A Favor From Those With 4link Software

[BillyShope]

I'm asking a small favor from those of you who have 4link software. I'm curious as to whether these packages can also handle a 4bar suspension. In other words, if the upper and lower links are parallel, do they give the right answer or do they "blow up."

Here's the input to try:


18 inch center of gravity height
100 inch wheelbase
16.935 inch upper and lower link lengths (yes, it's necessary to carry it out to the thousandths of an inch in order to force parallelism with the 100% antisquat line)
11 inch and 17 inch vertical to rear pivots
14 inch and 20 inch vertical to front pivots

If the software is able to handle 4bar suspensions, it will tell you that the antisquat is 100 percent and the instant center is an infinite distance forward and an infinite distance up.

Thanks for your time. Appreciate it.
http://www.racetec.cc/shope

[astroracer]

My Performance Trends program doesn't blow up but it requires a lot more info to get an accurate output. I used a 22" front tire and 30" rear with a 50% wieght bias. The output is close enough to your numbers to not matter.
Mark

[Norm Peterson]

It should be me thanking you, Billy.

Your sample problem has provided a measure of verification for a number of rear suspension spreadsheets, one of which I was updating only yesterday.

The sheet does choke when I use the exact dimensions that you provided, due to occurrences of divide-by-zero when the link slopes are identical (making their difference equal to zero).

That's a known issue, and making the arm lengths differ by only 0.000000000001" cleans everything up and does provide 100.00% A/S (to two decimal percentage places). The IC coordinates go to X=5.5E14 and Z=1E14; ~9 billion miles/95 astronomical units out being close enough to ∞ for me :) .

I get 99.9970% to four places, and working backwards suggests that if I were to use 16.9345144" instead (sqrt((100/6)^2 + 3^2)). that the A/S result becomes 100 and all zeros a little further out (99.99999997% → 100.0000000%).


Norm

[BillyShope]

After reading your post a few times I have to ask what your
understanding of 4 bar and parallel 4 bar is... These terms
describe the two suspensions looking from the top. A
parallel 4 bar has all four bars running directly forward,
"parallel to the car centerline and themselves. An "angled
" 4 bar has at least two bars, usually the uppers, angled
in or out as compared to the center of car or the lower
bars. In reading your post I think you are asking about
the side view of the suspension. A parallel 4 bar will not
necessarily be parallel in the side view, only in the top
view...

The software doesn't care where the links are "cross-car".
My Performance Trends program doesn't have an input for
"cross-car" locations so all of the data is generated on a
2d plane so the bars are parallel already. In fact the
inputs are for one set of bars only. I cannot input the
points for two sets of bars.
Your suspension doesn't really care were the points are
in the "cross-car" locations either. As long as the
suspension is free to articulate without binding an angled
4 bar should work along the same lines as a parallel 4 bar.
All the angled bars do is eliminate the need for a panhard
bar.Mark
Mark, each branch of motorsports seems to have its own
esoteric language. I was using terms peculiar to dragracing.
Unless the adjective "triangulated" is added, the term
"4link" is understood...among dragracers...to refer to a
configuration where the 4 links are parallel in plan, but
each is adjustable in side view.

The term "4bar," on the other hand, refers to a
configuration where the 4 links are still parallel in plan,
but, in addition, they are constrained to side view
parallelism...throughout suspension travel...by their end
brackets. Since there is no link bind with this
arrangement, it is often found in cars that are also
driven on the street, but, without the adjustment of a
4link, it is not as popular...for serious racers...at the
dragstrip.

Now that we're speaking the same language, you can
understand why I would consider your software to have
failed, since the parameters you changed have nothing to
do with the calculation.
http://www.racetec.cc/shope

[BillyShope]

The sheet does choke....
Norm
Norm, I avoid this in mine (Page 13) by first checking for parallelism and, when it is found, by switching to a different solution form. My source code is available for viewing.
http://www.racetec.cc/shope

[Norm Peterson]

I could probably add a couple of 'if' tests to the slope cell formulae to do the same sort of thing (probably would just add a tiny slope if parallelism was detected). Maybe I will, sounds easy enough.

And do the same sort of thing for plan view parallelism, as the roll steer calc side of it doesn't like parallel plan view link projections any better.

FWIW, what I added was something to account for ride height changes caused by either lowering the car or just getting into it. You typically make all your measurements with nobody in the car, and quite possibly before you've gone ahead and lowered it any (because you're using the sheet to help choose things). When roll steer varies by nearly 6%/in and A/S varies by 3%/in or so, it can make a difference.


Norm

[Bryce]

I could probably add a couple of 'if' tests to the slope cell formulae to do the same sort of thing (probably would just add a tiny slope if parallelism was detected). Maybe I will, sounds easy enough.

And do the same sort of thing for plan view parallelism, as the roll steer calc side of it doesn't like parallel plan view link projections any better.

FWIW, what I added was something to account for ride height changes caused by either lowering the car or just getting into it. You typically make all your measurements with nobody in the car, and quite possibly before you've gone ahead and lowered it any (because you're using the sheet to help choose things). When roll steer varies by nearly 6%/in and A/S varies by 3%/in or so, it can make a difference.


Norm

What about progressive antisquat. With over 100% A/S as the suspension lifts you get more and more A/S. The opposite is true for less than 100%

[Norm Peterson]

The opposite is true for less than 100%
Not sure about that. Link lengths are much shorter than wheelbase, so if the whole car moves up an inch, the tangent of the link inclination rises faster than the ratio of CGz to Wheelbase does.

Having plots from -2.5" to +2.5" does show a bit more than what's going on at just the static ride height . . . you might not expect some things that can happen with unequal length uppers and lowers if the inclinations and lengths are "just so", and in some cases seeing how the A/S varies may be nearly as valuable is knowing what the number is.

Seeing what happens as the car pitches under acceleration/braking would be probably more useful still, but that's a task for another time. Exterminating a couple of #DIV/0! bugs is enough for one day.


Norm

[astroracer]

Mark, each branch of motorsports seems to have its own
esoteric language. I was using terms peculiar to dragracing.
Unless the adjective "triangulated" is added, the term
"4link" is understood...among dragracers...to refer to a
configuration where the 4 links are parallel in plan, but
each is adjustable in side view.

The term "4bar," on the other hand, refers to a
configuration where the 4 links are still parallel in plan,
but, in addition, they are constrained to side view
parallelism...throughout suspension travel...by their end
brackets. Since there is no link bind with this
arrangement, it is often found in cars that are also
driven on the street, but, without the adjustment of a
4link, it is not as popular...for serious racers...at the
dragstrip.

Now that we're speaking the same language, you can
understand why I would consider your software to have
failed, since the parameters you changed have nothing to
do with the calculation.
http://www.racetec.cc/shope

Hey Billy,
I wrote that first response without realizing "who" I was talking with... You can see by my edited initial post I figured it out quickly. But not quickly enough to keep you from getting a copy of my ramblings...
I had never plugged in any dims for the old non-adjustable four bar suspensions before (yes, I do know what you were looking for and I changed no parameters that effected the outcome). Never considered it as it is, like you say, not a desirable set-up for hi-performance use.
Thanks for the insight.
Mark

[BillyShope]

Mark, I figured you must have changed your mind, but I used your comments as an opportunity to clear things up for those who might remain uncertain as to what I meant.

Instead of using your name, I should have simply expanded on my definitions.

Sorry. Didn't mean to offend.
http://www.racetec.cc/shope

[Norm Peterson]

Billy,

A little off-topic, but you may be able to use some of this input information on your page 21. Measurements were taken to bushing centers with the vehicle unoccupied, only a few lbs of stuff in the trunk, fuel level unknown. They should be good to within 1/8".

X is measured from the rear axle centerline, positive forward
Y is measured laterally from the vehicle longitudinal centerline
Z is the measurement above grade


3LinkPHB19
J & N Performance (06 May 2010)
Filename/Descr. 2008 Mustang (approx)
INPUTS
LCAchassisX . . . . . . . . . . . 18.38 in
LCAchassisY . . . . . . . . . . . 23.63 in
LCAchassisZ . . . . . . . . . . . 7.88 in
LCAaxleX . . . . . . . . . . . . 0.00 in
LCAaxleY . . . . . . . . . . . . 22.13 in
LCAaxleZ . . . . . . . . . . . . 8.25 in
CGHeight . . . . . . . . . . . . 21.00 in
Wheelbase . . . . . . . . . . . . 107.10 in
RearTireRollingRadius . . . . . . 13.00 in
RideHeightChange . . . . . . 0.00 in
LCAaxleZChange . . . . . . . . 0.00 in
FrontSeatLoad . . . . . . . . 185 lb
TotalWheelRate . . . . . . . . 600 lb/in

UCAchassisX . . . . . . . . . . . 8.50 in
UCAchassisY . . . . . . . . . . . 0.00 in
UCAchassisZ . . . . . . . . . . . 19.38 in
UCAaxleX . . . . . . . . . . . . . . 0.00 in
UCAaxleY . . . . . . . . . . . . . . 0.00 in
UCAaxleZ . . . . . . . . . . . . . 20.63 in
PHBchassisX . . . . . . . . . . . . . -5.00 in
PHBchassisY . . . . . . . . . . . . . 21.00 in
PHBchassisZ . . . . . . . . . . . . . 12.75 in
PHBaxleX . . . . . . . . . . . . . -7.00 in
PHBaxleY . . . . . . . . . . . . . -21.00 in
PHBaxleZ . . . . . . . . . . . . . 11.25 in


Norm

[BillyShope]

Thanks, Norm. I'll put these in as default values.
EDIT: It's been done.
http://www.racetec.cc/shope

[BillyShope]

Norm, please take another look at Page 21. I've pretty much completely redone it. When I added your dimensions, I realized that I hadn't a clue as to how I had solved the problem before. As an engineer, I see the computer program as a tool and not...as would the software designer...an end in itself. As a result, my source code documentation ranges from sloppy to non-existent.

So, I started with an equation set and went through the derivation again. Each time I do one of these, I gain a little more insight into the overall problem. Back in the fifties, when I did the first derivation for the Ramchargers 3link, I carried along a bunch of equations in the set which didn't contain dynamic terms and...without a computer...it was a nightmare!

Anyway, I have confidence in Page 21 now. All the equations check. I've also added a percent antisquat input and link loads for 1G acceleration.
http://www.racetec.cc/shope