction Software http://PerformanceSimulations.Com My little car sim screenshots: http://performancesimulations.com/scnshot4.htm

"J. Todd Wasson" <[email hidden]> wrote in message news:[email hidden]... >> It seems my reader missed the original post by Collin.. Could someone re-post it please? I think I've got answers to allthis, been at suspension setup stuff for awhile now.. << Here ya go - SB ------------------------- Camber, it is such an interesting term, thrown about in conversation willy nilly but what EXACTLY is it a product of? For example, if you have an SLA setup with the upper and lower arms actually the same length and the vertical distance between the inner and outer ball-joints the same. This will give a KPI of 0 degrees and assuming that the inner ball-joints are fixed at the same X offset from the CG (assuming that +Z points to the front of the car) this will give you 0 degrees camber. If we then wanted to add some camber we would either have to change the length of one of the arms or change the vertical distance between either the inner or outer ball-joints. Right? Take another example, if you had a more traditional SLA setup, long lower arm, short upper arm, upper arm has a larger X offset from CG, and about 5 degrees of KPI. Surely the camber for this system could possibly be different for when the tires are touching the ground as when the car is up on jacks? Also, wouldn't changing the spring rates or dampers change the camber? And now for the question, is camber merely a variable used to represent the angle between the wheel and the contact surface for a stationary vehicle? Stating that the front wheels on a vehicle have say 5 degrees of camber is only a valid statement when the vehicle is not in motion? I really need to get this straight in my head before I code it all up. TIA, Colin Reed ----------------------------------------------------- My homepage : http://www.cersoft.com -----------------------------------------------------

>Here ya go - SB Thanks for reposting this, Steve :0) >Camber, it is such an interesting term, thrown about in conversation >willy nilly but what EXACTLY is it a product of? > Everything! lol Especially the phase of the moon...

Quoted text (11 lines)

>For example, if you have an SLA setup with the upper and lower arms >actually the same length and the vertical distance between the inner >and outer ball-joints the same. This will give a KPI of 0 degrees and >assuming that the inner ball-joints are fixed at the same X offset >from the CG (assuming that +Z points to the front of the car) this >will give you 0 degrees camber. >If we then wanted to add some camber we would either have to change >the length of one of the arms or change the vertical distance between >either the inner or outer ball-joints. Right? >

Right. I've got two other adjustments in my suspension system in addition though that are important IMO. We're doing a little radio controlled car sim. The VectorNT chassis, for example, has two little screws in each front wheel that can be turned in and out to adjust camber and track width. So I've also got two additional "arms" that stay perpendicular to the mounts at the hub carrier/wheel assembly that can have their lengths adjusted as well. This is good to do because you can easily allow a user to add camber plates or other devices to their sim car if they want, which will change the track width slightly in addition to changing camber.

Quoted text (7 lines)

>Take another example, if you had a more traditional SLA setup, long >lower arm, short upper arm, upper arm has a larger X offset from CG, >and about 5 degrees of KPI. Surely the camber for this system could >possibly be different for when the tires are touching the ground as >when the car is up on jacks? Also, wouldn't changing the spring rates >or dampers change the camber? >

Yes, the camber will most likely be different in those two situations. In our sim, we're allowing the users to setup the cars much like they do with the real cars. They can set "elevated" camber, which is the camber when the car is on jacks, as well as setting "driving" or "resting" camber, which is the camber when it's not up on jacks (which is what all the current driving sims probably are approximating). "Elevated" camber can be set by changing all the arm lengths and positions and what-not. "Driving" camber can be set the same way as well, but changing spring rates, spring preload, and spring rest length alter the driving camber even further.. Actually, changing tire width or diameter alters the camber a little too. Changing dampers won't change the "driving/resting" camber though... In short, in our little sim, the user doesn't just adjust camber by clicking a button, they've got about 10 different ways to adjust it (2 for the tires, so really 8 suspension adjustments effect camber), and they'll all have different effects on the "dynamic" camber, or camber while driving around... And we all think GPL is hard to set up!! lol >And now for the question, is camber merely a variable used to >represent the angle between the wheel and the contact surface for a >stationary vehicle? Stating that the front wheels on a vehicle have >say 5 degrees of camber is only a valid statement when the vehicle is >not in motion? "Yes" to both of these.. >I really need to get this straight in my head before I code it all up. Sounds like you've got the idea just fine. Todd Wasson --- Performance Simulations Drag Racing and Top Speed Prediction Software http://PerformanceSimulations.Com My little car sim screenshots: http://performancesimulations.com/scnshot4.htm

"J. Todd Wasson" wrote: > > >If we then wanted to add some camber we would either have to change > >the length of one of the arms or change the vertical distance between > >either the inner or outer ball-joints. Right? > > Right... Not necessarily. With karts, the front spindles (axles) are made at an angle to the kingpin. Camber is built in even if the lateral KPI is zero. Camber is usually adustable on a kart via Heim joints, or a "pill" that adjusts lateral KPI. [Can you tell I started racing karts this year? :) ] I'm not totally familiar with the actual form of a suspended car's front end, but camber adjustment is surely possible without changing out a swingarm. -- Pat Dotson

Quoted text (9 lines)

>Interesting (read: baffling to me). Is anyone modeling tire casing/contact >patch deformation currently, or will that not get you >anything useful? > >Thks, > >SB > >

I've done this, but it's not incorporated into my sim. It's really slow 'cause there are so many calcs.. If you've got good Pacejka constants for the tires you want, that's better anyway probably. Todd Wasson --- Performance Simulations Drag Racing and Top Speed Prediction Software http://PerformanceSimulations.Com My little car sim screenshots: http://performancesimulations.com/scnshot4.htm

Camber doesn't mean much if the wheel isn't attached to a car, so tire test machines use their own axis system (SAE and ISO are different, both are commonly used). Afaik, definitions aren't standard (yet) but we use the following to help keep this straight: Camber angle is measured between wheel and chassis (positive = top of wheels tilted away from car) Inclination is measured between wheel and ground - the tire axis system (SAE positive = lean to the right, looking in the direction of travel) Confused yet? See Race Car Vehicle Dynamics pp61-63 and Chapter 4. -- Doug Milliken http://www.millikenresearch.com (hint--under Books is purchase info) On Fri, 11 Jan 2002, Ruud van Gaal wrote:

Quoted text (38 lines)

> On Fri, 11 Jan 2002 14:57:24 GMT, "Steve Blankenship" > <[email hidden]> wrote: > > >A Q from a non-physics guy; I understand how camber changes with suspension deflection, but that's relative to the tub/frame, right? > >If you're modeling in this detail, do you not then have to also look at roll to determine the actual angle of the tire to the track > >surface at a given moment? > > Roll and suspension deflection are often the same thing. > It all boils down to getting the roll angle of the tire (which you can > see as camber) relative to the *track surface*. So that means > including: > - rest camber angle > - orientation of the car chassis > - orientation of the track surface. > > And then do some angle stuff (haven't investigated this in detail yet, > so no formulae). > > > Or is there not enough camber variation from roll vs. suspension deflection to warrant doing this? > > As a body in a 10 degree angle roll has 10 degree 'camber' with > respect to the track, it certainly does make a difference. If you're > with 2 wheels of the ground, the wheels touching the ground will have > an added 10 degrees of camber, which is a lot ofcourse. > At that point though, you also start to wonder, 'hmm, shouldn't I have > 3 points at which I look at the tire-track collisions?', since one end > of the tire is probably in the air, the middle is so-so, and the outer > part firmly touches the ground. > Nice for those temperature calculations. > > </rambling> > > > Ruud van Gaal > Free car sim : http://www.marketgraph.nl/gallery/racer/ > Pencil art : http://www.marketgraph.nl/gallery/ > >

Some system's I've seen are shimming the inner attachment point of the lower wishbone and adjustment at the upper kingpin pivot via offset bushes. I'm sure there are many other (and more elegant!) methods out there. Ben "Pat Dotson" <[email hidden]> wrote in message news:[email hidden]...

Quoted text (21 lines)

> "J. Todd Wasson" wrote: > > > > >If we then wanted to add some camber we would either have to change > > >the length of one of the arms or change the vertical distance between > > >either the inner or outer ball-joints. Right? > > > > Right... > > Not necessarily. With karts, the front spindles (axles) are made at an > angle to the kingpin. Camber is built in even if the lateral KPI is > zero. Camber is usually adustable on a kart via Heim joints, or a > "pill" that adjusts lateral KPI. > > [Can you tell I started racing karts this year? :) ] > > I'm not totally familiar with the actual form of a suspended car's front > end, but camber adjustment is surely possible without changing out a > swingarm. > > -- > Pat Dotson

On Fri, 11 Jan 2002 18:00:12 -0000, "Jonny Hodgson" <[email hidden]> wrote:

Quoted text (14 lines)

> >Ruud van Gaal <[email hidden]> wrote in message news:[email hidden]... >> On Fri, 11 Jan 2002 14:40:51 -0000, "Jonny Hodgson" >> <[email hidden]> wrote: > >> >You can probably assume linearity for road/race cars; if you're >> >going off-road, you might want to consider at least a quadratic >> >approximation or even something more accurate. >> >> Good to know, thanks for that piece of info. > >Do bear in mind that unless I cite actual evidence / experience, >I'm only talking in approximate terms so don't take everything >I type as gospel ;-)

:) You'd say the camber change is really a sine; probably the wheel makes a circular motion when going through the suspension. But for small sine values, sin(x)~x. Ruud van Gaal Free car sim : http://www.marketgraph.nl/gallery/racer/ Pencil art : http://www.marketgraph.nl/gallery/

J. Todd Wasson <[email hidden]> wrote in message news:[email hidden]... > Right. I've got two other adjustments in my suspension system in addition > though that are important IMO. We're doing a little radio controlled car sim. > The VectorNT chassis, for example, has two little screws in each front wheel Ha! You wouldn't disclose who you were working for last time I asked... thank you very much! <eg> Is this for internal/development use, or is said company planning on releasing this as commercial software then? </pushing luck> Jonny

Ruud van Gaal <[email hidden]> wrote in message news:[email hidden]...

Quoted text (11 lines)

> >> On Fri, 11 Jan 2002 14:40:51 -0000, "Jonny Hodgson" > >> <[email hidden]> wrote: > > > >> >You can probably assume linearity for road/race cars; if you're > >> >going off-road, you might want to consider at least a quadratic > >> >approximation or even something more accurate. > :) > You'd say the camber change is really a sine; probably the wheel makes > a circular motion when going through the suspension. But for small > sine values, sin(x)~x.

Hmmm... you might be able to sell me a Fourier series, although I doubt a single sine comes anywhere near once you get to large angles. Point taken and agreed in practice, however ;-) Jonny

Quoted text (9 lines)

>Ha! You wouldn't disclose who you were working for last time I >asked... thank you very much! <eg> > >Is this for internal/development use, or is said company planning >on releasing this as commercial software then? ></pushing luck> > >Jonny >

LOL!! Damn! You got me (I remember that thread :-))... We're hopefully going to wrap up things in a couple months here. It'll be a free download (advertising for you know who) and I'll let you all know if y'all care too. Warning.... "It's no GPL"! Todd Wasson --- Performance Simulations Drag Racing and Top Speed Prediction Software http://PerformanceSimulations.Com My little car sim screenshots: http://performancesimulations.com/scnshot4.htm

J. Todd Wasson <[email hidden]> wrote in message news:[email hidden]...

Quoted text (14 lines)

> >Ha! You wouldn't disclose who you were working for last time I > >asked... thank you very much! <eg> > > > >Is this for internal/development use, or is said company planning > >on releasing this as commercial software then? > ></pushing luck> > > > >Jonny > > > > LOL!! Damn! You got me (I remember that thread :-))... We're hopefully > going to wrap up things in a couple months here. It'll be a free download > (advertising for you know who) and I'll let you all know if y'all care too. > Warning.... "It's no GPL"!

Having seen their 1/10th offering running in the wet with a Can-Am wedge bodyshell, trailing twin vortices in the spray as second gear slams in... hell yeah, I'm interested! Jonny

Quoted text (13 lines)

>> LOL!! Damn! You got me (I remember that thread :-))... We're hopefully >> going to wrap up things in a couple months here. It'll be a free download >> (advertising for you know who) and I'll let you all know if y'all care too. >> Warning.... "It's no GPL"! > >Having seen their 1/10th offering running in the wet with a Can-Am >wedge bodyshell, trailing twin vortices in the spray as second gear >slams in... hell yeah, I'm interested! > >Jonny > > >

Haha! Well, we don't have any wet weather stuff in. Lots of extra work and I'm only in charge of the driving model itself so... It's a bit buggy right now sometimes, these little cars are a real b**ch to model and keep stable. Although we've got a suspension setup system that I haven't seen before yet.. You can do most of your own suspension geometry :0) Todd Wasson --- Performance Simulations Drag Racing and Top Speed Prediction Software http://PerformanceSimulations.Com My little car sim screenshots: http://performancesimulations.com/scnshot4.htm