Koni Sport dampers and Evolution driving school..

[Norm Peterson]

Is any of that online? If so, where?

I get the print magazine, so I haven't really tried to find this stuff online.

Where would the initial oversteer come from?

Greater lateral load transfer would obviously occur initially at the rear with the higher rear geo RC. Actually, this is a component of understeer/oversteer, just one of several, and its individual effect would be oversteerish. It's a slip angle thing (difference in slip angles, actually), and applicable at any level of lateral acceleration (though obviously more noticeable at the higher lateral g's).

The front roll moment is greater than the rear roll moment. Therefore, at the point where the rear would be stabilized in roll due to the weight transfer in the rear, the front has not stabilized. The additional roll in the front is transferred to the rear via the chassis, which increases the amount of compression force on the outside rear tire while the amount of lateral force there remains the same, and therefore the rear gets greater compression force at the tire and thus greater adhesion. Greater adhesion in the rear than the front gets you understeer.

Unless the car really is a wet noodle, the front and rear will stabilize in roll at the same time.

LLT through the roll centers is fixed, consider it a function only of the sprung mass CG X-axis (longitudinal) location.

LLT through suspension roll stiffnesses is apportioned according to their relative stiffnesses (assuming a decently rigid chassis). With linear-rate springs and bars and motion ratios that don't shift much over suspension deflection, this is also a relatively fixed relationship. Damper behavior kind of messes with this as the roll rotation is actually in progress, but drops out at steady-state (yeah, I know this doesn't ever really occur, but it probably helps here to assume that it can).

The M3 might well behave like that.

I suspect my car does the way it sits (and I had an instructor specifically mention it recently); the M3 with a more level "roll axis" is going to roll like a boat in a beam sea and start out with a better balance between front and rear LLT due to RC heights.

What I've noticed from the Mustang's suspension is that if I go into a corner hard enough, I'll get a lot of initial understeer unless I go into it with some trail braking. Once the car is going around the corner, I can control the rear with the throttle. With the stock suspension without camber plates, I was able to make the car "tuck in" by lifting the throttle, but the rear wouldn't come out. With camber plates, I can make the rear come out simply by lifting off the throttle.

So it seems like the car's behavior is basically corner entry understeer (except when trail braking), and mid-corner neutrality (or darned close to it), and corner exit oversteer can be induced if you give it enough throttle on corner exit.

You should be able to easily throttle-steer through the middle with only small throttle adjustments or find a throttle setting that holds you on line through turns of 180° or more.

What changes can be made to the car to make corner entry neutral, but otherwise retain the characteristics it seems to have with camber plates?

Aside from front tires and wheel widths, tire pressures, and tinkering with the dampers, I'm thinking front control arm bushings (firmer).


Norm

 

[kcbrown]

Greater lateral load transfer would obviously occur initially at the rear with the higher rear geo RC. Actually, this is a component of understeer/oversteer, just one of several, and its individual effect would be oversteerish. It's a slip angle thing (difference in slip angles, actually), and applicable at any level of lateral acceleration (though obviously more noticeable at the higher lateral g's).

I guess you could call it "oversteerish" in terms of its contribution to the total picture, but it wouldn't really come into play until at least some of the tires start to break traction. That's why it's not clear to me if it will actually matter in terms of how the car behaves at the traction limits, since the overall geometry has changed at that point.

Unless the car really is a wet noodle, the front and rear will stabilize in roll at the same time.

Understood. I was using that to describe my thinking in terms of the force contributions, and why I said the car would have a tendency to understeer with a lower roll moment in the rear than in the front.

LLT through suspension roll stiffnesses is apportioned according to their relative stiffnesses (assuming a decently rigid chassis). With linear-rate springs and bars and motion ratios that don't shift much over suspension deflection, this is also a relatively fixed relationship. Damper behavior kind of messes with this as the roll rotation is actually in progress, but drops out at steady-state (yeah, I know this doesn't ever really occur, but it probably helps here to assume that it can).

But LLT isn't the whole picture as regards the traction equation. Vertical forces have to be accounted for as much as horizontal ones (indeed, it's because of the vertical force distribution in braking that trail braking results in oversteer in cars that are otherwise prone to understeer).

I suspect my car does the way it sits (and I had an instructor specifically mention it recently); the M3 with a more level "roll axis" is going to roll like a boat in a beam sea and start out with a better balance between front and rear LLT due to RC heights.

What changes have you made to your car? I know it has Koni Sport dampers, and seem to recall that you're still on stock springs and sway bars. I don't recall any other changes.

You should be able to easily throttle-steer through the middle with only small throttle adjustments or find a throttle setting that holds you on line through turns of 180° or more.

That is indeed the case with my car as it is now. The difference is in how much oversteer can be induced by lifting the throttle. The car is closer to neutral with the camber plates than without, which shouldn't be a surprise given what it does to the front contact patch.

Aside from front tires and wheel widths, tire pressures, and tinkering with the dampers, I'm thinking front control arm bushings (firmer).

The problem with changing the front tires and wheel widths is that it would affect the balance in all regimes, not just corner entry. But playing with the dampers might do it (here I presume I'd need more rebound in the rear. Right now the dampers are set to give about the same percentage of critical damping all the way around).


That said, there may be some wisdom in leaving it as it is, since changing the corner entry balance will almost certainly make the car more prone to oversteer on corner entry with trail braking. That's fine (and may even be desirable) on an autocross course, but probably isn't a wise thing to mess with on the track until driving skill has improved enough to handle it. I don't think I'm there yet. Perhaps in another couple of years...

 

[Norm Peterson]

I guess you could call it "oversteerish" in terms of its contribution to the total picture, but it wouldn't really come into play until at least some of the tires start to break traction. That's why it's not clear to me if it will actually matter in terms of how the car behaves at the traction limits, since the overall geometry has changed at that point.

LLT through the geo roll centers is always there and always affects the tires' slip angles. Understeer can be defined as [front slip angle] - [rear slip angle], so as long as there is more LLT through the rear RC than through the front RC it will be oversteerish. I suppose you could intentionally give your car such an adverse amount of tire size stagger or use tires of such different cornering properties that the LLT effect could be understeer even with a much higher rear than front RC and fairly even weight distribution. But then again, we probably wouldn't be discussing such a vehicle in this particular forum.

But LLT isn't the whole picture as regards the traction equation. Vertical forces have to be accounted for as much as horizontal ones (indeed, it's because of the vertical force distribution in braking that trail braking results in oversteer in cars that are otherwise prone to understeer).

LLT is very much about vertical tire forces, specifically the changes in them when the car undergoes lateral acceleration. Trail braking involves management of the vector sums of lateral and longitudinal forces by introducing an aX component and longitudinal load transfer to the aY and lateral load transfer.

What changes have you made to your car? I know it has Koni Sport dampers, and seem to recall that you're still on stock springs and sway bars. I don't recall any other changes.

Koni sports, OE springs, Sam's 35/22 bars (this was his original bar set, the 25 rear bar came later). Somewhere around -2° camber (I don't think it ever had -1.5° or less, even when brand new). Poly/spherical rear LCAs (the poly has been tweaked a bit). All three of the wheel & tire packages have wheel widths equal to max recommended for the tire size (even the 8.5" wide OE wheels were the max recommended for the OE 235/50-18's).


Norm

 

[kcbrown]

LLT through the geo roll centers is always there and always affects the tires' slip angles. Understeer can be defined as [front slip angle] - [rear slip angle], so as long as there is more LLT through the rear RC than through the front RC it will be oversteerish. I suppose you could intentionally give your car such an adverse amount of tire size stagger or use tires of such different cornering properties that the LLT effect could be understeer even with a much higher rear than front RC and fairly even weight distribution. But then again, we probably wouldn't be discussing such a vehicle in this particular forum.

I guess that's true, but as a driver, you won't really notice any of that until one of the ends starts to break loose, or gets close enough to the limits that the tires start to give feedback.

LLT is very much about vertical tire forces, specifically the changes in them when the car undergoes lateral acceleration. Trail braking involves management of the vector sums of lateral and longitudinal forces by introducing an aX component and longitudinal load transfer to the aY and lateral load transfer.

My apologies. I had a "blonde moment" or something, I guess. Of course LLT is about vertical tire forces. Indeed, it doesn't make sense for it to really be about anything else. Sigh...

Anyway, if a low roll moment results in most of the weight transfer occurring through the LCAs, and the LCAs are (for the purpose of the exercise) horizontal to the ground, and weight transfer really means a transfer of vertical forces to the tires, then how in the world is that vertical tire force exerted through what amounts to a horizontal member? It implies a torque has to be applied to the LCA, but how, when the only link that could do so in the front is the strut, and the only links that could in the rear are the springs and shocks? And since it's the chassis weight transfer we're talking about, force transfer through the struts/springs/shocks must occur through compression/extension of those members, and if we remove the dampers for the purpose of simplifying the exercise, then the transfer must result in compression/extension of the spring in proportion to the amount of weight transferred. But compression on one side and extension on the other is exactly what results in roll. If (for illustrative purposes) the roll moment on one end is zero (RC height and CG height are the same there), then you end up with a contradiction: body roll (because weight transfer is happening) when such roll shouldn't be possible.


That's why my initial thought was that weight transfer had to involve horizontal forces as well, but that doesn't make any sense either as there's no transfer involved there.

I am now thoroughly confused.

Koni sports, OE springs, Sam's 35/22 bars (this was his original bar set, the 25 rear bar came later). Somewhere around -2° camber (I don't think it ever had -1.5° or less, even when brand new). Poly/spherical rear LCAs (the poly has been tweaked a bit). All three of the wheel & tire packages have wheel widths equal to max recommended for the tire size (even the 8.5" wide OE wheels were the max recommended for the OE 235/50-18's).

What are your spring rates? I believe they differ from those of the Brembo-equipped GTs, don't they?

The Brembo cars have a 34.6mm front bar and a 24mm rear bar, which sounds similar to your sway bar setup (actually, probably stiffer in the rear than yours).

 

[Norm Peterson]

I guess that's true, but as a driver, you won't really notice any of that until one of the ends starts to break loose, or gets close enough to the limits that the tires start to give feedback.

If you feel anything at all - which isn't guaranteed, this will be part of why a car might feel like it progressively understeers a little more as you go deeper into the turn.

My apologies. I had a "blonde moment" or something, I guess. Of course LLT is about vertical tire forces. Indeed, it doesn't make sense for it to really be about anything else. Sigh...

Anyway, if a low roll moment results in most of the weight transfer occurring through the LCAs, and the LCAs are (for the purpose of the exercise) horizontal to the ground, and weight transfer really means a transfer of vertical forces to the tires, then how in the world is that vertical tire force exerted through what amounts to a horizontal member? It implies a torque has to be applied to the LCA, but how, when the only link that could do so in the front is the strut, and the only links that could in the rear are the springs and shocks? And since it's the chassis weight transfer we're talking about, force transfer through the struts/springs/shocks must occur through compression/extension of those members, and if we remove the dampers for the purpose of simplifying the exercise, then the transfer must result in compression/extension of the spring in proportion to the amount of weight transferred. But compression on one side and extension on the other is exactly what results in roll. If (for illustrative purposes) the roll moment on one end is zero (RC height and CG height are the same there), then you end up with a contradiction: body roll (because weight transfer is happening) when such roll shouldn't be possible.

At the rear axle, the LCAs have essentially nothing to do with lateral forces (to be picky, their tiny amount of plan view skew away from the purely longitudinal means that they would have some very tiny effect - that can safely be neglected). The PHB (or Watts link) lateral force is applied at axle height, which resolves statically to both vertical and lateral forces at the contact patches.

Up front, you probably need to think in terms of force resolution for a line drawn from each contact patch to the front view instant center for that patch's suspension. For a geo-RC above ground, there will necessarily be a vertical force component as well as the obvious lateral component. IOW, you can't think in terms of the LCA alone since the entire suspension, wheel, and tire rotates about the FVIC instead. (If you're thinking that locating the static geo-RC below-grade would be a good thing, you'd be introducing more roll, and reducing the initial front LLT would increase the initial oversteerish tendency. And you'd increase the shift toward understeer as roll develops assuming that you'd build even more front roll stiffness into the car to counteract the added roll).

What are your spring rates? I believe they differ from those of the Brembo-equipped GTs, don't they?

The Brembo cars have a 34.6mm front bar and a 24mm rear bar, which sounds similar to your sway bar setup (actually, probably stiffer in the rear than yours).

Most estimates I've seen seem to be based on the MVMS document that's been floating around, which lists 136 lb/in front and 142 lb/in rear for the springs themselves. However, the wheel rates listed in that document are actually slightly higher, which was explained to me once as being due to the listed wheel rates actually including the control arm bushing torsional stiffnesses about the pivot axes of rotation.

Yes, any 24mm rear bar is going to be stiffer than my 22, even at its stiffest setting (where mine has been sitting since very shortly after I installed it).

It's been nearly three years since the thumbnail picture was taken, and I don't have any better idea what the lateral acceleration might have been than to guess somewhere between maybe 0.8 and 0.95.


Norm

 

[kcbrown]

If you feel anything at all - which isn't guaranteed, this will be part of why a car might feel like it progressively understeers a little more as you go deeper into the turn.

I suspect understeer is easier to sense than oversteer once you know roughly what steering input you'd be putting in for a given corner prior to breaking traction. If you're putting more than that in while getting the same path through the corner, then you're getting understeer. In a sense, it seems to me you're detecting the slip angle. Oversteer requires that you sense the rear end coming loose. I don't know of a way to feel the slip angle change back there before the rear end simply breaks free.

At the rear axle, the LCAs have essentially nothing to do with lateral forces (to be picky, their tiny amount of plan view skew away from the purely longitudinal means that they would have some very tiny effect - that can safely be neglected). The PHB (or Watts link) lateral force is applied at axle height, which resolves statically to both vertical and lateral forces at the contact patches.

Right. But the principles are the same, in that vertical forces at the tires have to be transmitted through some vertical element in the rear suspension. For this, that would be the springs and shocks.

Up front, you probably need to think in terms of force resolution for a line drawn from each contact patch to the front view instant center for that patch's suspension. For a geo-RC above ground, there will necessarily be a vertical force component as well as the obvious lateral component. IOW, you can't think in terms of the LCA alone since the entire suspension, wheel, and tire rotates about the FVIC instead. (If you're thinking that locating the static geo-RC below-grade would be a good thing, you'd be introducing more roll, and reducing the initial front LLT would increase the initial oversteerish tendency. And you'd increase the shift toward understeer as roll develops assuming that you'd build even more front roll stiffness into the car to counteract the added roll).

But the problem with that is that at the end of the day, the forces have to be transmitted by some physical member. While the FVIC analysis may indicate an increased vertical force on the outside tires, if the roll moment is zeroed out (say, because the RC and the CG are at the same height), and the only vertical elements in the suspension involve the springs (i.e., the struts) and dampers, then how can load transfer (i.e., vertical force transfer from one side of the car to the other) possibly happen while simultaneously resulting in no roll? That seems like an impossibility, a contradiction.

Yes, any 24mm rear bar is going to be stiffer than my 22, even at its stiffest setting (where mine has been sitting since very shortly after I installed it).

Given the roll centers and CG locations, and the stock suspension geometry, is the thicker bar in the rear likely to yield more oversteer, or will it get you more understeer?

 

[Norm Peterson]

I suspect understeer is easier to sense than oversteer once you know roughly what steering input you'd be putting in for a given corner prior to breaking traction. If you're putting more than that in while getting the same path through the corner, then you're getting understeer. In a sense, it seems to me you're detecting the slip angle. Oversteer requires that you sense the rear end coming loose. I don't know of a way to feel the slip angle change back there before the rear end simply breaks free.

:dunno . . . . I've always found mild amounts of "looseness" to be more evident than an equal or slightly greater amount of understeer.

Right. But the principles are the same, in that vertical forces at the tires have to be transmitted through some vertical element in the rear suspension. For this, that would be the springs and shocks.

I think we've found the problem.

It's easiest to show geometric LLT in general via what looks like a rear axle with a PHB, but if you can imagine the front strut suspension having a virtual PHB running through its geo roll center it'd be the same picture (albeit with the "PHB" drawn much closer to the ground).

The red arrow is the chassis inertial force and the blue arrows are how the reaction to that force resolves statically down at the contact patches. Nothing else need be involved***.

(*** Unless you're trying to account for effects like moment stiffness in the shock/strut attachment points, separate springs having moment resistances at their seats, and net lateral forces developed in tie rods, toe rods, or any sta-bar that might be present.)


Norm

 

[Norm Peterson]

I intentionally split this response so that the thumbnail in the other post would more closely follow the text it pertains to.

But the problem with that is that at the end of the day, the forces have to be transmitted by some physical member. While the FVIC analysis may indicate an increased vertical force on the outside tires, if the roll moment is zeroed out (say, because the RC and the CG are at the same height), and the only vertical elements in the suspension involve the springs (i.e., the struts) and dampers, then how can load transfer (i.e., vertical force transfer from one side of the car to the other) possibly happen while simultaneously resulting in no roll? That seems like an impossibility, a contradiction.

I'm hoping that the thumbnail in the above post clears this up for you. Consider what would happen if the CG of the "Body" rectangle coincided with the red arrow PHB . . . there would be no "body roll" relative to the wheels.

Given the roll centers and CG locations, and the stock suspension geometry, is the thicker bar in the rear likely to yield more oversteer, or will it get you more understeer?

It can't help but "loosen" the handling, aka have an oversteer effect. There will be a very small reduction in roll, which in a stick axle car will benefit the front tires more than the rear tires, and assuming no other changes will drag a little more of the elastic LLT rearward. Both effects are in the direction of reducing understeer - which would include shifting the balance into oversteer or even increasing oversteer if that's what you started with.

It's not exactly like the front bar situation, where the bar is much more effective in reducing roll even as it increases front elastic LLT (the bar effects are now acting in opposition). A big front bar is worth more in maintaining camber when the springs are soft (IOW, OE or with most lowering spring rates) than when the springs are much stiffer and account for greater roll stiffness.


Norm

 

[claudermilk]

I suspect understeer is easier to sense than oversteer once you know roughly what steering input you'd be putting in for a given corner prior to breaking traction.

:dunno . . . . I've always found mild amounts of "looseness" to be more evident than an equal or slightly greater amount of understeer.

I've been lurking, reading the thread ans was just going to point out the same thing. I seem to be able to feel oversteer much quicker than understeer. I seem to be able to instantly feel every wiggle from the rear, but the car needs to plow a bit before I clue in.

 

[kcbrown]

I intentionally split this response so that the thumbnail in the other post would more closely follow the text it pertains to.


I'm hoping that the thumbnail in the above post clears this up for you. Consider what would happen if the CG of the "Body" rectangle coincided with the red arrow PHB . . . there would be no "body roll" relative to the wheels.

Yeah, that clears it up. The vertical force vectors are necessary to resolve the torque equation that you'd have to set up at the wheel hubs (or, really, wherever you wish).

I've noticed that a lot of effort goes into controlling the camber curve of the front tires, but if you can geometrically change the roll center and move it closer to the CG, then wouldn't that have beneficial effects on control of the camber, seeing how a zero roll moment results in LLT without any suspension movement?

It can't help but "loosen" the handling, aka have an oversteer effect. There will be a very small reduction in roll, which in a stick axle car will benefit the front tires more than the rear tires, and assuming no other changes will drag a little more of the elastic LLT rearward. Both effects are in the direction of reducing understeer - which would include shifting the balance into oversteer or even increasing oversteer if that's what you started with.

It's not exactly like the front bar situation, where the bar is much more effective in reducing roll even as it increases front elastic LLT (the bar effects are now acting in opposition). A big front bar is worth more in maintaining camber when the springs are soft (IOW, OE or with most lowering spring rates) than when the springs are much stiffer and account for greater roll stiffness.

Since it seems that camber changes are the result of body roll during cornering, I still don't quite understand why lowering the car has a substantial net beneficial effect on the handling of the car when all else is kept equal (i.e., between a car with much stiffer springs but stock ride height versus a car with much stiffer springs and a lowered ride height). I was also under the impression that stiffer springs make it more difficult for the suspension to maintain contact with the road over larger (lower frequency) bumps, but may help with maintaining contact over smaller bumps, so it winds up being a balancing act such that you'd want to select the spring rate based primarily on the features of the road surface you'll be up against and use other means to control the camber curve. The resulting spring rate may be better for controlling the camber curve, but it would be a secondary benefit, while it seems that most people change the spring rates primarily for the purpose of camber control and let the question of surface interaction fall by the wayside.

 

[Norm Peterson]

Yeah, that clears it up. The vertical force vectors are necessary to resolve the torque equation that you'd have to set up at the wheel hubs (or, really, wherever you wish).

I've noticed that a lot of effort goes into controlling the camber curve of the front tires, but if you can geometrically change the roll center and move it closer to the CG, then wouldn't that have beneficial effects on control of the camber, seeing how a zero roll moment results in LLT without any suspension movement?

If it were only that simple . . . but the car becomes less "tuneable" (more LLT through the RCs in some fixed proportion, less through the suspension elastic elements that you do your front vs rear roll stiffness tuning with by stiffening or softening). Never mind that the FVSA gets too short and the tendency for 'jacking' increases. The FVSA length also affects the rate of camber gain, which would at least affect your static camber settings.

Since it seems that camber changes are the result of body roll during cornering, I still don't quite understand why lowering the car has a substantial net beneficial effect on the handling of the car when all else is kept equal (i.e., between a car with much stiffer springs but stock ride height versus a car with much stiffer springs and a lowered ride height).

In this specific comparison, the total LLT is lower for the lowered car, meaning that mechanical grip on the more heavily loaded outside tires goes up. This is always true, though I think people tend to credit it for being a bigger effect than it really is. An inch lowering only buys ~5% reduction in LLT, and that in turn buys less than a 5% grip improvement, all else held constant.

Obviously, a lower CG becomes more important when you're pulling higher lateral g's - at some point overturning becomes a real risk, even if you aren't running over apex curbs and getting a little extra upward kick on the "wrong" side of the car.

The resulting spring rate may be better for controlling the camber curve, but it would be a secondary benefit, while it seems that most people change the spring rates primarily for the purpose of camber control and let the question of surface interaction fall by the wayside.

Most race tracks and autocross lots are indeed smooth enough over most of their length to allow you to worry more about camber control than surface interaction.

Don't forget about retaining adequate pitch control. Running out of either bump or rebound travel can't be good things, certainly not if it puts you so hard into the stops that their effective rates have skyrocketed and suspension motion abruptly ceases.


Norm

 

[frank s]

Does this article contribute anything, or is it just a re-hash of something everyone already knows?

 

[kcbrown]

Does this article contribute anything, or is it just a re-hash of something everyone already knows?

Didn't know about the bump stop material differences. That's interesting stuff. Wonder what the stock stops are made of...

 

[lost won]

Bump stops...

Didn't know about the bump stop material differences. That's interesting stuff. Wonder what the stock stops are made of...

I dunno. But they work okay. (All stock suspension with A6s. Left front in the air). Been trying to convert to Street class from Stock. Ugh! No more Purple crack = no more traction. Like trying to learn to drive on ice. Help me find a Hoosier with 200 tread wear!

 

[csamsh]

ESP.....

 

[lost won]

No $$$ racing.....kinda sucks, but better than not.

ESP.....

Yeah, I considered it. Be just fine until someone came along with a properly set up ESP car. (lotsa $$$$).

At least in the old F Stock or F Stock R classes, I could get by with just the Purple Crack. And by the time you make the Pony into a spot on ESP car, it's not much of a daily driver any more.

Still having fun, just need to brush up on my ice driving...

Best Regards
John

 

[csamsh]

Yeah, I considered it. Be just fine until someone came along with a properly set up ESP car. (lotsa $$$$).

At least in the old F Stock or F Stock R classes, I could get by with just the Purple Crack. And by the time you make the Pony into a spot on ESP car, it's not much of a daily driver any more.

Still having fun, just need to brush up on my ice driving...

Best Regards
John

Everything you say...I have lived...and it's all true.

 

[Vorshlag-Fair]

I dunno. But they work okay. (All stock suspension with A6s. Left front in the air). Been trying to convert to Street class from Stock. Ugh! No more Purple crack = no more traction. Like trying to learn to drive on ice. Help me find a Hoosier with 200 tread wear!

YEEEEEE-HAAAAAWWWW!

 

[lost won]

Good One!

YEEEEEE-HAAAAAWWWW!

I dunno, Terry. I think this was more like YEEEEEE-HAAAAWWWW! :

River Running in Qualcomm stadium. With the wrong kinda boat......

 

[kcbrown]

Apologies. Responded to the wrong thread and can't delete this message.