Developing Suspension
2 views
Skip to first unread message
jaho...@googlemail.com
Feb 17, 2008, 3:50:56 PM2/17/08
to Puk7
Posted this onto LocostBuilder and LocostUSA in an effort to gain some
constructive criticism of the approach:
Suspension Design - how to?
I've been iterating through different ways of attacking the design of
the suspension. I've read most of the reference books excepting Kimini
one and the Miliken one (Santa proved to be a disappointment on that
front - but thanks for the jumper mate). Rather than continue to lie
awake in bed pondering this, can I ask for some feedback:
My thoughts are that you'd want the tyre to be as close to upright as
possible under conditions of maximum load and it would be ok to
compromise on this when the tyre was under little load.
If that is an acceptable idea (pipe up anyone with some experience)
then the following ought to hold true:
1. When cornering the outside wheels would be upright, or leaning in
slightly at the top like a motorbike type (but definitely not leaning
out) AND it would be an acceptable compromise for the inside wheel to
be less than upright because it is lightly loaded.
2. Under brakes the front tyres are most heavily loaded, so aim to
keep them upright or leaning inwards at the top slightly (but avoid
leaning out) AND it would be an acceptable compromise for the rear
wheels to be less than upright - again because they're lightly loaded.
3. Under acceleration keep the rear wheels upright or leaning inwards
at the top slightly (but avoid leaning out) AND it would be an
acceptable compromise for the front wheels to be less than upright -
again because they're lightly loaded.
So if all that hold true then it is just the small matter (gulp) of
establishing the dip of the chassis at the loaded end under these 3
cases of maximum acceleration. If I know the target acceleration in
any direction (1.5G in all directions is about the best I would except
from a set of race tyres on car without downforce) then if I also know
the spring rates, the location of the CofG, the tracks and wheel base,
then it ought to be possible to calculate the roll of the chassis
under acceleration (acceleration being positive when the car gets
quicker, negative under brakes and lateral in a corner).
Knowing the dip of the chassis tells me the amount that the loaded
wheel has been compressed by. For the sake of argument if the car dips
50mm braking at 1.5g then I want the front wheels to be upright when
the suspension is compressed by 50mm.
Ideally I'd like the outside wheel in a bend to be compressed be the
same amount when achieving 1.5g of lateral acceleration. Though quite
how that is achieved is a bridge that is yet to arrive - maybe that is
where rollbars come in?
Ok fire away!
constructive criticism of the approach:
Suspension Design - how to?
I've been iterating through different ways of attacking the design of
the suspension. I've read most of the reference books excepting Kimini
one and the Miliken one (Santa proved to be a disappointment on that
front - but thanks for the jumper mate). Rather than continue to lie
awake in bed pondering this, can I ask for some feedback:
My thoughts are that you'd want the tyre to be as close to upright as
possible under conditions of maximum load and it would be ok to
compromise on this when the tyre was under little load.
If that is an acceptable idea (pipe up anyone with some experience)
then the following ought to hold true:
1. When cornering the outside wheels would be upright, or leaning in
slightly at the top like a motorbike type (but definitely not leaning
out) AND it would be an acceptable compromise for the inside wheel to
be less than upright because it is lightly loaded.
2. Under brakes the front tyres are most heavily loaded, so aim to
keep them upright or leaning inwards at the top slightly (but avoid
leaning out) AND it would be an acceptable compromise for the rear
wheels to be less than upright - again because they're lightly loaded.
3. Under acceleration keep the rear wheels upright or leaning inwards
at the top slightly (but avoid leaning out) AND it would be an
acceptable compromise for the front wheels to be less than upright -
again because they're lightly loaded.
So if all that hold true then it is just the small matter (gulp) of
establishing the dip of the chassis at the loaded end under these 3
cases of maximum acceleration. If I know the target acceleration in
any direction (1.5G in all directions is about the best I would except
from a set of race tyres on car without downforce) then if I also know
the spring rates, the location of the CofG, the tracks and wheel base,
then it ought to be possible to calculate the roll of the chassis
under acceleration (acceleration being positive when the car gets
quicker, negative under brakes and lateral in a corner).
Knowing the dip of the chassis tells me the amount that the loaded
wheel has been compressed by. For the sake of argument if the car dips
50mm braking at 1.5g then I want the front wheels to be upright when
the suspension is compressed by 50mm.
Ideally I'd like the outside wheel in a bend to be compressed be the
same amount when achieving 1.5g of lateral acceleration. Though quite
how that is achieved is a bridge that is yet to arrive - maybe that is
where rollbars come in?
Ok fire away!
jaho...@googlemail.com
Feb 17, 2008, 4:19:47 PM2/17/08
to Puk7
Well I pasted the previous text here:
http://www.locostbuilders.co.uk/viewthread.php?tid=80566
and
http://www.locostusa.com/forums/viewtopic.php?t=3125&postdays=0&postorder=asc&highlight=suspension+design&start=0
It drew some comments but nothing which rally confirmed this approach
- and much which suggested that it was a compromise anyway. So I spoke
with the race departments at Toyo and Yokohama - and both basically
said you want -2.5deg camber to generate max lateral grip, but 0deg fr
max tyre life.
Now assuming that I am not about to invent some new type of constant
camber suspension mechanism (definitely out of scope for this project)
I can't achieve -2.5deg under cornering loads without having some
negative camber under braking. Which would mean that the tyres
longitudinal grip capability would be reduced. But I think that I can
live with that firstly because the brake load is shared between both
tyres on an axle (where as cornering load is focused on the outside
tyre) and secondly this is the compromise that all wishbone suspension
set ups have to balance anyway.
I discussed weight transfer with a doctor of physics (an old mate of
mine) and we came up with a set of formula for calculating weight
transfer under longitudinal and lateral acceleration. Next job is to
check that they agree with Miliken and then program them into MathCAD.
The objective is to assemble a suit of formula that allow the affect
of different wheelbases, tracks, front:back weight distribution & CG
height to be iterated through to find something that gives a feel for
the weight transfer under different load cases (brake, accel,
cornering).
That data will be used to iterate through different ride rates, anti-
roll rates, anti-dive and anti-squat values to try and identify upper
and lower boundary values for suspension compression for each load
case.
That data can be used to drive geometry design. I'm not expecting a
single set of geometry, but rather a basic set and an amount of
"adjustability". That is how much camber variation is required, how
much rake adjustment...
I'd be very surprised if I don't end up with a set of geometry that
doesn't match current thinking (Caterham CSR data if I can get it) -
but the point is that I'll now have a process for designing the
suspension.
On Feb 17, 9:50 pm, "ja...@netbriller.com" <jahob...@googlemail.com>
wrote:
http://www.locostbuilders.co.uk/viewthread.php?tid=80566
and
http://www.locostusa.com/forums/viewtopic.php?t=3125&postdays=0&postorder=asc&highlight=suspension+design&start=0
It drew some comments but nothing which rally confirmed this approach
- and much which suggested that it was a compromise anyway. So I spoke
with the race departments at Toyo and Yokohama - and both basically
said you want -2.5deg camber to generate max lateral grip, but 0deg fr
max tyre life.
Now assuming that I am not about to invent some new type of constant
camber suspension mechanism (definitely out of scope for this project)
I can't achieve -2.5deg under cornering loads without having some
negative camber under braking. Which would mean that the tyres
longitudinal grip capability would be reduced. But I think that I can
live with that firstly because the brake load is shared between both
tyres on an axle (where as cornering load is focused on the outside
tyre) and secondly this is the compromise that all wishbone suspension
set ups have to balance anyway.
I discussed weight transfer with a doctor of physics (an old mate of
mine) and we came up with a set of formula for calculating weight
transfer under longitudinal and lateral acceleration. Next job is to
check that they agree with Miliken and then program them into MathCAD.
The objective is to assemble a suit of formula that allow the affect
of different wheelbases, tracks, front:back weight distribution & CG
height to be iterated through to find something that gives a feel for
the weight transfer under different load cases (brake, accel,
cornering).
That data will be used to iterate through different ride rates, anti-
roll rates, anti-dive and anti-squat values to try and identify upper
and lower boundary values for suspension compression for each load
case.
That data can be used to drive geometry design. I'm not expecting a
single set of geometry, but rather a basic set and an amount of
"adjustability". That is how much camber variation is required, how
much rake adjustment...
I'd be very surprised if I don't end up with a set of geometry that
doesn't match current thinking (Caterham CSR data if I can get it) -
but the point is that I'll now have a process for designing the
suspension.
On Feb 17, 9:50 pm, "ja...@netbriller.com" <jahob...@googlemail.com>
wrote:
jaho...@googlemail.com
Mar 5, 2008, 4:07:15 AM3/5/08
to Puk7
Have been reading Miliken's Race Car Vehicle Dynamics over the last
few weeks. In my opinion its the definitive book, managing to combine
physics and personal insights to educate the reader. Its a lot easier
going than the aeronautical tomes that we used at university.
I wanted to jot down a couple of key items that will drive suspension
design:
1 Dynamic suspension load can be established from free body analyses
of the car. The variables are wheelbase, track, CG position (including
vertical).
2 The resulting suspension displacement is a function of RC position
and suspension stiffness.
3 RC position - an axle with a pair of equally loaded wheels generates
more lateral force than one with unequal load. A low RC reduces this
weight transfer affect. A high RC introduces a jacking affect -
raising the chassis about its RC. Consider the lever arm between the
CG and the RC - equal length lever arms front and back reduce the
tendency of the chassis to twist and lift a wheel of the road under
hard cornering. So low RCs, with the higher one being at the end with
the highest mass centroid is the target. I have not done enough
analyses to discover whether a subterranean RC is necessarily a bad
thing.
4 Suspension stiffness is a function of spring rates, motion ratios -
I need to do some more work here to establish whether it is sufficient
to know only the motion ratios and establish if the actual suspension
geometry (wishbone lengths and mounting points) has any affect on the
vertical displacement of the wheel under load (I know that it affects
the angle of the wheel to the road under the displacement)
5 Assuming that displacement is independent of geometry we can use the
displacement to drive the design of the geometry:
If we know that the outside front wheel compresses 25mm under target
cornering, and we also know that we need the camber of of the wheel to
be -2.5deg (say) at this lateral acelration then we can contrive the
geometry to achieve this.
Another great insight was that you can use caster to increase camber
when the wheel is turned. Carrol Smith states that a race car front
wheel never turns more than 8 degrees (once it is on the track that
is). So what if turning 8degrees generated -2.5deg camber.
Generally caster is between 5 & 9 degrees. For arguments sake lets
choose 9. So if you turned the front wheel through 90 degrees, the
camber would be -9 degrees. If you turn through 10 degrees - camber
would be 10/90 * 9 = 1degree. Ok its not -2.5 - but it would
contribute.
But as caster increases so does trail - more trail means that it takes
more effort to turn the wheel from the straight ahead. I have no idea
what the ideal force at the steering wheel is - but if we can mimic
what ever a Lotus 7 (not a Locost) does then we won't be far off -
and if we can achieve this with more caster (but similar or less
trail) then we can generate negative camber just when we need it.
In a straight line, for maximum tyre life, camber needs to be 0. Lots
of static negative camber creates a darty car. So, for a road car, we
need to minimise the initial negative camber value.
Turning a wheel around its camber axis drops the outside wheel and
raises the inside one (relative to the chassis). If the net affect is
to raise the CG then this will contribute to the cars self
straightening tendency (good). Droping the outside wheel relative to
the inside one will also twist the chassis - banking it into the turn
(good).
Other ideas: On the track you can sacrifice ride height to reduce CG
height (track being smoother). Part of this can be acheived by
swapping to lower profile tyres, or possible smalled diameter wheels.
Also can wind the springs down (coil overs with adjustable spring
collars) or if using rocker arms - change the rocker. This would also
moves the wheel to a different part of the suspension locust. Maybe we
can combine these two affects to give a different wheel locust for
track than for road...
Next step - model the mass centroid for the car.
King pin inclination reduces camber as the wheel is turned - so KPI
needs to be minimised.
On Feb 17, 10:19 pm, "ja...@netbriller.com" <jahob...@googlemail.com>
wrote:
few weeks. In my opinion its the definitive book, managing to combine
physics and personal insights to educate the reader. Its a lot easier
going than the aeronautical tomes that we used at university.
I wanted to jot down a couple of key items that will drive suspension
design:
1 Dynamic suspension load can be established from free body analyses
of the car. The variables are wheelbase, track, CG position (including
vertical).
2 The resulting suspension displacement is a function of RC position
and suspension stiffness.
3 RC position - an axle with a pair of equally loaded wheels generates
more lateral force than one with unequal load. A low RC reduces this
weight transfer affect. A high RC introduces a jacking affect -
raising the chassis about its RC. Consider the lever arm between the
CG and the RC - equal length lever arms front and back reduce the
tendency of the chassis to twist and lift a wheel of the road under
hard cornering. So low RCs, with the higher one being at the end with
the highest mass centroid is the target. I have not done enough
analyses to discover whether a subterranean RC is necessarily a bad
thing.
4 Suspension stiffness is a function of spring rates, motion ratios -
I need to do some more work here to establish whether it is sufficient
to know only the motion ratios and establish if the actual suspension
geometry (wishbone lengths and mounting points) has any affect on the
vertical displacement of the wheel under load (I know that it affects
the angle of the wheel to the road under the displacement)
5 Assuming that displacement is independent of geometry we can use the
displacement to drive the design of the geometry:
If we know that the outside front wheel compresses 25mm under target
cornering, and we also know that we need the camber of of the wheel to
be -2.5deg (say) at this lateral acelration then we can contrive the
geometry to achieve this.
Another great insight was that you can use caster to increase camber
when the wheel is turned. Carrol Smith states that a race car front
wheel never turns more than 8 degrees (once it is on the track that
is). So what if turning 8degrees generated -2.5deg camber.
Generally caster is between 5 & 9 degrees. For arguments sake lets
choose 9. So if you turned the front wheel through 90 degrees, the
camber would be -9 degrees. If you turn through 10 degrees - camber
would be 10/90 * 9 = 1degree. Ok its not -2.5 - but it would
contribute.
But as caster increases so does trail - more trail means that it takes
more effort to turn the wheel from the straight ahead. I have no idea
what the ideal force at the steering wheel is - but if we can mimic
what ever a Lotus 7 (not a Locost) does then we won't be far off -
and if we can achieve this with more caster (but similar or less
trail) then we can generate negative camber just when we need it.
In a straight line, for maximum tyre life, camber needs to be 0. Lots
of static negative camber creates a darty car. So, for a road car, we
need to minimise the initial negative camber value.
Turning a wheel around its camber axis drops the outside wheel and
raises the inside one (relative to the chassis). If the net affect is
to raise the CG then this will contribute to the cars self
straightening tendency (good). Droping the outside wheel relative to
the inside one will also twist the chassis - banking it into the turn
(good).
Other ideas: On the track you can sacrifice ride height to reduce CG
height (track being smoother). Part of this can be acheived by
swapping to lower profile tyres, or possible smalled diameter wheels.
Also can wind the springs down (coil overs with adjustable spring
collars) or if using rocker arms - change the rocker. This would also
moves the wheel to a different part of the suspension locust. Maybe we
can combine these two affects to give a different wheel locust for
track than for road...
Next step - model the mass centroid for the car.
King pin inclination reduces camber as the wheel is turned - so KPI
needs to be minimised.
On Feb 17, 10:19 pm, "ja...@netbriller.com" <jahob...@googlemail.com>
wrote:
> Well I pasted the previous text here:
>
> http://www.locostbuilders.co.uk/viewthread.php?tid=80566
>
> and
>
> http://www.locostusa.com/forums/viewtopic.php?t=3125&postdays=0&posto...
>
> http://www.locostbuilders.co.uk/viewthread.php?tid=80566
>
> and
>
jaho...@googlemail.com
Mar 19, 2008, 4:16:09 AM3/19/08
to Puk7
Interesting stuff here:
http://locostusa.com/forums/viewtopic.php?t=3030&postdays=0&postorder=asc&start=45
Including
http://puk7.googlegroups.com/web/scrub.jpg?gda=v4hWdToAAACVcKeOksVgN102gmdl4NMoGM1Ehml-pn_qLtm740s01GG1qiJ7UbTIup-M2XPURDRI6uvu86m9-MY0Xr4t5_QM
On Mar 5, 10:07 am, "ja...@netbriller.com" <jahob...@googlemail.com>
wrote:
http://locostusa.com/forums/viewtopic.php?t=3030&postdays=0&postorder=asc&start=45
Including
http://puk7.googlegroups.com/web/scrub.jpg?gda=v4hWdToAAACVcKeOksVgN102gmdl4NMoGM1Ehml-pn_qLtm740s01GG1qiJ7UbTIup-M2XPURDRI6uvu86m9-MY0Xr4t5_QM
On Mar 5, 10:07 am, "ja...@netbriller.com" <jahob...@googlemail.com>
wrote:
Reply all
Reply to author
Forward
0 new messages