McPherson suspension geometry

[Kozy]

Originally Posted by nrostedt

The toe settings have always been at 0 in all four corners. But during sessions we have fiddled with camber settings, and after that we didn`t have the equipment to reset toe. So they may have been a bit off.

IME altering the camber has a significant effect on toe angles, I went from -1° to -2.5° on ym double wishbine car and got about 2° total toe in, a friend with MacPhersons went from about -.5° to -1.5° and got about 2.5° toe in. That's over 15mm if you prefer it that way. I got mine re-aligned the same day so it was no issue, my friend didn't and wore out his brand new AD08s in a few thousand miles.

Originally Posted by nrostedt

I haven`t really taken this into concideration, and it actually helps me quite a lot. With a little steering input and right amount of caster, I`m actually getting camber net loss quite close to zero.

Too bad the program doesn`t let me do combined bumb, roll and steer analysis.

If you have the measurements, I can do this for you. I have a 3D model that combine steering, bump, roll and even zero-droop roll to give dynamic camber and toe angles. Unlike a LOT of supposedly professional (i.e. pricey) software, my model is built on 'natural roll' i.e. it doesn't 'roll around the roll centre'. The roll center is simply a number that it gives you, and then it's only height, the lateral position is irrelevant.

Here's a screen shot of the model for my autocross car, simulating front end lift, roll and steering on corner exit (where I was trying to optimise front end grip).



Let me know if you would like me to have a look at this problem for your car.

[nrostedt]

Originally Posted by Kozy

Let me know if you would like me to have a look at this problem for your car.

I would be great if you would!!

Can you do something with Lotus shark file (.shk)?
Or can I send you the x,y,z cordinates of suspension hardpoints?

It would be good to know how the tyres react when you combine bumb, roll and steering....

Did you make this program yourself?

[Kozy]

Originally Posted by nrostedt

I would be great if you would!!

Can you do something with Lotus shark file (.shk)?
Or can I send you the x,y,z cordinates of suspension hardpoints?

It would be good to know how the tyres react when you combine bumb, roll and steering....

Did you make this program yourself?

The Lotus file is of no use to me unfortunately, XYZ co-ordinates I can work with though.

It is essentially just a wireframe CAD model with a lot of constraints to make it behave like a suspension system, but yes I created it myself.

I shall PM you my email address for the data.

[Flight909]

Originally Posted by Kozy

Unlike a LOT of supposedly professional (i.e. pricey) software, my model is built on 'natural roll' i.e. it doesn't 'roll around the roll centre'. The roll center is simply a number that it gives you, and then it's only height, the lateral position is irrelevant.


Let me know if you would like me to have a look at this problem for your car.

What is "natural roll"? How do you define this?

[Kozy]

A real car rolls according to how much load the springs see, not around any pre-defined point. To roll my model you change the lengths of the spring/damper assembly, and the car takes a set as it would in a corner. There is no input for roll angle, you need to know the suspension loads and the roll angle is resultant.

[phoenix]

Originally Posted by Kozy

A real car rolls according to how much load the springs see, not around any pre-defined point. To roll my model you change the lengths of the spring/damper assembly, and the car takes a set as it would in a corner. There is no input for roll angle, you need to know the suspension loads and the roll angle is resultant.

A real car does roll about it's roll axis! This is the virtual line between front and rear roll centres.

However, by your method you can certainly calculate the camber changes, roll angle etc. from the suspension displacement, so your method is absolutely fine for what you want to do in your model.

But what you cannot calculate is what the suspension displacement will be in roll, or the roll resistance of the springs and any roll bars, without knowing where the roll centre is.

Having said that, although I wouldn't like to do the maths, I am sure that you could work back from the suspension deflection and determine where the roll centre is at either end of the car, if you know the mean height and the location for/aft of the centre of mass as well as the data you already need for the model to do what is does now.

[Kozy]

I should have perhaps elaborated a bit. The 3D model is supplemented by a fairly comprehensive spreadsheet detailing load transfers (it will even account for wheel lift and zero droop setups), the roll center is found for the suspension geometry in the model, and the measurement added to the spreadsheet along with the various data for springs, ARBs and other relevant parameters. For a given acceleration level, recorded in action from a datalogger, I can then work out the suspension loads and input that data back into the model as spring deflections, which then rolls the model and gives me back the dynamic alignments. It's a bit long winded I know, however I could not afford one of the proper programs so I've had a shot at making one myself. Just looking for a way to integrate the spreadsheet into the model now!

I can assure you though, that with all this calculated and modelled, the suspension does not roll around the roll center. I rolled the model using my method, and rolled it around the roll center, and the results were different.

[phoenix]

Originally Posted by Kozy

I can assure you though, that with all this calculated and modelled, the suspension does not roll around the roll center. I rolled the model using my method, and rolled it around the roll center, and the results were different.

I think the reason for that is the roll centre is a moving target! As the car rolls, the angles of the suspension links change and so the position of the roll centre location changes. Maybe you could confirm this: If you calculate the position of the roll centre with your spreadsheet once the car has adopted a certain roll angle, is should agree with the 'real' roll centre you are seeing in your other model. I can't see why it wouldn't...

[Kozy]

Originally Posted by phoenix

I think the reason for that is the roll centre is a moving target! As the car rolls, the angles of the suspension links change and so the position of the roll centre location changes. Maybe you could confirm this: If you calculate the position of the roll centre with your spreadsheet once the car has adopted a certain roll angle, is should agree with the 'real' roll centre you are seeing in your other model. I can't see why it wouldn't...

My spreadsheet cannot calculate the roll centre location, it is an input which is taken from the model. The roll center does indeed move with displacement, although it stays pretty much at the same height on the model for my car. Modelling roll for a car with significant vertical movement may be tricky!

[nrostedt]

A question about ackermann...

What would be a good value to have in a racing car and what kind of angles do they have in normal road-going vehicles?

If I put 20mm of steering input in my current model, the inside wheel turns ~11,5 deg and outside turns ~10,9 deg. With that the turning circle is 12,5m.
The program also gives a percentage for ackermann and with those values it`s 40,2%.

Can somebody tell me if this is over or under normal values?

[GORDON STREETER]

A lot depends on what type of racing is going to be done, whether on a twisty circuit or a fast flowing one.
If you can make it adjustable you would be surprised at the difference it can make from circuit to circuit if the cars are very similar in power and type.
A good web site is this http://www.smithees-racetech.com.au/ackerman.html

[Cameron Winton]

Is part of the answer more static caster? Would at least help minimise the loss of camber as the wheel turns?

[GORDON STREETER]

The thing is with adjusting "anything" to do with steering/handling is that there is always a trade off somewhere, if you have a car with loads of downforce then you are into another world of set ups.
Although there are "ballpark" figures ultimately it comes down to how the driver prefers the car to handle in different situations.