Suspension designing from scratch

[TEAM78]

Hi, ive got a clean sheet of paper for designing suspension on a single seater FS car. I need to design the front and rear suspension and there are no wings (hence only drag forces here). How do designers go about designing suspension from scratch. What aspect of suspension do they approach first. any comments please as im a little lost. but have so far got this in mind.

1) decide what tyres are to be used front and rear and figure out their force capabilities
2) locate the C of G then figure out the roll moment distribution
3) Do a kinematic analysis (including RC)
4) do a force analysis
5) itnerate between 3 & 4
6) ride analysis

what do you guys reckon then???

[boyracer]

Yeah, sounds simple when you put it like that. List looks to be in the right order, but I think the actual practice could be a little more interesting. Be sure to allow lots of scope for adjustment if things go pear shape.
Rather then start from scratch I'd look at a design that you know works, or at least kinda works figure out what it's doing, reason to yourself about what it does right and what it does wrong and it does it because of this, then set about fixing it while trying not to fall into any of the other traps.
If you'r attempting this sort of project I presume you know about the handling triangle and it all bein g a compromise, you just have to figure out what you want to compromise on
Good luck

[ss_collins]

I reckon by reading Decembers racecar, we have a piece on tyre usage on FS cars, basically there are some widely held misconceptions about FS suspension design that perhaps are dispelled... One result amazed me, because its totally not what I thought.

Read it take it in, and you may pick up some good design points, remember yopu should be able to justify everything you have done. Tyres are really key - check PVV on the contact patch work done by Georgia Tech (I think it was them - yellow car)

[HiRich]

I would suggest Step 0: Define your basic performance targets and parameters.
- Estimates for sprung & unsprung weight
- CoG axis
- Design load (at max bump travel)
- Bump and rebound travel allowances (full travel and "normal operating range")
- Acceptable roll
- Target lateral G
- Acceptable camber change ("normal operating range" and also "bump & roll (brake & turn)", "rebound & roll (power-on & turn)" and "kerb clip (step input while already in roll), as these cover different driving situations)
There may be more, but it's a long time since I did it.
For starters these are only estimates, and may change if circumstances change (e.g. a better engine has a lower CoG), or you feel you can increase or decrease the envelope. But do:
- Write them down, and justify them
- If you do decide to change them, stop work, calculate the impacts, and reset everything.

[TEAM78]

Hi, thanks for the quick responses and advice. whats the handling triangle boyracer? Ive heard of it as being a trade of between three aspects of performance. n yep i will be digging out my december issue of Racecar eng as im open to any advice and hope that i dont get carried away into thinking a certain value is a must have. Ive also been building a picture library of suspensions on race cars to help give me ideas. there are people saying that you need full camber compensation but I need to prove to my self if this is right or not.

[ubrben]

Big subject!

First a quick correction to Sam's post - it was Universty of Western Australia who did the contact patch pressure prints.

Second, a suggestion - register at www.fsae.com and read the forums regularly as well as the archive.

Third, Go to Learn to Win at Silverstone on 29th - 30th of this month and talk to some design judges. This design judge will be there on the 30th wearing some yellow tyre-related clothing :-)

Without writing an essay on the subject I would suggest that the most important thing is that your car works well as a complete system. No one suspension parameter is worth sacrificing the rest of the car for. Make sure you consider how everything else must fit around the suspension.

In terms of actual design I'd worry more about camber control than obsessing about kinematic roll centre migrations. The kinematic roll centre is not a motion centre.

My philosophy was something like the following:

- Long swing arm length at the front to keep the wheel upright during braking
- Large caster angle (we had 8 on my second car to get camber compensation in corners
- Shorter rear swing arm length (as much as half if you can package it) for camber compensation. Keeping the wheel upright on the brakes is less important at the rear because they only generate 30% of the force
- Linear installation ratios for simplicity
- Rear roll centre height 2 - 2.5 times higher than the front to ensure decent phasing of front and rear load transfer
- As much Ackerman geometry as you can package - you'll find what the limiting factor is fairly quickly

As I said, I would use static roll centre height as a guide to the inelastic load transfer and not obsess about force based modelling - you won't have time anyway unless you are already an expert user in ADAMS/Car. Building an integrated car on time and testing it will always gain you more points in the competition.

Apologies - I think that turned into an essay. Let me know what you think.

Ben

[GordonG]

Ben

You mention using a lot of camber compensation at the rear as camber change in braking is not so much of a problem - what about camber change in squat under power - I would have thought getting the power down is more important in a lot of classes (and especially FS) than braking and arguably even cornering.

Having little in the way of camber compensation at the front may mean less camber change in dive, but will you not end up with more static camber anyway to make the tyres work in the corner, so you end up with similar braking camber either way?

I spose the large caster recovers the camber to some extent in corners, but that will vary camber greatly depending on corner radius/steering angle, while the G (and hence presumably the camber required) at best stays the same as corner radius decreases and on a downforce car will get less, as camber induced from the caster increases.

I'm not qualified to say you're wrong and that's not what I'm trying to do - just trying to further my understanding. Thanks

G

[ubrben]

Originally Posted by GordonG

Ben

You mention using a lot of camber compensation at the rear as camber change in braking is not so much of a problem - what about camber change in squat under power - I would have thought getting the power down is more important in a lot of classes (and especially FS) than braking and arguably even cornering.

Having little in the way of camber compensation at the front may mean less camber change in dive, but will you not end up with more static camber anyway to make the tyres work in the corner, so you end up with similar braking camber either way?

I spose the large caster recovers the camber to some extent in corners, but that will vary camber greatly depending on corner radius/steering angle, while the G (and hence presumably the camber required) at best stays the same as corner radius decreases and on a downforce car will get less, as camber induced from the caster increases.

I'm not qualified to say you're wrong and that's not what I'm trying to do - just trying to further my understanding. Thanks

G

Good points Gordon.

The important thing to remember is that a RWD FS car is traction limited to about 0.9-1.0 G but can brake at 1.5-1.6G. So any longitudinal load transfer will be higher on the brakes than under acceleration. Th other thing we did was to set up the car with a load of positive camber at the rear for the accleration event so that under about 1G accel they would tend to zero camber with the squat.

You may have a point on the camber with steer. Some of the benefit of the caster could well have been the de-wedging effect giving us good turn-in response. I never had the time or the data logging resources to separate the effects.

Regardless of the pros and cons of a given approach it is vital to have a consistant philosophy underpinning it. The OP has got the right idea on that front. Remember that why you did something is as important as what you actually did as far as the design judges are concerned.

BTW - we were running cross plies that don't like static negative camber. Should have mentioned that - sorry.

Ben

[TEAM78]

Thanks for the advice guys. I will be going to Silverstone next weekend on Saturday as that’s when the seminars are and I am looking forward to the December issue of Racecar Engineering. I also found in an earlier copy (March) a lot of info on suspension design, great article.

One more question though and that is what processes are there available to help aid a designer in prioritizing what’s important and what’s not for particular circuits? I’m aware there are track simulation packages that can help decide where the biggest improvements can be had but are often limited to basic things. Are there probability methods that can help maybe??

[ubrben]

Originally Posted by TEAM78

Thanks for the advice guys. I will be going to Silverstone next weekend on Saturday as that’s when the seminars are and I am looking forward to the December issue of Racecar Engineering. I also found in an earlier copy (March) a lot of info on suspension design, great article.

One more question though and that is what processes are there available to help aid a designer in prioritizing what’s important and what’s not for particular circuits? I’m aware there are track simulation packages that can help decide where the biggest improvements can be had but are often limited to basic things. Are there probability methods that can help maybe??

I wouldn't worry too much about track optimisation. That's a bit subtle for Formula Student. A bit of lap time simulation wouldn't hurt for the design event to show you're aware of the technique, but the biggest gain in lap time will be from the driver if you get a lot of testing done.

Ben

[nopanon]

Hi!

Interesting thread.

Apart from the design steps of the suspension geometry itself, I would like to share some experience on the more practical stuff from last year when I designed the suspension for an FS car.

First, the suspension and spring/damper geometry has big influence on the design of the rest of the car, and thus a lot of people are dependent on you specifying a geometry within reasonable time. To be as helpful as possible to your fellow team mates(and yourself) I would go about in the following way:

1. Decide wheel (rim) diameter and preferably also width.

2. Decide on overall parameters such as wheelbase and trackwidth

3. Start with the geometry in the wheels. I think you can decide upon scrub radius and trail without having bothered about the rest of the suspension. With some estimate on caster and max allowed kingpin you and your friends can go about the packaging quest of the upright assemblies. deciding upon steering throw on this point could be a good idea to give constraints on how "wide" your wishbones can be.

4. With the wheel layout roughly decided you can see approximately where the wishbones will be heading. This will give the frame group something to start with.

5. Look into damper mounting together with frame guys. At this point it's a good idea to know what dampers you will be using as to know how much space is required

6. Do some force calculations. Now there should be information enough to look into rodends, bolts and other practical stuff. This is very important as you can't really design the suspension in detail without knowing how things will interact in reality, and things like rodends can impose constraints on your design with regards to angular displacement for example

7. Go about the suspension in more detail in parallel with the design of the suspension parts.

"Some" iteration 3-7. Maybe not so much required on no. 6

In general i think that one shouldn't be too afraid of giving preliminary and in the end maybe not correct information to other team members on the suspension. They, just as you, will probably have to rework their design several times, and its better they get something concrete in their hands so they can start thinking and working, rather than doing nothing.

Information like reasonably exact CoG location and roll moment distribution you will not be able to obtain until you've designed the whole car probably. And it's a lot of work cading everything "correct" and to put the correct density everywhere. Of course you can have wishes regarding these parameters(and should,at least the CoG) but for it to end up exactly where you wish is alot of work.

Uncertain component delivery is very bad both for the design process and the enthusiasm among project participants. Then it's better to directly settle for a component you certainly can obtain in time even though the performance on paper is worse. For the project it's the better component.

This is just from my experience, if you have thoughts about it please let me know!

As you maybe have guessed, I did it almost completely the other way around

Good luck!

Author is in no way responsible for the consequences of applying suggestions stated in this text

[ubrben]

Good post.

I think the best point made there was the importance of iterating the design. Give the other guys something to work with and then debate the next move when you've got something to look at.

The best example we had of this was I designed the rear suspension geometry and another guy designed the diff assembly and we put them in a full car CATIA model. Then with me on an ADAMS model and him on the CATIA model he would measure the clearances between the components and suggest where to move a wishbone pickup. I'd then do an ADAMS kinematics run and suggest whether it was ok or offer an alternative option, he'd then change the CATIA model...

In the end I had to settle for a longer swing arm length than I wanted, but the packaging was extremely tidy.

A good working relationship with your fellow designers is vital.

Ben

[ss_collins]

I'll be there at some point during the weekend - though I won't have the Dec issue with me I don't think...

any feedback or ideas for the mag will be appreciated

[ubrben]

Originally Posted by ss_collins

I'll be there at some point during the weekend -

Saturday or Sunday? I'm planning for Sunday (would like to see some cars running)

Ben

[ss_collins]

was thinking sunday too but then I've seen them all run before - and I'd be interested in the seminars too...

[chezza]

You only wanna be there on the Saturday to have a go with Syd on the main circuit!!

[ubrben]

Saturday sounds like a plan then.

Ben

[TEAM78]

Thanks for the advice guys

[TEAM78]

I will be there Saturday myself

[ss_collins]

not sure.

[ubrben]

I'm doing Sunday now. Have a meeting with Brian Robinson the organisor and Sunday was best for him. Need Saturday to sort life out before final race of season.

Ben

[ss_collins]

I think saturday because the seminars are more relevant to what I need

[ss_collins]

oops! tyre testing feature is in Jan's issue now... apologies for the confusion

[Goran Malmberg]

I like to build a car as neutral as possible, to become easely tuned for different track condition. What has been left out of the balancing discussion is the influece of Sai angle, scrub distance and upper A-arm angle on wheel rate. Wheelrate also become non-linear and therfore has a bad influence on balance during suspension travel. It is therfore a good idea to keep those figures the same for both front and rear geometry. But we seldome see any discussion about, for example, rear scrub distance.
Goran Malmberg

[ger80]

Maybe we should start a open car developing project?
Step files are readable for many persons, but not parametric. Should be use only one CAD software? Witch?
Whats the best suspension software?