cornering and braking-induced oversteer

I am in the process of building a vehicle simulation game, and before I add any assists and/or hacks to make the game easier to drive, I want to get very realistic behavior. Things are going very well so far, as we have the driveline simulated, realistic tire curves, independant suspension, a rigid-body base, and all the basics that a good simulator should have. The behavior is thus far very realistic, except for a major flaw...I think. If we are in a steady-state corner at or near the limit of lateral friction, and you apply the brakes, it creates pretty heavy understeer in our simulation. The car is set up with no ABS, and front/rear brake bias such that the fronts and rears will lock up at the same brake input in straight-line braking. Almost everyone I've talked to thinks that the car should oversteer (completely lose its rear) in this situation. I can see two major new torques acting on the body once the brakes are applied: 1) The weight-shift forward due to braking would cause an oversteer torque because the front tires have a greater load and therefore greater lateral force than the rears. 2) Since the outside-front tire has the greatest load, the longitudinal (braking) force of the tires will cause an understeer torque (this one is not as intuitive, but draw a diagram of a vehicle cornering, and consider the braking force/torque of all tires on the body, and the loads on each). In our simulator, the torque of (2) is greater than that of (1), and therefore the car understeers in this situation. I am wondering if any of you out there that have run into this problem in your own simulators. Or maybe someone knows what a real car would do. Any advice or input here would be greatly appreciated! Great newsgroup by the way...I have just recently discovered it :) -Dave Gierok PS: If I increase the rear brake-bias, I will get oversteer, but people that I've asked swear that it should oversteer regardless.

Dave Gierok wrote:

Quoted text (26 lines)

> I am in the process of building a vehicle simulation game, and before > I add any assists and/or hacks to make the game easier to drive, I > want to get very realistic behavior. Things are going very well so > far, as we have the driveline simulated, realistic tire curves, > independant suspension, a rigid-body base, and all the basics that a > good simulator should have. The behavior is thus far very realistic, > except for a major flaw...I think. > > If we are in a steady-state corner at or near the limit of lateral > friction, and you apply the brakes, it creates pretty heavy understeer > in our simulation. The car is set up with no ABS, and front/rear > brake bias such that the fronts and rears will lock up at the same > brake input in straight-line braking. Almost everyone I've talked to > thinks that the car should oversteer (completely lose its rear) in > this situation. I can see two major new torques acting on the body > once the brakes are applied: > > 1) The weight-shift forward due to braking would cause an oversteer > torque because the front tires have a greater load and therefore > greater lateral force than the rears. > > 2) Since the outside-front tire has the greatest load, the > longitudinal (braking) force of the tires will cause an understeer > torque (this one is not as intuitive, but draw a diagram of a vehicle > cornering, and consider the braking force/torque of all tires on the > body, and the loads on each).

Some more competing effects: If the car is rear wheel drive: Depending on what you have modeled, the motoring torques from the engine (if the clutch is still engaged) will also slow down the rear tires. This would act similarly to braking and make the brake bias calculations harder to get correct. If it is a race car, you can heavily load up the front tires with braking force and depending on how you implement "friction circle" types of effects in your tire model you can lose the side force capability as you approach locking up the front tires. Depending on how you model the effects of that heavy normal force, the fronts could also lose sideforce capability from that effect to the point where they still have less than the rears and the result is final understeer. Depending on how you model the profile of lateral force changing with normal force, you can also change the sideforce capability of the fronts through this effect. (They might have been just under their sideforce peak before braking but after loading them could perhaps be farther under the peak if you move it farther out with heavier normal force.) - Matt

I've never had this problem. How are you modeling the brakes themselves? I question your point number 2. In an ideal world, the pressure on both left- and right-front brake calipers should be the same, so if no wheels are locked up, with a straight coulomb friction model the brake torque will be symetrical left/right even though the the vertical tire load is asymmetrical--the outside tires will have more _capability_ for longitudinal friction than the inside tires, but without ABS or some sort of unequal left/right brake proportioning system the brake torques should be the same. Also, another thing could be if you have so much understeer in the chassis because of spring settings or bad suspension geometry that it swamps any other effect. The other things to check are your combined tire slip functions. Another good test is to set the car up with unequal ride heights (left/right or front/rear) and make sure there aren't any "unexpected" forces due to math errors in camber and geometry... -Dave "Dave Gierok" <[email hidden]> wrote in message news:[email hidden]...

Quoted text (39 lines)

> I am in the process of building a vehicle simulation game, and before > I add any assists and/or hacks to make the game easier to drive, I > want to get very realistic behavior. Things are going very well so > far, as we have the driveline simulated, realistic tire curves, > independant suspension, a rigid-body base, and all the basics that a > good simulator should have. The behavior is thus far very realistic, > except for a major flaw...I think. > > If we are in a steady-state corner at or near the limit of lateral > friction, and you apply the brakes, it creates pretty heavy understeer > in our simulation. The car is set up with no ABS, and front/rear > brake bias such that the fronts and rears will lock up at the same > brake input in straight-line braking. Almost everyone I've talked to > thinks that the car should oversteer (completely lose its rear) in > this situation. I can see two major new torques acting on the body > once the brakes are applied: > > 1) The weight-shift forward due to braking would cause an oversteer > torque because the front tires have a greater load and therefore > greater lateral force than the rears. > > 2) Since the outside-front tire has the greatest load, the > longitudinal (braking) force of the tires will cause an understeer > torque (this one is not as intuitive, but draw a diagram of a vehicle > cornering, and consider the braking force/torque of all tires on the > body, and the loads on each). > > In our simulator, the torque of (2) is greater than that of (1), and > therefore the car understeers in this situation. I am wondering if > any of you out there that have run into this problem in your own > simulators. Or maybe someone knows what a real car would do. > > Any advice or input here would be greatly appreciated! Great > newsgroup by the way...I have just recently discovered it :) > > -Dave Gierok > > PS: If I increase the rear brake-bias, I will get oversteer, but > people that I've asked swear that it should oversteer regardless.

I meant coulomb friction for the brake pad-to-brake rotor contact, not the tire-ground contact, of course! ;) "Dave Pollatsek" <[email hidden]> wrote in message news:WQ0y9.70755$[email hidden]... > I've never had this problem. How are you modeling the brakes themselves?I > question your point number 2. In an ideal world, the pressure on bothleft- > and right-front brake calipers should be the same, so if no wheels are > locked up, with a straight coulomb friction model the brake torque will be > symetrical left/right even though the the vertical tire load is > asymmetrical--the outside tires will have more _capability_ forlongitudinal > friction than the inside tires, but without ABS or some sort of unequal > left/right brake proportioning system the brake torques should be thesame.

Quoted text (54 lines)

> > Also, another thing could be if you have so much understeer in the chassis > because of spring settings or bad suspension geometry that it swamps any > other effect. > > The other things to check are your combined tire slip functions. Another > good test is to set the car up with unequal ride heights (left/right or > front/rear) and make sure there aren't any "unexpected" forces due to math > errors in camber and geometry... > -Dave > > "Dave Gierok" <[email hidden]> wrote in message > news:[email hidden]... > > I am in the process of building a vehicle simulation game, and before > > I add any assists and/or hacks to make the game easier to drive, I > > want to get very realistic behavior. Things are going very well so > > far, as we have the driveline simulated, realistic tire curves, > > independant suspension, a rigid-body base, and all the basics that a > > good simulator should have. The behavior is thus far very realistic, > > except for a major flaw...I think. > > > > If we are in a steady-state corner at or near the limit of lateral > > friction, and you apply the brakes, it creates pretty heavy understeer > > in our simulation. The car is set up with no ABS, and front/rear > > brake bias such that the fronts and rears will lock up at the same > > brake input in straight-line braking. Almost everyone I've talked to > > thinks that the car should oversteer (completely lose its rear) in > > this situation. I can see two major new torques acting on the body > > once the brakes are applied: > > > > 1) The weight-shift forward due to braking would cause an oversteer > > torque because the front tires have a greater load and therefore > > greater lateral force than the rears. > > > > 2) Since the outside-front tire has the greatest load, the > > longitudinal (braking) force of the tires will cause an understeer > > torque (this one is not as intuitive, but draw a diagram of a vehicle > > cornering, and consider the braking force/torque of all tires on the > > body, and the loads on each). > > > > In our simulator, the torque of (2) is greater than that of (1), and > > therefore the car understeers in this situation. I am wondering if > > any of you out there that have run into this problem in your own > > simulators. Or maybe someone knows what a real car would do. > > > > Any advice or input here would be greatly appreciated! Great > > newsgroup by the way...I have just recently discovered it :) > > > > -Dave Gierok > > > > PS: If I increase the rear brake-bias, I will get oversteer, but > > people that I've asked swear that it should oversteer regardless. > >

Hi Dave, from ample experience <g> ...an average car in most cases in such a situation would begin to oversteer when you apply the brake lightly or moderately. If you fully slam the brakes and quickly enough, it will usually understeer (if you don't slam them fast enough, its body will go in rapid motion through the movement that accompanies the weight transfers and announces the oversteer that would happen if you'd not slammed the brakes but only applied them normally). An anomaly from this behaviour might in my experience occur in a real car if the front outside tire doesn't have enough grip relative to the rear tires, like not having enough weight on the (especially outside) front wheel(s) or too much weight on the (especially inside) rear wheel(s) (heavy cargo in trunk?) especially if the road surface is slippery. An exception could also occur if the outside front tire suspends all the transferred weight instead of increasing the friction at the contact patch between asphalt and tire, like if the tire were underinflated. Too much braking force at the rear wheels could have a similar effect, as that would reduce the weight transfers. Hope this helps - apologies for not being able to make more qualified and knowledgeable contributions. Achim

On Wed, 6 Nov 2002 07:55:24 -0000, "Joachim Trensz" <[email hidden]> wrote: >If you fully slam the brakes and quickly enough, it will usually understeer I'm possibly off the mark here, but if you fully lock the brakes rapidly enough, and all 4 wheels stop spinning simultaneously, I'm guessing that the car would understeer because of centrifigual force? Without having traction at any tire, I would assume that the car would want to travel in a straight line... As far as mid-corner braking without locking causing oversteer, I would think that you would be shifting more weight to the front wheels, increasing the relative traction at that end of the car and increasing the traction available for steering (plus the decrease in speed would allow the car to take the turn with a smaller radius using the same amount of wheel lock). Probably off again there, too, but that's what I get when I attempt to visualize it. Jason

Quoted text (11 lines)

>If we are in a steady-state corner at or near the limit of lateral >friction, and you apply the brakes, it creates pretty heavy understeer >in our simulation. The car is set up with no ABS, and front/rear >brake bias such that the fronts and rears will lock up at the same >brake input in straight-line braking. Almost everyone I've talked to >thinks that the car should oversteer (completely lose its rear) in >this situation. I can see two major new torques acting on the body >once the brakes are applied: > >PS: If I increase the rear brake-bias, I will get oversteer, but >people that I've asked swear that it should oversteer regardless.

Certainly braking can cause understeer sometimes. It depends on the usual things, weight distribution, brake bias, setup, cornering load, I've experienced it in NASCAR2002 and in real front wheel drive cars. I think it's related to simply overloading the front tires. Whether or not the understeer you see is realistic, I haven't a clue.

Hi Jason, in the cars I've driven, when driving around a corner near the limit, simply reducing the throttle a bit infallibly made the rear end come around. I've used this often in real cars to help steering, and it works in N2002 and GPL as well. The cars behaviour of course also depends on factors I mentioned in my other message, or the car's setup - brake balance, weight distribution, which axle is driven, diff settings etc. But I'll leave the theoretical discussion to those who know more about it than I do :-) Achim "Jason Moyer" <[email hidden]> wrote in message news:[email hidden]... > On Wed, 6 Nov 2002 07:55:24 -0000, "Joachim Trensz" > <[email hidden]> wrote: > > >If you fully slam the brakes and quickly enough, it will usuallyundersteer

Quoted text (16 lines)

> > I'm possibly off the mark here, but if you fully lock the brakes > rapidly enough, and all 4 wheels stop spinning simultaneously, I'm > guessing that the car would understeer because of centrifigual force? > Without having traction at any tire, I would assume that the car would > want to travel in a straight line... > > As far as mid-corner braking without locking causing oversteer, I > would think that you would be shifting more weight to the front > wheels, increasing the relative traction at that end of the car and > increasing the traction available for steering (plus the decrease in > speed would allow the car to take the turn with a smaller radius using > the same amount of wheel lock). Probably off again there, too, but > that's what I get when I attempt to visualize it. > > Jason

Quoted text (8 lines)

> I've never had this problem. How are you modeling the brakes themselves? I > question your point number 2. In an ideal world, the pressure on both left- > and right-front brake calipers should be the same, so if no wheels are > locked up, with a straight coulomb friction model the brake torque will be > symetrical left/right even though the the vertical tire load is > asymmetrical--the outside tires will have more _capability_ for longitudinal > friction than the inside tires, but without ABS or some sort of unequal > left/right brake proportioning system the brake torques should be the same.

The brakes are modeled by applying a friction torque to the wheels. And you're right that both left/right brake torques are equal. But with the wheels turned, even if the braking FRICTION is the same on the front Left/Right tires, their torques on the CAR are not equal because the outside front tire will have a greater lever-arm. Also, there is a lot of weight transfer to the outside wheels (maybe more than there should be), and therefore it looks like the front inside tire is locking up with very little braking. The rear inside tire does not lock up because this car is RWD and set up with a limitted-slip-differential. > Also, another thing could be if you have so much understeer in the chassis > because of spring settings or bad suspension geometry that it swamps any > other effect. The front springs are SLIGHTLY stiffer than the rear (no sway-bars yet) and the car tends to be fairly neutral, but leans towards understeer. > The other things to check are your combined tire slip functions. Another > good test is to set the car up with unequal ride heights (left/right or > front/rear) and make sure there aren't any "unexpected" forces due to math > errors in camber and geometry... > -Dave I'll check that out...thanks for the advice

From: [email hidden] (Dave Gierok) Newsgroups: rec.autos.simulators Subject: Re: cornering and braking-induced oversteer References: <[email hidden]> <WQ0y9.70755$[email hidden]> NNTP-Posting-Host: 131.107.3.86 Message-ID: <[email hidden]>

Quoted text (8 lines)

> I've never had this problem. How are you modeling the brakes themselves? I > question your point number 2. In an ideal world, the pressure on both left- > and right-front brake calipers should be the same, so if no wheels are > locked up, with a straight coulomb friction model the brake torque will be > symetrical left/right even though the the vertical tire load is > asymmetrical--the outside tires will have more _capability_ for longitudinal > friction than the inside tires, but without ABS or some sort of unequal > left/right brake proportioning system the brake torques should be the same.

The brakes are modeled by applying a friction torque to the wheels. And you're right that both left/right brake torques are equal. But with the wheels turned, even if the braking FRICTION is the same on the front Left/Right tires, their torques on the CAR are not equal because the outside front tire will have a greater lever-arm. Also, there is a lot of weight transfer to the outside wheels (maybe more than there should be), and therefore it looks like the front inside tire is locking up with very little braking. The rear inside tire does not lock up because this car is RWD and set up with a limitted-slip-differential. > Also, another thing could be if you have so much understeer in the chassis > because of spring settings or bad suspension geometry that it swamps any > other effect. The front springs are SLIGHTLY stiffer than the rear (no sway-bars yet) and the car tends to be fairly neutral, but leans towards understeer. > The other things to check are your combined tire slip functions. Another > good test is to set the car up with unequal ride heights (left/right or > front/rear) and make sure there aren't any "unexpected" forces due to math > errors in camber and geometry... > -Dave I'll check that out...thanks for the advice.

Contributing to the debate... I find that if I apply the brakes when half way round a corner, several things happen in my RWD, limited slip diff car.. Normally I don't go from full throttle to full brake, in reality it's more a case of balance on the throttle, lift off, then hit the brakes... In the above situation, you've already induced oversteer by lifting off, and having engine torque try and slow the rear wheels down, on a big V8 (not mine!) I would presume that it's a lot more evident. Anyway, if you then whack on the brakes you're exacerbating the situation, as you're slowing all wheels down just as the body roll comes in. Basically, the car wants to go past the wheels, so hitting the brakes will cause the heaviest part of the car to want to continue on. If you've induced oversteer with the engine breaking then slapping on the brakes will either continue to induce oversteer (as in the Porsche news group!) or if you've got a car that's dart like in it's balance, you could see understeer. I believe that a lot of how the car handles is down to weight distribution and the point that the effort (+ve or -ve) is being transmitted to the tarmac, stiffer springs and anti-roll bars just help damp or reduce the effects that would be happening, to get the car physics good I'd be running a car that had almost no dampers so that I could ensure that the basic physics are sound, then stiffen up, and tweak. Regards, AndyC "Dave Gierok" <[email hidden]> wrote in message news:[email hidden]...

Quoted text (39 lines)

> I am in the process of building a vehicle simulation game, and before > I add any assists and/or hacks to make the game easier to drive, I > want to get very realistic behavior. Things are going very well so > far, as we have the driveline simulated, realistic tire curves, > independant suspension, a rigid-body base, and all the basics that a > good simulator should have. The behavior is thus far very realistic, > except for a major flaw...I think. > > If we are in a steady-state corner at or near the limit of lateral > friction, and you apply the brakes, it creates pretty heavy understeer > in our simulation. The car is set up with no ABS, and front/rear > brake bias such that the fronts and rears will lock up at the same > brake input in straight-line braking. Almost everyone I've talked to > thinks that the car should oversteer (completely lose its rear) in > this situation. I can see two major new torques acting on the body > once the brakes are applied: > > 1) The weight-shift forward due to braking would cause an oversteer > torque because the front tires have a greater load and therefore > greater lateral force than the rears. > > 2) Since the outside-front tire has the greatest load, the > longitudinal (braking) force of the tires will cause an understeer > torque (this one is not as intuitive, but draw a diagram of a vehicle > cornering, and consider the braking force/torque of all tires on the > body, and the loads on each). > > In our simulator, the torque of (2) is greater than that of (1), and > therefore the car understeers in this situation. I am wondering if > any of you out there that have run into this problem in your own > simulators. Or maybe someone knows what a real car would do. > > Any advice or input here would be greatly appreciated! Great > newsgroup by the way...I have just recently discovered it :) > > -Dave Gierok > > PS: If I increase the rear brake-bias, I will get oversteer, but > people that I've asked swear that it should oversteer regardless.