Definition of Anti Squat 3

[Milanoguy]

Okay, first a dumb question and then hopefully a less dumb question, both about anti squat.

First question: Lots of references are made to anti squat in terms of percentage like 40% or 100%. But what does the percentage refer to? The total weight of the car? just the weight carried by the back wheels? If we have a 1,000 kg car with equal weight on each wheel(250 Kg's) and a 100% anti squat what is the weight on each rear wheel when the car is accelerating?

Next question: how does anti squat relate to a de Dion rear suspension, specifically the kind found in Alfa Romeo's cars like the Alfetta and it's successors the GTV6 and the 75/ Milano? Does the usual rule that limits the available anti squat in a I.R.S. suspension to about 25%, apply to a de dion suspension. How do you calculate anti squat in de Dion suspension?

To see a drawing of the de Dion suspension used in the Alfa, go to this link for some scans

Http://photos.yahoo.com/digestingtheinternet

 

[GregLocock]

If the SVSA IC is a above the wheel centre, and b forward of it then

%antisquat=100*a/b/(height of cg above ground/wheelbase)

antisquat does not affect the load transfer during uniform acceleration (actually there is a tiny effect)

I got that from Milliken.


What rule limits it to 25%?



Cheers

Greg Locock

 

[Warpspeed]

Greg is quite right about the weight transfer not being appreciably altered.

There is a far too prevalent myth that anti-squat pushes the wheels down increasing traction.

There is another prevalent myth that lots of rear squat pushes the wheels down increasing traction. hmmm....

Antisquat only offers some advantage in reducing suspension vertical travel (squat) thus sometimes improving the rear suspension geometry.

With an IRS that is set up for huge negative camber gain in bump, under acceleration, squat can cause massive camber change and even some track change at the contact patch. With wide flat tyres, this can lose a lot of traction.

With a di Dion rear end, there will be no geometry related problems, so I cannot really see the advantage of changing the original anti-squat geometry.

 

[Milanoguy]

Hi Guys

Thanks for the response. Okay here are my responses and some new questions

Responses
“What rule limits it to 25%?”
“…but a value of about 25% is about the practical limit. This compares to a well designed live rear axle rear suspension which often has over 100% anti-squat.” P.74 of Chassis Engineering by Herb Adams HP Books. Also “About 20% seems to be the maximum before we get into tire compliance problems.” P. 36 Tune to Win by Carroll Smith.


Questions

(A.) I have been operating under the assumption of “Squat good, Anti-Squat Better” meaning that squat increased the weight on the rear tires giving better traction and anti squat was better still, placing even more weight on the rear tires. I got this idea from the Adams book at p.64 he states
“Anti-Squat can counteract the squat force and it can be made strong enough to actually raise the rear of the car during acceleration…Because any force that can raise the rear of the car will need to have an equal and opposite force pushing against the pavement you can use anti squat to increase tire loading during acceleration.”

Warpspeed, am I understanding you correctly that the above quote is incorrect and that there is no relation between traction/grip and squat or anti squat.?

(b.) Greg, you stated “If the SVSA IC is…” I’m guessing that the IC term stands for Instant Center but what does SVSA stand for.?

Bye for now

 

[Warpspeed]

My understanding of weight transfer under acceleration, is that it depends on wheelbase and centre of gravity height. Something like a motorcycle is going to have a lot more weight transfer than an open wheel formula car.

Nothing you can do to the spring rates is going to change the static weight of the car.

Likewise changing spring rates is only going to change total dynamic weight transfer if the centre of gravity height changes as well (which it probably will).

Anti dive and anti squat use torque reaction to effective increase spring rates. It cannot alter the mass of the vehicle.

If you take an extreme case of a wheel-stand, with the front wheels completely off the ground. Total vehicle mass will then be on the rear tyres. Changing rear spring rates, or suspension geometry is not going to further increase the weight on the rear tyres beyond total vehicle mass.

 

[GregLocock]

SVSA IC is Side View Swing Arm Instantaneous Centre. The geometrical construction to establish that varies wildly with suspension type.

In a steady acceleration the increase in vertical load on the rear axle is always given by (m*a/g)CG height/wheelbase (or something like that). Note antisquat does not come into that equation, you don't even need to know which axle is driving.

Therefore we can assume that squat or antisquat provides a dynamic modifier. What goes up must come down. The instantaneous increase in vertical load will be paid for, some time later, by a corresponding decrease. Unless you are very cunning you will probably lose more than you gain, due to the non linear relationship between grip and vertical loading.

Hmm, OK, that is the theory. I'll put together a CarSim model and see what the effect actually is.

I agree there are practical limits to antisquat. But as someone once remarked (paraphrased) "building a succesful suspension is not just a process of applying rules of thumb to every aspect, the reasons behind the rules of thumb need to be understood"



Cheers

Greg Locock

 

[GregLocock]

http://www.geocities.com/greglocock

Look in the gallery, first two links.

cg height /wheelbase is about 0.203



Cheers

Greg Locock

 

[BillyShope]

I would only add that Greg's quote from the Millikens refers to an IRS. For a live axle, "a" would be measured from the tire patch.

(Actually, if you want to be picky and take the rear suspension unsprung weight into consideration, "a" would be measured from a point slightly below the wheel centerline for an IRS and slightly below...about an inch and a half...the tire patch for a live axle. And, the CG height would be the CG height for all but the unsprung weight at the rear.)

Because the no squat/no rise line, for an IRS, is displaced upwards by a distance equal to the effective radius of the rear tire, it is difficult...but certainly not impossible...to sufficiently angle the trailing links to achieve 100% anti-squat. Much easier with a live axle. Chrysler dragrace cars of the sixties, having leaf springs with a relatively short distance between axle and front eye, would "jump" off the line like frogs. Lots of favorable transient loading, but, as Greg correctly points out, what goes up must come down.

 

[GregLocock]

For a De Dion it is the same as an IRS, ie measured from the wheelcente

Cheers

Greg Locock

 

[MoreWing]

Milanoguy,

Let me take a shot at your question.

A.) I got the exact some idea from Herb Adams book when it first came out. Unfortunately, he's a quack. Normally, I wouldn't be some offensive, but when trying to talk to him about the issue he treated me like I had leprosy. So I don't feel too badly about it.

Warpspeed said that longitudinal load transfer is a function of wheelbase length and C.G. height. He's right on. There are 2 basic ways that one can choose to accept that load transfer. It can either be in the springs or in the control arms. When you accept the load in the springs, there is a certain time lag before the motion of the body can sort itself out. It tends to load the tire 'slowly'. Antisquat lets the suspension arms take the longitudinal load transfer. Suspension arms are made specifically not to defect. The load transfer happens instanteous. It loads the tire very quickly.

Here's how it works in practice. If you apply 100% throttle to a car that has no antisquat it will accelerate, pitch back, and then load the contact patch of the tires. It will all happen rather slowly. If you accelerate with a car that has 100% antisquat it will accelerate and at the same time load the rear tires. The front will lift, but the rear will not squat. The rear has essentially been locked solid. If you go over a small irregularity in the road, the rear tires will not follow the road and will probably get wheelspin. On an exceptionally grippy and smooth piece of pavement (launching pad on a drag strip) it might be fastest because it reacts so quickly, but no where else is it good.

Incidently, after reading Adams' book, I built a car that had a 100% anti-squat option. His book makes you think it would be great for rear grip on throttle. If you were turning the steering wheel at all it was impossible to apply any throttle. When you would apply the throttle the car would lock the rear suspension. That would not only take longitudinal load transfer, but it would take lateral load transfer. After the huge change in rear stiffness the car would either get very sideways or spin....almost on cue. To make the rear work well I ended up going to nearly 0% antisquat. It was much more controllable and had significantly better grip.

For my money, all 'anti's' should be used sparingly. It's probably reasonable to use a little antidive and a little antisquat, but think of them as spice in the soup. A little goes a long way. Concentrate on good camber and instant center control. Make sure bumpsteer is eliminated. Make sure caster doesn't change much. Then, if you want to use some anti to allow you to run softer springs, knock yourself out. Just don't build a car around it, and don't believe Herb Adams about it.

 

[BillyShope]

MoreWing, I'm afraid I can't entirely agree with your analysis. You say:

"If you accelerate with a car that has 100% antisquat it will accelerate and at the same time load the rear tires. The front will lift, but the rear will not squat."

I'm with you to this point. All of what the dragracers call "weight transfer" is being carried through the links and there is no change in spring loading.

Continuing: "The rear has essentially been locked solid. If you go over a small irregularity in the road, the rear tires will not follow the road and will probably get wheelspin."

But, I can't reach this conclusion. The situation, as far as the links are concerned, is essentially no different than when the car was sitting in the pits with the engine off. A force and moment balance exists in both cases. The only difference, during launch, is that an inertial force has been added to the mix. But, a force is a force. The suspension is no more "locked solid" during launch than it was in the pits. Other forces are simply additive.

As for an explanation of your experiences, I would have to look elsewhere. Specifically, I would strongly suspect that the additional loading resulted in a "stickiness" at certain suspension pivot points.

 

[Warpspeed]

MoreWing is quite right when he says the suspension effectively locks solid with 100% anti-squat.

If you have 100% anti squat, that suggests that there is zero suspension movement with a considerable rear weight transfer. That must mean the suspension rate has actually risen to infinity. In other words zero suspension movement with considerable extra weight applied.

A spring with infinite rate might be thought of as going solid. He is also right saying that with the suspension effectively non functional (locked solid), traction is going to be less than wonderful with normal road tyres.

Try making a triangular structure with needle roller bearings at the joints. It will be solid. Why ? because all movement is restricted by resisting forces in the links, even if the pivot points are free. Likewise torque reaction in your anti-squat linkage opposes all vertical movement.

If a wheel tries to move upward, that generates higher downward anti squat forces, and the whole thing effectively locks solid.

 

[GregLocock]

What you say may be true at a practical level with certain suspensions, but I can easily design a suspension that has any amount of antisquat and yet will not lock up.

For instance, let's use a twist beam De Dion. That is, a single trailing arm at each side of the car, joined by a torsional beam typically behind the wheelcentre. It is a perfectly sensible suspension, not that I've ever seen one.

The anstisquat is controlled by the SV angle of the trailing arms, yet there is no mechanism there to lock up. The wheel will always move up and down in response to bumps, won't it?


Cheers

Greg Locock

 

[Warpspeed]

For such a suspension to have anti-squat properties, would not the arms need to be at the front, and slope steeply upwards ?

To have 100% anti-squat those arms would need to be at a very steep upward angle indeed, probably not far off vertical !

If that were true, bump loads would be trying to compress those arms as much as swing them around the torsion bar axis. A pretty harsh setup.

If it was set up with the arms nearly horizontal with little or no built in anti-squat, it would probably work fine. If you could get the driveshaft past the torsion bar that is.

I have not seen one of those either Greg, but I have a good imagination.

 

[MoreWing]

OK, Warps and I are on agree on this one. Let's put some numbers on it and maybe I can explain my point.

From Greg's long. load transfer (I hate the term weight transfer) equation let's say we have a 3000# car with a 110 inch wheelbase and a 24 CG height. It's accelerating at 0.5 G. We transfer 327 pounds off the front axle and the rear has to accept it.

If we have 100% AS, the suspension links are pushing vertically with 327# of force to counteract that load transfer. In a perfect world, if you were to run over washboard pavement at this point, the car would soak up all the bumps in the springs, but the rear wouldn't have any more (or less) squat in it due to constant long. acceleration. We don't live in this world.

If you are anywhere near the limit of tire grip and go over rough pavement, the tires will see small bits of wheelspin due to the transient loadings. Even on a simple city street, these transients can be very large. The load in the suspension arms that are reacting to the load transfer are directly related to the longitudinal force on the tire contact patches (i.e. your AS force ultimately comes from the force of accelerating the car). As soon as you get a touch of wheelspin, the anti-squat force goes away. When the tire hooks back up, the AS force is restored. All this loading and unloading acts in a very short time period and is pretty severe, and it's a lot of force we're talking about 300 some odd pound going on and off the tires not to mention the loads from the bumps. When faced with rapid loading in pretty much any direction, a tire's first inclination is to slide or spin. They just don't react well to rapid inputs.

This is why I said that on a perfect launch pad, it might work, but in any environment I've ever worked in, high levels of anti-squat were a nightmare.

I only did the 100% anti-squat on one car. That was an option, but there were several other combinations available. It was a racecar and the links were monoball bearings. There wasn't excess friction in the system. Reducing the AS to 0 was the next step. 100% didn't work, so instead of screwing around with all sorts of other options, we just took it all out. This probably was not a perfect test. It was years ago, and I don't remember the exact numbers of the car. It did change motion ratios very slightly. It was as close to being a straight back-to-back test as I could do, though. No spring/shock/bar/setup changes. The rear of the car was massively better to work with. Yes, it had some squat. That didn't seem to matter. It would hook up nicely and was very docile to drive after that change. Mind you, it wasn't like a shade of gray change...I'm talking black and white.

One of the things I worried about on that car was braking. Excess AS will make the car prone to brake hop due to short SVSA lengths. Because of that, I had the brake system decoupled from the drive. There were seperate leading links that the calipers transmitted their forces through (with very little anti-lift ~10% or less). Braking of that car was always pretty good. I guess that was something I didn't overlook, but even at that, it was competely different from what I expected. In that respect, it was a very good learning experience even if not a particularly successful one.

 

[Milanoguy]

Hi Morewing

Thanks for your post. While I like the Adams book I have always had my doubts about it, for instance on page(p.35) he recommends cutting off one coil of a coil spring. All the other books, articles, about suspension modification I have read, condemn this practice.

Are you familiar with “How to Make Your Car Handle” by Fred Puhn ? If so, what do you think of it?

Bye for now

 

[BillyShope]

WarpSpeed says, "If you have 100% anti squat, that suggests that there is zero suspension movement with a considerable rear weight transfer. That must mean the suspension rate has actually risen to infinity."

No, it simply means that a force and moment balance exists without deflection of a spring. Any new forces...such as those which would occur if an irregularity in the road surface was encountered...would be additive, requiring a change in spring loading.

Greg's post should be read more closely. He explains that the locating links are trailing. Perhaps a more common design would be clearer. Consider two trailing links on each side. Each link pair would be in a plane parallel to the XZ plane. And, the members of each link pair would be parallel to each other. Finally, all links, in side view, are angled upwards, from the axle, at an angle with a tangent equal to Greg's "a" over "b." Since parallel lines meet at infinity, this would mean 100% anti-squat. This also means that all inertial loads, transmitted back through the links, are at the same angle. Any additional load, parallel to the Z axis and received at the tire patch, would require a change in spring loading, since there is no component of vector forces in the links available to balance it.

(Obviously, there is also nothing available to overcome the moment generated by the inertial force of the unsprung mass of the rear axle assembly, which is why I made the earlier comment about the no squat/no rise line actually passing a short distance below the rear tire patch...or axle centerline for an IRS or DeDion.)

 

[GregLocock]

This idea you have for compensating for the unsprung mass is neat. In practice we get around having to work it out by modeling the whole vehicle, and then running a straight line acceleration event, and measuring the pitch, and the wheel loads. We also run straight line decelerations to set the antidive/antilift compromise. It is probably worth pointing out that for real production suspensions the contribution to these antis by the bush compliances is very significant indeed, which starts to make hand calculations very difficult. Obviously the basic principle still applies.

I'd have liked to extend the CarSim model to include an acceleration over bump, sadly that program seems to be able to cope with bumpy roads, or horizontally accelerating vehicles, but not both at once.



Cheers

Greg Locock

 

[BillyShope]

We're of different generations, Greg. I'm still clinging to the simplified analyses I was taught half a century ago, before whole-vehicle modeling became available. This unsprung mass refinement of mine has, in effect, "fallen into a hole." A few of my old cronies might be interested, but most are too busy with their golf or grandchildren. Young people, like yourself, simply go to the whole-vehicle model. As for myself, I find an occasional suspension analysis more interesting than a crossword puzzle.

(Incidentally, I see a danger in the blind trust of a comprehensive model. I have, for instance, asked if the engine modeling software takes into account the "water hammer" principle in the modeling of the manifolding. Nobody seems to know. Everybody seems to be placing their faith in the consideration of flux at an increment of flow passage length, but I'm not at all certain that this adequately "captures" the water hammer phenomenon. In other words, once the model goes beyond the initial development phase, basic assumptions are no longer questioned. This could be very dangerous.)

 

[MoreWing]

Since parallel lines meet at infinity, this would mean 100% anti-squat. This also means that all inertial loads, transmitted back through the links, are at the same angle. Any additional load, parallel to the Z axis and received at the tire patch, would require a change in spring loading, since there is no component of vector forces in the links available to balance it.

------------------------------------------------------

Billy,

Regardless, If we have trailing links angled up at something like 45 degrees and a spring coming off them, you have to see that the trailing links are going to take a meaningful portion of a jounce load even if the car is coasting (no AS force). Since the tire is going to naturally move backwards (make the wheelbase longer) to accept the load it might not be as bad as I think it would be, but it's still going to be pretty severe.

I know what you're thinking in terms of any change in tire loading would require a change in spring loading, but let's do a thought experiment.

Assume the spring in this scenario is a link with pin joints, infinitely stiff. If this were the case, would you know say that the trailing link is taking none of the load? Of course not, a simple bit of statics shows you that is taking a verical load that corresponds to the longitudinal load it is being asked to take. This change depending on the angle of the arms, spring, loading, etc. but it's very real. The trailing link will take a portion of the verical load. I suppose if the load were directly in line with the spring, then you could argue that the spring takes all the load, but again, we don't live in this world.

I'm not trying to rain on a parade here, but what I'm saying is that large amounts of AS just do not work well in practical application. If someone wants to build one, more power to them. Just do yourself a favor and make sure that's not the only scenario you give yourself.

----------------------------------------------

As far as books go, I've seen Fred Puhn's book, but I don't own it. There are a bunch of books that cover pretty much the same material, but in slightly different manners. I think Greg has a list of good books to read, you might want to check that out. Van Valkenburg's book is pretty good. Smith's are pretty good. Puhn's has been in cirulation long enough that it must be pretty good. A fair bet is to read them all. Start with the oldest and easiest to read and progress into the more scientific. You'll get something from all of the them.