rec.autos.simulators

Please can anyone with the necessary background and experience please
help me and a friend settle an ongoing debate about the weight
transfer effects of ARBs?  The discussion started when I claimed that
fitting front and rear ARBs would not affect the handling balance of a
car provided that the _percentage_ increase in roll stiffness at each
end was the same.  My friend (an experienced test driver) says that it
is common knowledge that ARBs increase overall lateral load transfer,
and refers to respected sources for backup.

The trouble is, this seems to be at odds with the basic physics.  In
steady state, the weight transfer must be determined by track width,
C/G position and cornering radius/speed only.  A car that rolls less
will not lean its C/G so far towards the outside of the turn, which
will slightly *decrease* the weight over the outside wheels, the
opposite of what my friend argues.

So, where have I gone wrong?  And if I haven’t gone wrong, why
do so many people confirm that ARBs *increase* the lateral load
transfer?  For example, here is a quote that Simon sent me from
Carroll Smith’s “Tune To Win”...

“The greater the resistance of springs, the less roll will
result - but there will be no significant effect on the amount of
lateral load transfer because the roll couple has not been changed and
there is no physical connection between the springs on opposite sides
of the car.  The same cannot be said of the resistance of anti-roll
bars.  In this case, because the bar is a direct physical connection
between the outside wheel and the inside wheel, increasing the
stiffness of the anti-roll bar will both decrease roll angle and
increase lateral load transfer.”

This just seems plain wrong to me, as any attempt at lifting the
inside wheel by the ARB will just cause the car to “settle
back” onto that wheel until the loadings are in keeping with Mr.
Newton’s requirements for the cornering acceleration, C/G
location and track width.  This will lead to less body roll, which is
just what we expect.  All makes perfect sense to me – only
problem is, I’m arguing against a legend!  Help!

TIA

-= mike =-

You're buddy is correct. ARBs will add to the weight trasnfer. You just
about got it with the explanation of it picking the inside wheel up to help
keep the car level, but i think you're putting to much stress on the effect
of it keeping the c/g planted. In essense, ARBs are is trying to pick the
inside wheel up, but that takes weight off that wheel...the weight has to go
somewhere, so it gets transferred to the opposite side.

The way i learned to keep it straight in my head was to think of a go-kart
with no suspension and a solid rear axle. If we could apply enough lateral
force to tip the kart on two wheels, what would we have just done? We put
all the weight on the two wheels remaining on the ground. Take the same kart
only this time instead of the rear axle being a rigid part of the chassis,
free it up with only a single swivel point in the axle's center connecting
it to the chassis and apply the same lateral force to tip it on it's side,
this time both wheels will remain planted and we'll still have weight on the
inside wheel. Extremely simplistic look at it, but i think it's a good way
to "see" it.
--
In memory of DE,
- Will
- M1MS
- http://mach1.simhq.com

you are absolutely correct. The ARBS might decrease body roll, and as
such responsiveness of the car, but, given a fixed lateral acceleration,
and under the assumptions you put forward, the total weight left-right
weight transfer will not be affected, as that is just a function of the
c.g. height and track width.

If you, however, are using different ARBs, what will change is the
relative front to right load transfer, meaning that, while the total
left-right transfer is the same, the front tire may take a larger part
of it should the ARB be relatively stiffer at that end. I think the load
transfer in the  quote you used refers to only the rear or only the
front part of the car, and not the car as a whole.

-Gregor

side,

No, sorry.  If you apply the *same* lateral force, at the *same*
height above the ground, it will still lift the inside wheels.  You
can't 'lose' force except by using inertial dampers (a la Star Trek)
;-)

Jonny

Agree with Gregor, but note that adding ARBs will increase the roll
stiffness which will make the load transfer more *rapid* during
transients...

Jonny

I think you missed the point of the explanation, or i didn't explain it very
well. See, the first time, the axle is rigidly bolted to the chassis...the
axle doesn't move. If you could then mount the rear axle by a single
mounting point in the center and allow it to swivle around that point this
would leave the axle free to stay planted squarly on the ground as the kart
tipped over. Hard to explain apparently...the book i read it from dedicated
a whole chapter to explaining the effect of weight transfer, so i can hardly
go into depth in a mere paragraph.
--
In memory of DE,
- Will
- M1MS
- http://mach1.simhq.com

But in terms of the MAXIMUM transfer of weight from inside to outside
wheels, you are right. ARBs do not effect the factors that determine the
maximum lateral weight transfer which are track width, center of gravity
height, and roll axis height, at a given corner speed (lateral
acceleration).

True, but a negligible amount.

Looks like poor editing. Seems to be contradictory unless the the last
sentence is referring to increasing only one (front or rear) ARB rate, in
which case the lateral weight transfer will increase only at that end of
the car, reducing the weight transfer on the other end, but not changing
the total weight transfer of both inside wheels to both outside wheels.

Not to worry. You and the legend are in total agreement.

Marty

the kart

But you wouldn't be *able* to apply the same force, since the kart
would simply end up spinning around this pivot in a continuous barrel
roll.  If it had limited travel, it would hit the stops before the
force built up and *then* it would lift wheels.

Ultimately, any force applied to the vehicle (including, for these
purposes, imaginary ones like centrifugal force) has to be resisted by
the tyres.  There's just no way round this.

Jonny

  I agree with your claim.  As (I think it was Marty) pointed out though, the
time it takes for the load transfer to happen would decrease, but the handling
balance and total weight transfer would remain identical.  The first thing that
comes to mind that would mess all this up is that with increased roll
stiffness, you may end up with a different tire camber situation that could
change the handling balance, but if your friend and you are not including this
in the argument, I agree with your claim about equal load transfers.

  >My friend (an experienced test driver) says that it

  As the others pointed out, increasing anti-roll stiffness at only one end of
the car will do exactly what your friend says, but doing it to both ends the
same amount will keep everything identical as you argued.  This is a common
argument.

  Exactly right, although as someone else pointed out, it may be a negligable
effect and unrelated to the point your friend seems to be arguing anyway.

 >So, where have I gone wrong?  And if I haven t gone wrong, why

  This may interest you.  From the same book, page 36:

"  One of the most widespread misconceptions in racing is that the amount of
load transfer taking place is directly related to chassis roll.  Two opposing
theories are revalent:

 (1)  The car that rolls a lot transfers more load and develops more cornering
force.

 (2)  The car that is strongly restricted from rolling doesn't transfer as much
weight and so develops more cornering force."

  Neither of these exactly pick out what you're arguing about, but close.  

  A couple of paragraphs before this, Carrol Smith lays out the formula for
lateral load transfer:

Load transfer = (Lateral Acceleration(g) * weight(lbs) * c.g. height(inches)) /
track width (inches)

  Where are the anti-roll bar settings and spring constants?  Not there,
because they don't matter when looking at overall load transfer as you
described.  That only matters when the front/rear settings are different.

  >This just seems plain wrong to me, as any attempt at lifting the

  You've got it right.  If the car generated more load transfer than dictated
by cg height, track width, and cornering acceleration, it would continue to
roll forever.  The torques are in balance once steady state is reached if the
car has no rolling velocity.

  The only thing I can say on your friend's behalf is that he is right if he
means at some specific roll angle, total weight transfer is higher with the
ARB's.  At say 3 degrees, the load transfer is higher, for instance.  However,
it doesn't sound to me like that's what the argument is about, but rather, it's
about total weight transfer through a given corner, letting roll angle fall
where it may.

Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://PerformanceSimulations.Com

--
In memory of DE,
- Will
- M1MS
- http://mach1.simhq.com

Will, which book did you read for the weight transfer stuff? I'm glad
in a way that you picked the kart example, as I have used that as a
reference point in recent discussions. You mention the effect of
tipping the vehicle up onto the outside wheels, but that will happen
regardless of roll angle when there is enough cornering force to "pull
the wheels out from under the car". Body roll only plays a part
insofar as it moves the C/G towards the outside, which will increase
the tendency to overturn. I'm willing to bet your book contains a
diagram that shows the various forces and moments involved in this
process.

Gregor, Jonny, Marty and Todd, glad to see you all think I'm on the
right track. Also glad that people mentioned the transient effects.
There was a much longer message that I was originally going to post
that included some of this stuff, but Simon reviewed it and said it
was overlong so I cut it down to the raw essentials. These were my
original comments on the transient behaviour changes if you simply add
"balanced" ARBs to an existing setup:

1. Initial transient weight transfer increased on turn-in, because the
higher roll rate means that the outside wheels will load up harder and
faster in comparison with the softer setup.  This would also be true
if the same roll rate increase was obtained by upping the main spring
rates, however.
2. If the suspension is under-damped in roll, you get a delayed weight
transfer effect to the outside following turn-in as the springs halt
the outward swing of the body. A softer setup allows the body to swing
further, so that it builds up more momentum, while the stiffer
ARB-equipped setup will exhibit this effect to lesser degree.
3. Transient effects resulting from braking/tractive forces while
cornering will follow the same basic patterns.  Dabbing the brakes
mid-corner will load the outside front more suddenly in the
ARB-equipped car for example.

This last point is interesting, because Simon is willing to agree on
the pure steady-state situation but says that it is never applicable
in real life because there is almost always some braking or
acceleration going on, even if a minute amount, which alters the
situation. In his words: "as soon as weight shifts forward or
backwards the ARB at the unladen end will cause more weight transfer
even if the roll resistance at both ends are the same."

I don't agree with that. I think if you achieve the same roll rates by
means of the main springs, you will see the same corner loadings,
because front/rear load transfer is governed by C/G position,
wheelbase and longitudinal force. The only difference is with ARBs and
softer springs, you will have more pitch movement which will lead to
the similar transient weight transfer effects to those we have already
agreed on in the roll situation. I.e., an increase in the "negligible"
movement of the C/G due to pitching, plus slower load transfer and
perhaps (if we're low on damping) a higher peak magnitude. Camber
angles and other aspects of the geometry might also be influenced by
the larger wheel displacements, but I think it's fine to treat those
as a separate issue for now.

Simon, I think the "accepted truth" about ABRs and weight transfer is
out of context here, because it is only true if you increase the rate
at just ONE END. This is a point I tried to make earlier, and has been
confirmed by others here.

Guys, who here is a dynamics engineer with practical experience who
can maybe give some anecdotal evidence or other information that
establishes the truth of all this in real-life situations as well as
just theory?

-= mike =-

I'm Mike's friend (who's getting a slating!! :-))

The main cause of the difference of opinion between Mike and I has
been the difference between the isolated and incredibly limited
discussion of true steady state cornering and the application of the
effects of ARBs to the HUGE subject of the entire handling package.

In true steady state (if such a thing exists in the real world) I
agree that load transfer is not a function of roll resistance except
to say that the *difference* between front and rear roll resistance
will obviously affect it.  But..

Roll resistance describes the amount of force required to roll the
sprung weight a given amount.

So if you had exactly the same roll resistance (through ARBs) front
and rear and perfect 50/50 weight distribution, horizontal roll axis
and true steady state cornering..  The car would roll equally at both
ends, and weight transfer would be equal at both ends..

But..

Move the load rearward (acceleration, etc) and you will roll the rear
suspension more than the front = front inside wheel lifts = more load
transfer at the front.

The more roll resistance created in this way, the more greater this
effect and the greater the load transfer.

Handling is unfortunately (or fortunately otherwise it would be
boring) substantially more complicated than steady state cornering.

All the best

Simon

Welcome aboard ;-)

Aha - I've just spotted the aberrant assumption!  The chassis is
generally regarded as torsionally stiff for the purposes of suspension
analysis, so *roll angle is the same at both ends*.  Wheel lifting
happens when one end either has less travel than the other, or is
significantly stiffer in roll such the the load on one wheel drops
below zero (in practice the latter doesn't happen with coil springs,
since they're not designed to load in tension).

This assumption may actually be false for '60s GP cars, but I don't
know whether or not it holds for GPL... ;-)

HTH,
Jonny

So with ARBs fitted (even if they are equal) at any time that the
sprung mass is rolled towards one single wheel, the oposite wheel
loses load to it's adjacent axle partner.  The stiffer the arb the
greater this transfer.

All the best

Simon

It is the torsionally stiff properties of the chassis which *cause*
this.  Although, the fact that in real life they are not completely
stiff actually REDUCES this effect.

I was not saying that the roll *angle* was more at the rear than the
front, but that the *forces* were increased at the rear.  As the front
has less forces on it, but is still being forced to roll (by the
chassis), it will effectively sit up higher on it's laiden spring and
the ARB will lift the unlaiden wheel (admitedly not necessarily off
the ground, but load will be lifted off that wheel and transfered onto
the laiden one).

The problem with all this is that there are very few hard and fast
rules in vehicle dynamics.  I know of / have been involved in
situations where the spookiest things have a happened...

Sierra Cosworth Race car.  1000lb springs on the front... dire
understeer.  Fitted 1200lb springs and the understeer went away????
Now that information isolated would be argued.  If I posted a message
saying "how can I reduce the understeer of my race car" and someone
said "increase the front springs", they would be laughed at.  But they
would have also (in this case) have been right.

The biggest trap you can fall into is the one where you believe you
understand it all.  There are very few hard and fast rules as there is
such a chaotic aspect to ride and handling once a car is on the move
in the real world.  The maths and formulas are invaluable for making a
first stab at it, and then for ideas as to what to test in order to
tweak it.  But after that, there are accepted truths which generally
work, and LOTS and LOTS of testing.

Cheers

Simon