Trying to Learn squat....

[Grant 302]

Hope you're feeling better! Our whole family got hit this winter.

Do you still have a lot of prep to do for your first event? Sounds like everything is working for street use.

 

[Mad Hatter]

Feeling alive now! I just have a few suspension details to finish up. With luck next week the car goes into a friends shop for the extension of the roll bar. Will send pics when its done.

 

[Grant 302]

I've been driving the Boss and GT back to back lately and have noticed differences how much better the Boss puts down power in and coming out of turns. I've thought in the past that it was mostly due to differences from the Torsen vs. T-lok...but I'm thinking now, that the squat has been a larger part of it. Since I've been paying attention to it, it's very noticeable how the GT lifts and the Boss squats. GT certainly launches and rolls-out better.

 

[Mad Hatter]

Just to put the thought in this Squat thread... As I was riding shotgun on a friends first track venture in a 2013 GT with H&R springs and nothing else.... You could see the LCA (Lower trailing arms for the Word Police!) where way into Squat territory.

Since he was new to the track I was sure that he would have some tail happy activities but the car stayed very planted despite having a auto box that made all the changes in the wrong places..

Yes the car did tend to go wide in the curves but that could be the street alignment as well as the Roll Understeer.

 

[Grant 302]

Would be nice if one of us could do some back to back testing on this. Trouble might be learning to drive each setup reasonably well. All these years of learning to drive smoothly works well for momentum cars, but probably not the fastest for our current Mustangs.

 

[Mad Hatter]

Some one has two mustangs!! On the 23/24 I will have two days at the track so I can try swapping the LCA angles and see what my times give....

 

[Norm Peterson]

Just to put the thought in this Squat thread... As I was riding shotgun on a friends first track venture in a 2013 GT with H&R springs and nothing else.... You could see the LCA (Lower trailing arms for the Word Police!) where way into Squat territory.

But did you see where the upper was pointed? It matters just as much.


Norm

 

[Mad Hatter]

But did you see where the upper was pointed? It matters just as much.


Norm

I would need to get very up and personal!! Its kind of like asking hows your prostate!!! But I would guess that it would be about 14 degrees would have to get it near a pit. None at our tracks.

Sent from my SM-G900M using Tapatalk

 

[Grant 302]

But did you see where the upper was pointed? It matters just as much.


Norm

Certainly matters for figuring out exact AS/% lift values, but I think it can be reasonably ballparked using 'stock' geometry lowered by the H&R rear 'drop' values.

 

[Norm Peterson]

I would need to get very up and personal!! Its kind of like asking hows your prostate!!! But I would guess that it would be about 14 degrees would have to get it near a pit. None at our tracks.

For LCA chassis-side pickups 2" lower than the axle-side, actual car measurements at stock height and a spreadsheet to figure what happens when you lower the car enough to put that 2" difference in finds that the UCA would be inclined downhill toward the chassis at ~19.7°. Anti-squat would be about 29%. Even though the LCA points downward, the SVIC is still about 2.7" above grade and about 54" ahead of the rear axle centerline, so it's still an "anti-squat" situation rather than "pro-squat". Just not a very strong "anti" value.

Here's the free-body diagram, simplified for purposes of illustration. Force from the tire to the ground, lets give it a 1000 lb value rearward (about 0.3g acceleration). For simplicity, let's put the LCAs 8" above the ground and the UCA at 16". Summing moments about the UCA axle-side pickup puts the LCA force at 2000 lbs (1000 lbs per LCA). Equilibrium requires the UCA force to be 1000 lbs rearward. Otherwise the car acceleration computed from the LCA and UCA forces would not match the car acceleration computed from the traction force at the ground.




Norm

 

[Grant 302]

Your car is in (R)everse.

 

[Grant 302]

so it's still an "anti-squat" situation rather than "pro-squat". Just not a very strong "anti" value.

I think the conventional terms and especially percentages all confuse people. 29% A.S. *is* 'pro-squat'.

 

[Norm Peterson]

The force shown at the pavement is the force from the tire on the pavement (trying to push the pavement backward). The other two forces are forces from the linkage to the car. It's one way of showing how force and moment equilibrium of the axle is satisfied, with the LCA force shown pushing the car forward. Showing it by force and axle torque at axle level is harder for me to draw and would take more than one picture to show clearly, so I chose the easier/quicker way.

Anti-squat is any resistance to squat happening, IOW any positive value. In the 29% case, the car squats about 71% as far as it would with zero load transfer taken through the suspension links (think infinitely long perfectly horizontal LCAs and UCA, putting the SVIC at "infinity").

"Pro-squat" would be a sort of "negative anti-squat", where load transfer through the linkage results in more squat than if there was no load transfer through them at all. The SVIC would have to be below grade and ahead of the rear axle or above grade and behind the axle.


Norm

 

[Mad Hatter]

Point taken! Here is what I guess that GT's IC angle is at.... Its still antisquat but very low.... Leaning towards Squat if you like.

 

[Grant 302]

I get what you did with the diagram and the 'reaction' forces. But that's almost certain to confuse people about how to use and draw free body diagrams, in my opinion.

I've said before that I dislike the conventional terms related to squat and anti-squat. The terms are somewhat misleading. Anything under 100% A.S. will squat. And our suspensions will dynamically increase squat (and decrease A.S. values) as weight is transferred. I believe this is a big point that many or most do not understand because of the terminology.

Add the angles and the spring reactions, and it becomes waaay more clear about what actually happens. There should be a torque reaction and a net force (1,000 lb. in your example), not a balance or equilibrium...otherwise the axle assembly doesn't *go* anywhere.

 

[Norm Peterson]

Once you've got the equilibrium satisfied, you can then see a net forward force to the chassis of 1000 lbs. Which is the net reaction to the 1000 lbs trying to push the pavement backward. Strictly speaking, the link forces would be reduced slightly because some force is required to accelerate the axle/rotors/wheel/tires forward at the same rate as the rest of the car.

I think there's some difficulty resolving % antisquat against the fact that anything less than 100% antisquat still allows some squat to occur. Antisquat is not a binary some squat/no squat parameter; it's a variable that can be anywhere from less than 0% (the axle starts being carried by the car) to more than 100% (where the tail actually rises - a phenomemon that's readily visible at the 5th wheel area of an 18-wheeler's tractor when it pulls away from a stop).

Maybe if you look at %antisquat as being the % or rearward load transfer carried geometrically and compute the amount of squat from the remaining portion of rearward load transfer divided by the total rear axle spring rate it'd be easier to follow.

Incidentally, the OE antisquat geometry also starts out around 30%, drops a couple percent as the rear suspension compresses about 1.5", and then rises again. It reverses because the shorter UCA changes inclination faster than the longer LCAs do (and the side view swing arm starts getting pretty short pretty quickly). It looks something like this ↓↓↓




Norm

 

[Grant 302]

Maybe if you look at %antisquat as being the % or rearward load transfer carried geometrically and compute the amount of squat from the remaining portion of rearward load transfer divided by the total rear axle spring rate it'd be easier to follow.

That might produce reasonable results. Would be nice if it were that simple.

I've done a fair amount of geometrical analysis, but never approached it with the intent to produce theoretical values. It was always easier to quantify and validate changes with timeslips/60' times.

Making models for road racing might prove a bit complicated.

 

[Norm Peterson]

I think theoretical values can provide insight before you make any changes, and having insight tends to cut down on the number of iterations you end up making with the car on ramps and jackstands and you're underneath it with the wrenches. Here, you probably want squat to start adding anti-squat percent from the get-go rather than start out by throwing some of it away. Slight lowering has been identified as being helpful for launch issues by other people, and I think the above plot is consistent with that finding.

Of course, you still have to drive it to confirm that what you did was at least in the right direction. Even the mfrs don't expect final answers to come solely from their engineering departments' computer analyses.

A road course simulation even for just rear suspension geometry taken in isolation is a good bit more complex. Axle roll steer varies with ride height (the easy part) and most likely with the amount of roll (not so easy), and once you get those modeled you'd still have to merge whatever you got there with the anti-squat stuff (which won't be equal side to side due to that roll). I haven't even attempted this.


Norm

 

[Grant 302]

Anti-squat is any resistance to squat happening, IOW any positive value. In the 29% case, the car squats about 71% as far as it would with zero load transfer taken through the suspension links (think infinitely long perfectly horizontal LCAs and UCA, putting the SVIC at "infinity").

"Pro-squat" would be a sort of "negative anti-squat", where load transfer through the linkage results in more squat than if there was no load transfer through them at all. The SVIC would have to be below grade and ahead of the rear axle or above grade and behind the axle.

This highlights the problem with the definition of and parameters for "anti-squat". Any IC below the lower arm connection point(s) have no "anti-squat" value or contribution. Or at least *I* can't think of a way any for forces to resist squat below that. Your horizontal trailing arm example shows it well enough. There are no components from the arms resisting the inherent squat properties with that setup. This condition could or should define 0% anti-squat, *not* an IC at ground level at the front axle.

Antisquat is not a binary some squat/no squat parameter; it's a variable that can be anywhere from less than 0% (the axle starts being carried by the car) to more than 100%

And again, that's the problem with the terms and definitions. While not binary, there is a change of conditions that occur exactly at 100% AS/'neutral'. Above it, the rear will lift and below it the rear will squat.

I'd still have to see a diagram to show me how the axle can be 'carried' by the car. If you're talking about the 5th wheel rising while pulling a trailer, I think that's just the reaction from the high CG load it's pulling.

I think theoretical values can provide insight before you make any changes, and having insight tends to cut down on the number of iterations you end up making with the car on ramps and jackstands and you're underneath it with the wrenches. Here, you probably want squat to start adding anti-squat percent from the get-go rather than start out by throwing some of it away. Slight lowering has been identified as being helpful for launch issues by other people, and I think the above plot is consistent with that finding.

Of course, you still have to drive it to confirm that what you did was at least in the right direction. Even the mfrs don't expect final answers to come solely from their engineering departments' computer analyses.

While I like numbers and calcs, I don't think they provide feedback like data recording, lap times and 60' times. Understanding of the directions of the reactions *does* help with tuning damping which is far more important (IMO). I've even used video to see the lift/rise reactions and then help tune shocks for drag launch remotely.

A road course simulation even for just rear suspension geometry taken in isolation is a good bit more complex. Axle roll steer varies with ride height (the easy part) and most likely with the amount of roll (not so easy), and once you get those modeled you'd still have to merge whatever you got there with the anti-squat stuff (which won't be equal side to side due to that roll). I haven't even attempted this.

In concept, that's *exactly* what I've been talking about since the beginning of this thread!

I'm still adjusting from Hawaiian time...forgive me if any of the above doesn't make sense!

 

[Norm Peterson]

This highlights the problem with the definition of and parameters for "anti-squat". Any IC below the lower arm connection point(s) have no "anti-squat" value or contribution. Or at least *I* can't think of a way any for forces to resist squat below that. Your horizontal trailing arm example shows it well enough. There are no components from the arms resisting the inherent squat properties with that setup. This condition could or should define 0% anti-squat, *not* an IC at ground level at the front axle.

Absolutely correct.

But that particular sketch was only intended to show the force equilibrium side of the discussion. The SVIC is "virtual" in this very specific situation - out there "at infinity" in a purely horizontal direction for 0% antisquat.

If you start with that sketch and drop only the chassis-side pickup elevations down by the same amount (lowering the ride height a little), the UCA will change slope faster than the LCAs. The SVIC now has some real location and anti-squat now has some non-zero value (the situation that starts out with all of the arms/links horizontal actually goes into "pro-squat" with lowering).

And again, that's the problem with the terms and definitions. While not binary, there is a change of conditions that occur exactly at 100% AS/'neutral'. Above it, the rear will lift and below it the rear will squat.

Yes, it can get confusing when there's that 100% numerical difference between zero anti-squat as a parameter and zero squat as an observable phenomenon. I usually work with anti-squat and have to stop and think that the amount of squat is 100% minus the antisquat%. This still works for "pro-squat" or negative anti-squat percentages and even for antisquat% above 100% (ex. 100% - 125% being -25% squat which would be 'rise').

I'd still have to see a diagram to show me how the axle can be 'carried' by the car. If you're talking about the 5th wheel rising while pulling a trailer, I think that's just the reaction from the high CG load it's pulling.

In reality, it would only be a momentary unweighting of axle weight off the ground. I think you'd have to look for a very brief decrease in vertical tire deflection down at the ground before the real, physical weight over the rear axle drives it back down. Kind of self-correcting.

While I like numbers and calcs, I don't think they provide feedback like data recording, lap times and 60' times. Understanding of the directions of the reactions *does* help with tuning damping which is far more important (IMO). I've even used video to see the lift/rise reactions and then help tune shocks for drag launch remotely.

Agreed, feedback has to come from actually driving it. Numbers and calcs - if they're good enough and you can visualize what they mean - can point you the right direction before it's time to build (let alone drive). Feedback can't tell you that you went in the wrong direction until you've already built it and tuned it wrong.

OE mfrs and maybe a few others use $$$$$-level software such as ADAMS for their suspension modeling. Well beyond Suspension Analyzer-level tools.


Norm